From 00017c90ae82d3237f5a4784caa695459c5897d2 Mon Sep 17 00:00:00 2001 From: Øyvind Harboe Date: Tue, 15 Jun 2010 22:47:52 +0200 Subject: jim: -Wshadow warning fix MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Øyvind Harboe --- src/helper/#jim.c# | 12823 +++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 12823 insertions(+) create mode 100644 src/helper/#jim.c# (limited to 'src/helper/#jim.c#') diff --git a/src/helper/#jim.c# b/src/helper/#jim.c# new file mode 100644 index 00000000..c8583de8 --- /dev/null +++ b/src/helper/#jim.c# @@ -0,0 +1,12823 @@ +/* Jim - A small embeddable Tcl interpreter + * + * Copyright 2005 Salvatore Sanfilippo + * Copyright 2005 Clemens Hintze + * Copyright 2005 patthoyts - Pat Thoyts + * Copyright 2008,2009 oharboe - yvind Harboe - oyvind.harboe@zylin.com + * Copyright 2008 Andrew Lunn + * Copyright 2008 Duane Ellis + * Copyright 2008 Uwe Klein + * Copyright 2008 Steve Bennett + * Copyright 2009 Nico Coesel + * Copyright 2009 Zachary T Welch zw@superlucidity.net + * Copyright 2009 David Brownell + * + * The FreeBSD license + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE JIM TCL PROJECT ``AS IS'' AND ANY + * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, + * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A + * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE + * JIM TCL PROJECT OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, + * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES + * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF + * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * The views and conclusions contained in the software and documentation + * are those of the authors and should not be interpreted as representing + * official policies, either expressed or implied, of the Jim Tcl Project. + **/ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define __JIM_CORE__ +#define JIM_OPTIMIZATION /* comment to avoid optimizations and reduce size */ + +#ifdef __ECOS +#include +#include +#include + +typedef CYG_ADDRWORD intptr_t; + +#include +#include +#include +#include +#include +#include +#include +#endif +#ifndef JIM_ANSIC +#define JIM_DYNLIB /* Dynamic library support for UNIX and WIN32 */ +#endif /* JIM_ANSIC */ + +#include +#include + +/* Include the platform dependent libraries for + * dynamic loading of libraries. */ +#ifdef JIM_DYNLIB +#if defined(_WIN32) || defined(WIN32) +#ifndef WIN32 +#define WIN32 1 +#endif +#ifndef STRICT +#define STRICT +#endif +#define WIN32_LEAN_AND_MEAN +#include +#if _MSC_VER >= 1000 +#pragma warning(disable:4146) +#endif /* _MSC_VER */ +#else +#include +#endif /* WIN32 */ +#endif /* JIM_DYNLIB */ + +#ifdef __ECOS +#include +#else +#include "jim.h" +#endif + +#ifdef HAVE_BACKTRACE +#include +#endif + +/* ----------------------------------------------------------------------------- + * Global variables + * ---------------------------------------------------------------------------*/ + +/* A shared empty string for the objects string representation. + * Jim_InvalidateStringRep knows about it and don't try to free. */ +static char *JimEmptyStringRep = (char*) ""; + +/* ----------------------------------------------------------------------------- + * Required prototypes of not exported functions + * ---------------------------------------------------------------------------*/ +static void JimChangeCallFrameId(Jim_Interp *interp, Jim_CallFrame *cf); +static void JimFreeCallFrame(Jim_Interp *interp, Jim_CallFrame *cf, int flags); +static void JimRegisterCoreApi(Jim_Interp *interp); + +static Jim_HashTableType *getJimVariablesHashTableType(void); + +/* ----------------------------------------------------------------------------- + * Utility functions + * ---------------------------------------------------------------------------*/ + +static char * +jim_vasprintf(const char *fmt, va_list ap) +{ +#ifndef HAVE_VASPRINTF + /* yucky way */ +static char buf[2048]; + vsnprintf(buf, sizeof(buf), fmt, ap); + /* garentee termination */ + buf[sizeof(buf)-1] = 0; +#else + char *buf; + int result; + result = vasprintf(&buf, fmt, ap); + if (result < 0) exit(-1); +#endif + return buf; +} + +static void +jim_vasprintf_done(void *buf) +{ +#ifndef HAVE_VASPRINTF + (void)(buf); +#else + free(buf); +#endif +} + + +/* + * Convert a string to a jim_wide INTEGER. + * This function originates from BSD. + * + * Ignores `locale' stuff. Assumes that the upper and lower case + * alphabets and digits are each contiguous. + */ +#ifdef HAVE_LONG_LONG_INT +#define JimIsAscii(c) (((c) & ~0x7f) == 0) +static jim_wide JimStrtoll(const char *nptr, char **endptr, register int base) +{ + register const char *s; + register unsigned jim_wide acc; + register unsigned char c; + register unsigned jim_wide qbase, cutoff; + register int neg, any, cutlim; + + /* + * Skip white space and pick up leading +/- sign if any. + * If base is 0, allow 0x for hex and 0 for octal, else + * assume decimal; if base is already 16, allow 0x. + */ + s = nptr; + do { + c = *s++; + } while (isspace(c)); + if (c == '-') { + neg = 1; + c = *s++; + } else { + neg = 0; + if (c == '+') + c = *s++; + } + if ((base == 0 || base == 16) && + c == '0' && (*s == 'x' || *s == 'X')) { + c = s[1]; + s += 2; + base = 16; + } + if (base == 0) + base = c == '0' ? 8 : 10; + + /* + * Compute the cutoff value between legal numbers and illegal + * numbers. That is the largest legal value, divided by the + * base. An input number that is greater than this value, if + * followed by a legal input character, is too big. One that + * is equal to this value may be valid or not; the limit + * between valid and invalid numbers is then based on the last + * digit. For instance, if the range for quads is + * [-9223372036854775808..9223372036854775807] and the input base + * is 10, cutoff will be set to 922337203685477580 and cutlim to + * either 7 (neg == 0) or 8 (neg == 1), meaning that if we have + * accumulated a value > 922337203685477580, or equal but the + * next digit is > 7 (or 8), the number is too big, and we will + * return a range error. + * + * Set any if any `digits' consumed; make it negative to indicate + * overflow. + */ + qbase = (unsigned)base; + cutoff = neg ? (unsigned jim_wide)-(LLONG_MIN + LLONG_MAX) + LLONG_MAX + : LLONG_MAX; + cutlim = (int)(cutoff % qbase); + cutoff /= qbase; + for (acc = 0, any = 0;; c = *s++) { + if (!JimIsAscii(c)) + break; + if (isdigit(c)) + c -= '0'; + else if (isalpha(c)) + c -= isupper(c) ? 'A' - 10 : 'a' - 10; + else + break; + if (c >= base) + break; + if (any < 0 || acc > cutoff || (acc == cutoff && c > cutlim)) + any = -1; + else { + any = 1; + acc *= qbase; + acc += c; + } + } + if (any < 0) { + acc = neg ? LLONG_MIN : LLONG_MAX; + errno = ERANGE; + } else if (neg) + acc = -acc; + if (endptr != 0) + *endptr = (char *)(any ? s - 1 : nptr); + return (acc); +} +#endif + +/* Glob-style pattern matching. */ +static int JimStringMatch(const char *pattern, int patternLen, + const char *string, int stringLen, int nocase) +{ + while (patternLen) { + switch (pattern[0]) { + case '*': + while (pattern[1] == '*') { + pattern++; + patternLen--; + } + if (patternLen == 1) + return 1; /* match */ + while (stringLen) { + if (JimStringMatch(pattern + 1, patternLen-1, + string, stringLen, nocase)) + return 1; /* match */ + string++; + stringLen--; + } + return 0; /* no match */ + break; + case '?': + if (stringLen == 0) + return 0; /* no match */ + string++; + stringLen--; + break; + case '[': + { + int not, match; + + pattern++; + patternLen--; + not = pattern[0] == '^'; + if (not) { + pattern++; + patternLen--; + } + match = 0; + while (1) { + if (pattern[0] == '\\') { + pattern++; + patternLen--; + if (pattern[0] == string[0]) + match = 1; + } else if (pattern[0] == ']') { + break; + } else if (patternLen == 0) { + pattern--; + patternLen++; + break; + } else if (pattern[1] == '-' && patternLen >= 3) { + int start = pattern[0]; + int end = pattern[2]; + int c = string[0]; + if (start > end) { + int t = start; + start = end; + end = t; + } + if (nocase) { + start = tolower(start); + end = tolower(end); + c = tolower(c); + } + pattern += 2; + patternLen -= 2; + if (c >= start && c <= end) + match = 1; + } else { + if (!nocase) { + if (pattern[0] == string[0]) + match = 1; + } else { + if (tolower((int)pattern[0]) == tolower((int)string[0])) + match = 1; + } + } + pattern++; + patternLen--; + } + if (not) + match = !match; + if (!match) + return 0; /* no match */ + string++; + stringLen--; + break; + } + case '\\': + if (patternLen >= 2) { + pattern++; + patternLen--; + } + /* fall through */ + default: + if (!nocase) { + if (pattern[0] != string[0]) + return 0; /* no match */ + } else { + if (tolower((int)pattern[0]) != tolower((int)string[0])) + return 0; /* no match */ + } + string++; + stringLen--; + break; + } + pattern++; + patternLen--; + if (stringLen == 0) { + while (*pattern == '*') { + pattern++; + patternLen--; + } + break; + } + } + if (patternLen == 0 && stringLen == 0) + return 1; + return 0; +} + +int JimStringCompare(const char *s1, int l1, const char *s2, int l2, + int nocase) +{ + unsigned char *u1 = (unsigned char*) s1, *u2 = (unsigned char*) s2; + + if (nocase == 0) { + while (l1 && l2) { + if (*u1 != *u2) + return (int)*u1-*u2; + u1++; u2++; l1--; l2--; + } + if (!l1 && !l2) return 0; + return l1-l2; + } else { + while (l1 && l2) { + if (tolower((int)*u1) != tolower((int)*u2)) + return tolower((int)*u1)-tolower((int)*u2); + u1++; u2++; l1--; l2--; + } + if (!l1 && !l2) return 0; + return l1-l2; + } +} + +/* Search 's1' inside 's2', starting to search from char 'index' of 's2'. + * The index of the first occurrence of s1 in s2 is returned. + * If s1 is not found inside s2, -1 is returned. */ +int JimStringFirst(const char *s1, int l1, const char *s2, int l2, int index_t) +{ + int i; + + if (!l1 || !l2 || l1 > l2) return -1; + if (index_t < 0) index_t = 0; + s2 += index_t; + for (i = index_t; i <= l2-l1; i++) { + if (memcmp(s2, s1, l1) == 0) + return i; + s2++; + } + return -1; +} + +int Jim_WideToString(char *buf, jim_wide wideValue) +{ + const char *fmt = "%" JIM_WIDE_MODIFIER; + return sprintf(buf, fmt, wideValue); +} + +int Jim_StringToWide(const char *str, jim_wide *widePtr, int base) +{ + char *endptr; + +#ifdef HAVE_LONG_LONG_INT + *widePtr = JimStrtoll(str, &endptr, base); +#else + *widePtr = strtol(str, &endptr, base); +#endif + if ((str[0] == '\0') || (str == endptr)) + return JIM_ERR; + if (endptr[0] != '\0') { + while (*endptr) { + if (!isspace((int)*endptr)) + return JIM_ERR; + endptr++; + } + } + return JIM_OK; +} + +int Jim_StringToIndex(const char *str, int *intPtr) +{ + char *endptr; + + *intPtr = strtol(str, &endptr, 10); + if ((str[0] == '\0') || (str == endptr)) + return JIM_ERR; + if (endptr[0] != '\0') { + while (*endptr) { + if (!isspace((int)*endptr)) + return JIM_ERR; + endptr++; + } + } + return JIM_OK; +} + +/* The string representation of references has two features in order + * to make the GC faster. The first is that every reference starts + * with a non common character '~', in order to make the string matching + * fater. The second is that the reference string rep his 32 characters + * in length, this allows to avoid to check every object with a string + * repr < 32, and usually there are many of this objects. */ + +#define JIM_REFERENCE_SPACE (35 + JIM_REFERENCE_TAGLEN) + +static int JimFormatReference(char *buf, Jim_Reference *refPtr, jim_wide id) +{ + const char *fmt = ".%020" JIM_WIDE_MODIFIER ">"; + sprintf(buf, fmt, refPtr->tag, id); + return JIM_REFERENCE_SPACE; +} + +int Jim_DoubleToString(char *buf, double doubleValue) +{ + char *s; + int len; + + len = sprintf(buf, "%.17g", doubleValue); + s = buf; + while (*s) { + if (*s == '.') return len; + s++; + } + /* Add a final ".0" if it's a number. But not + * for NaN or InF */ + if (isdigit((int)buf[0]) + || ((buf[0] == '-' || buf[0] == '+') + && isdigit((int)buf[1]))) { + s[0] = '.'; + s[1] = '0'; + s[2] = '\0'; + return len + 2; + } + return len; +} + +int Jim_StringToDouble(const char *str, double *doublePtr) +{ + char *endptr; + + *doublePtr = strtod(str, &endptr); + if (str[0] == '\0' || endptr[0] != '\0' || (str == endptr)) + return JIM_ERR; + return JIM_OK; +} + +static jim_wide JimPowWide(jim_wide b, jim_wide e) +{ + jim_wide i, res = 1; + if ((b == 0 && e != 0) || (e < 0)) return 0; + for (i = 0; i < e; i++) {res *= b;} + return res; +} + +/* ----------------------------------------------------------------------------- + * Special functions + * ---------------------------------------------------------------------------*/ + +/* Note that 'interp' may be NULL if not available in the + * context of the panic. It's only useful to get the error + * file descriptor, it will default to stderr otherwise. */ +void Jim_Panic(Jim_Interp *interp, const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + /* + * Send it here first.. Assuming STDIO still works + */ + fprintf(stderr, JIM_NL "JIM INTERPRETER PANIC: "); + vfprintf(stderr, fmt, ap); + fprintf(stderr, JIM_NL JIM_NL); + va_end(ap); + +#ifdef HAVE_BACKTRACE + { + void *array[40]; + int size, i; + char **strings; + + size = backtrace(array, 40); + strings = backtrace_symbols(array, size); + for (i = 0; i < size; i++) + fprintf(fp,"[backtrace] %s" JIM_NL, strings[i]); + fprintf(fp,"[backtrace] Include the above lines and the output" JIM_NL); + fprintf(fp,"[backtrace] of 'nm ' in the bug report." JIM_NL); + } +#endif + + /* This may actually crash... we do it last */ + if (interp && interp->cookie_stderr) { + Jim_fprintf(interp, interp->cookie_stderr, JIM_NL "JIM INTERPRETER PANIC: "); + Jim_vfprintf(interp, interp->cookie_stderr, fmt, ap); + Jim_fprintf(interp, interp->cookie_stderr, JIM_NL JIM_NL); + } + abort(); +} + +/* ----------------------------------------------------------------------------- + * Memory allocation + * ---------------------------------------------------------------------------*/ + +/* Macro used for memory debugging. + * In order for they to work you have to rename Jim_Alloc into _Jim_Alloc + * and similary for Jim_Realloc and Jim_Free */ +#if 0 +#define Jim_Alloc(s) (printf("%s %d: Jim_Alloc(%d)\n",__FILE__,__LINE__,s),_Jim_Alloc(s)) +#define Jim_Free(p) (printf("%s %d: Jim_Free(%p)\n",__FILE__,__LINE__,p),_Jim_Free(p)) +#define Jim_Realloc(p,s) (printf("%s %d: Jim_Realloc(%p,%d)\n",__FILE__,__LINE__,p,s),_Jim_Realloc(p,s)) +#endif + +void *Jim_Alloc(int size) +{ + /* We allocate zero length arrayes, etc. to use a single orthogonal codepath */ + if (size == 0) + size = 1; + void *p = malloc(size); + if (p == NULL) + Jim_Panic(NULL,"malloc: Out of memory"); + return p; +} + +void Jim_Free(void *ptr) { + free(ptr); +} + +void *Jim_Realloc(void *ptr, int size) +{ + /* We allocate zero length arrayes, etc. to use a single orthogonal codepath */ + if (size == 0) + size = 1; + void *p = realloc(ptr, size); + if (p == NULL) + Jim_Panic(NULL,"realloc: Out of memory"); + return p; +} + +char *Jim_StrDup(const char *s) +{ + int l = strlen(s); + char *copy = Jim_Alloc(l + 1); + + memcpy(copy, s, l + 1); + return copy; +} + +char *Jim_StrDupLen(const char *s, int l) +{ + char *copy = Jim_Alloc(l + 1); + + memcpy(copy, s, l + 1); + copy[l] = 0; /* Just to be sure, original could be substring */ + return copy; +} + +/* ----------------------------------------------------------------------------- + * Time related functions + * ---------------------------------------------------------------------------*/ +/* Returns microseconds of CPU used since start. */ +static jim_wide JimClock(void) +{ +#if (defined WIN32) && !(defined JIM_ANSIC) + LARGE_INTEGER t, f; + QueryPerformanceFrequency(&f); + QueryPerformanceCounter(&t); + return (long)((t.QuadPart * 1000000) / f.QuadPart); +#else /* !WIN32 */ + clock_t clocks = clock(); + + return (long)(clocks*(1000000/CLOCKS_PER_SEC)); +#endif /* WIN32 */ +} + +/* ----------------------------------------------------------------------------- + * Hash Tables + * ---------------------------------------------------------------------------*/ + +/* -------------------------- private prototypes ---------------------------- */ +static int JimExpandHashTableIfNeeded(Jim_HashTable *ht); +static unsigned int JimHashTableNextPower(unsigned int size); +static int JimInsertHashEntry(Jim_HashTable *ht, const void *key); + +/* -------------------------- hash functions -------------------------------- */ + +/* Thomas Wang's 32 bit Mix Function */ +unsigned int Jim_IntHashFunction(unsigned int key) +{ + key += ~(key << 15); + key ^= (key >> 10); + key += (key << 3); + key ^= (key >> 6); + key += ~(key << 11); + key ^= (key >> 16); + return key; +} + +/* Identity hash function for integer keys */ +unsigned int Jim_IdentityHashFunction(unsigned int key) +{ + return key; +} + +/* Generic hash function (we are using to multiply by 9 and add the byte + * as Tcl) */ +unsigned int Jim_GenHashFunction(const unsigned char *buf, int len) +{ + unsigned int h = 0; + while (len--) + h += (h << 3)+*buf++; + return h; +} + +/* ----------------------------- API implementation ------------------------- */ +/* reset an hashtable already initialized with ht_init(). + * NOTE: This function should only called by ht_destroy(). */ +static void JimResetHashTable(Jim_HashTable *ht) +{ + ht->table = NULL; + ht->size = 0; + ht->sizemask = 0; + ht->used = 0; + ht->collisions = 0; +} + +/* Initialize the hash table */ +int Jim_InitHashTable(Jim_HashTable *ht, Jim_HashTableType *type, + void *privDataPtr) +{ + JimResetHashTable(ht); + ht->type = type; + ht->privdata = privDataPtr; + return JIM_OK; +} + +/* Resize the table to the minimal size that contains all the elements, + * but with the invariant of a USER/BUCKETS ration near to <= 1 */ +int Jim_ResizeHashTable(Jim_HashTable *ht) +{ + int minimal = ht->used; + + if (minimal < JIM_HT_INITIAL_SIZE) + minimal = JIM_HT_INITIAL_SIZE; + return Jim_ExpandHashTable(ht, minimal); +} + +/* Expand or create the hashtable */ +int Jim_ExpandHashTable(Jim_HashTable *ht, unsigned int size) +{ + Jim_HashTable n; /* the new hashtable */ + unsigned int realsize = JimHashTableNextPower(size), i; + + /* the size is invalid if it is smaller than the number of + * elements already inside the hashtable */ + if (ht->used >= size) + return JIM_ERR; + + Jim_InitHashTable(&n, ht->type, ht->privdata); + n.size = realsize; + n.sizemask = realsize-1; + n.table = Jim_Alloc(realsize*sizeof(Jim_HashEntry*)); + + /* Initialize all the pointers to NULL */ + memset(n.table, 0, realsize*sizeof(Jim_HashEntry*)); + + /* Copy all the elements from the old to the new table: + * note that if the old hash table is empty ht->size is zero, + * so Jim_ExpandHashTable just creates an hash table. */ + n.used = ht->used; + for (i = 0; i < ht->size && ht->used > 0; i++) { + Jim_HashEntry *he, *nextHe; + + if (ht->table[i] == NULL) continue; + + /* For each hash entry on this slot... */ + he = ht->table[i]; + while (he) { + unsigned int h; + + nextHe = he->next; + /* Get the new element index */ + h = Jim_HashKey(ht, he->key) & n.sizemask; + he->next = n.table[h]; + n.table[h] = he; + ht->used--; + /* Pass to the next element */ + he = nextHe; + } + } + assert(ht->used == 0); + Jim_Free(ht->table); + + /* Remap the new hashtable in the old */ + *ht = n; + return JIM_OK; +} + +/* Add an element to the target hash table */ +int Jim_AddHashEntry(Jim_HashTable *ht, const void *key, void *val) +{ + int index_t; + Jim_HashEntry *entry; + + /* Get the index of the new element, or -1 if + * the element already exists. */ + if ((index_t = JimInsertHashEntry(ht, key)) == -1) + return JIM_ERR; + + /* Allocates the memory and stores key */ + entry = Jim_Alloc(sizeof(*entry)); + entry->next = ht->table[index_t]; + ht->table[index_t] = entry; + + /* Set the hash entry fields. */ + Jim_SetHashKey(ht, entry, key); + Jim_SetHashVal(ht, entry, val); + ht->used++; + return JIM_OK; +} + +/* Add an element, discarding the old if the key already exists */ +int Jim_ReplaceHashEntry(Jim_HashTable *ht, const void *key, void *val) +{ + Jim_HashEntry *entry; + + /* Try to add the element. If the key + * does not exists Jim_AddHashEntry will suceed. */ + if (Jim_AddHashEntry(ht, key, val) == JIM_OK) + return JIM_OK; + /* It already exists, get the entry */ + entry = Jim_FindHashEntry(ht, key); + /* Free the old value and set the new one */ + Jim_FreeEntryVal(ht, entry); + Jim_SetHashVal(ht, entry, val); + return JIM_OK; +} + +/* Search and remove an element */ +int Jim_DeleteHashEntry(Jim_HashTable *ht, const void *key) +{ + unsigned int h; + Jim_HashEntry *he, *prevHe; + + if (ht->size == 0) + return JIM_ERR; + h = Jim_HashKey(ht, key) & ht->sizemask; + he = ht->table[h]; + + prevHe = NULL; + while (he) { + if (Jim_CompareHashKeys(ht, key, he->key)) { + /* Unlink the element from the list */ + if (prevHe) + prevHe->next = he->next; + else + ht->table[h] = he->next; + Jim_FreeEntryKey(ht, he); + Jim_FreeEntryVal(ht, he); + Jim_Free(he); + ht->used--; + return JIM_OK; + } + prevHe = he; + he = he->next; + } + return JIM_ERR; /* not found */ +} + +/* Destroy an entire hash table */ +int Jim_FreeHashTable(Jim_HashTable *ht) +{ + unsigned int i; + + /* Free all the elements */ + for (i = 0; i < ht->size && ht->used > 0; i++) { + Jim_HashEntry *he, *nextHe; + + if ((he = ht->table[i]) == NULL) continue; + while (he) { + nextHe = he->next; + Jim_FreeEntryKey(ht, he); + Jim_FreeEntryVal(ht, he); + Jim_Free(he); + ht->used--; + he = nextHe; + } + } + /* Free the table and the allocated cache structure */ + Jim_Free(ht->table); + /* Re-initialize the table */ + JimResetHashTable(ht); + return JIM_OK; /* never fails */ +} + +Jim_HashEntry *Jim_FindHashEntry(Jim_HashTable *ht, const void *key) +{ + Jim_HashEntry *he; + unsigned int h; + + if (ht->size == 0) return NULL; + h = Jim_HashKey(ht, key) & ht->sizemask; + he = ht->table[h]; + while (he) { + if (Jim_CompareHashKeys(ht, key, he->key)) + return he; + he = he->next; + } + return NULL; +} + +Jim_HashTableIterator *Jim_GetHashTableIterator(Jim_HashTable *ht) +{ + Jim_HashTableIterator *iter = Jim_Alloc(sizeof(*iter)); + + iter->ht = ht; + iter->index = -1; + iter->entry = NULL; + iter->nextEntry = NULL; + return iter; +} + +Jim_HashEntry *Jim_NextHashEntry(Jim_HashTableIterator *iter) +{ + while (1) { + if (iter->entry == NULL) { + iter->index++; + if (iter->index >= + (signed)iter->ht->size) break; + iter->entry = iter->ht->table[iter->index]; + } else { + iter->entry = iter->nextEntry; + } + if (iter->entry) { + /* We need to save the 'next' here, the iterator user + * may delete the entry we are returning. */ + iter->nextEntry = iter->entry->next; + return iter->entry; + } + } + return NULL; +} + +/* ------------------------- private functions ------------------------------ */ + +/* Expand the hash table if needed */ +static int JimExpandHashTableIfNeeded(Jim_HashTable *ht) +{ + /* If the hash table is empty expand it to the intial size, + * if the table is "full" dobule its size. */ + if (ht->size == 0) + return Jim_ExpandHashTable(ht, JIM_HT_INITIAL_SIZE); + if (ht->size == ht->used) + return Jim_ExpandHashTable(ht, ht->size*2); + return JIM_OK; +} + +/* Our hash table capability is a power of two */ +static unsigned int JimHashTableNextPower(unsigned int size) +{ + unsigned int i = JIM_HT_INITIAL_SIZE; + + if (size >= 2147483648U) + return 2147483648U; + while (1) { + if (i >= size) + return i; + i *= 2; + } +} + +/* Returns the index of a free slot that can be populated with + * an hash entry for the given 'key'. + * If the key already exists, -1 is returned. */ +static int JimInsertHashEntry(Jim_HashTable *ht, const void *key) +{ + unsigned int h; + Jim_HashEntry *he; + + /* Expand the hashtable if needed */ + if (JimExpandHashTableIfNeeded(ht) == JIM_ERR) + return -1; + /* Compute the key hash value */ + h = Jim_HashKey(ht, key) & ht->sizemask; + /* Search if this slot does not already contain the given key */ + he = ht->table[h]; + while (he) { + if (Jim_CompareHashKeys(ht, key, he->key)) + return -1; + he = he->next; + } + return h; +} + +/* ----------------------- StringCopy Hash Table Type ------------------------*/ + +static unsigned int JimStringCopyHTHashFunction(const void *key) +{ + return Jim_GenHashFunction(key, strlen(key)); +} + +static const void *JimStringCopyHTKeyDup(void *privdata, const void *key) +{ + int len = strlen(key); + char *copy = Jim_Alloc(len + 1); + JIM_NOTUSED(privdata); + + memcpy(copy, key, len); + copy[len] = '\0'; + return copy; +} + +static void *JimStringKeyValCopyHTValDup(void *privdata, const void *val) +{ + int len = strlen(val); + char *copy = Jim_Alloc(len + 1); + JIM_NOTUSED(privdata); + + memcpy(copy, val, len); + copy[len] = '\0'; + return copy; +} + +static int JimStringCopyHTKeyCompare(void *privdata, const void *key1, + const void *key2) +{ + JIM_NOTUSED(privdata); + + return strcmp(key1, key2) == 0; +} + +static void JimStringCopyHTKeyDestructor(void *privdata, const void *key) +{ + JIM_NOTUSED(privdata); + + Jim_Free((void*)key); /* ATTENTION: const cast */ +} + +static void JimStringKeyValCopyHTValDestructor(void *privdata, void *val) +{ + JIM_NOTUSED(privdata); + + Jim_Free((void*)val); /* ATTENTION: const cast */ +} + +static Jim_HashTableType JimStringCopyHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + JimStringCopyHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + NULL /* val destructor */ +}; + +/* This is like StringCopy but does not auto-duplicate the key. + * It's used for intepreter's shared strings. */ +static Jim_HashTableType JimSharedStringsHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + NULL, /* key dup */ + NULL, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + NULL /* val destructor */ +}; + +/* This is like StringCopy but also automatically handle dynamic + * allocated C strings as values. */ +static Jim_HashTableType JimStringKeyValCopyHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + JimStringCopyHTKeyDup, /* key dup */ + JimStringKeyValCopyHTValDup, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + JimStringKeyValCopyHTValDestructor, /* val destructor */ +}; + +typedef struct AssocDataValue { + Jim_InterpDeleteProc *delProc; + void *data; +} AssocDataValue; + +static void JimAssocDataHashTableValueDestructor(void *privdata, void *data) +{ + AssocDataValue *assocPtr = (AssocDataValue *)data; + if (assocPtr->delProc != NULL) + assocPtr->delProc((Jim_Interp *)privdata, assocPtr->data); + Jim_Free(data); +} + +static Jim_HashTableType JimAssocDataHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + JimStringCopyHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + JimAssocDataHashTableValueDestructor /* val destructor */ +}; + +/* ----------------------------------------------------------------------------- + * Stack - This is a simple generic stack implementation. It is used for + * example in the 'expr' expression compiler. + * ---------------------------------------------------------------------------*/ +void Jim_InitStack(Jim_Stack *stack) +{ + stack->len = 0; + stack->maxlen = 0; + stack->vector = NULL; +} + +void Jim_FreeStack(Jim_Stack *stack) +{ + Jim_Free(stack->vector); +} + +int Jim_StackLen(Jim_Stack *stack) +{ + return stack->len; +} + +void Jim_StackPush(Jim_Stack *stack, void *element) { + int neededLen = stack->len + 1; + if (neededLen > stack->maxlen) { + stack->maxlen = neededLen*2; + stack->vector = Jim_Realloc(stack->vector, sizeof(void*)*stack->maxlen); + } + stack->vector[stack->len] = element; + stack->len++; +} + +void *Jim_StackPop(Jim_Stack *stack) +{ + if (stack->len == 0) return NULL; + stack->len--; + return stack->vector[stack->len]; +} + +void *Jim_StackPeek(Jim_Stack *stack) +{ + if (stack->len == 0) return NULL; + return stack->vector[stack->len-1]; +} + +void Jim_FreeStackElements(Jim_Stack *stack, void (*freeFunc)(void *ptr)) +{ + int i; + + for (i = 0; i < stack->len; i++) + freeFunc(stack->vector[i]); +} + +/* ----------------------------------------------------------------------------- + * Parser + * ---------------------------------------------------------------------------*/ + +/* Token types */ +#define JIM_TT_NONE -1 /* No token returned */ +#define JIM_TT_STR 0 /* simple string */ +#define JIM_TT_ESC 1 /* string that needs escape chars conversion */ +#define JIM_TT_VAR 2 /* var substitution */ +#define JIM_TT_DICTSUGAR 3 /* Syntax sugar for [dict get], $foo(bar) */ +#define JIM_TT_CMD 4 /* command substitution */ +#define JIM_TT_SEP 5 /* word separator */ +#define JIM_TT_EOL 6 /* line separator */ + +/* Additional token types needed for expressions */ +#define JIM_TT_SUBEXPR_START 7 +#define JIM_TT_SUBEXPR_END 8 +#define JIM_TT_EXPR_NUMBER 9 +#define JIM_TT_EXPR_OPERATOR 10 + +/* Parser states */ +#define JIM_PS_DEF 0 /* Default state */ +#define JIM_PS_QUOTE 1 /* Inside "" */ + +/* Parser context structure. The same context is used both to parse + * Tcl scripts and lists. */ +struct JimParserCtx { + const char *prg; /* Program text */ + const char *p; /* Pointer to the point of the program we are parsing */ + int len; /* Left length of 'prg' */ + int linenr; /* Current line number */ + const char *tstart; + const char *tend; /* Returned token is at tstart-tend in 'prg'. */ + int tline; /* Line number of the returned token */ + int tt; /* Token type */ + int eof; /* Non zero if EOF condition is true. */ + int state; /* Parser state */ + int comment; /* Non zero if the next chars may be a comment. */ +}; + +#define JimParserEof(c) ((c)->eof) +#define JimParserTstart(c) ((c)->tstart) +#define JimParserTend(c) ((c)->tend) +#define JimParserTtype(c) ((c)->tt) +#define JimParserTline(c) ((c)->tline) + +static int JimParseScript(struct JimParserCtx *pc); +static int JimParseSep(struct JimParserCtx *pc); +static int JimParseEol(struct JimParserCtx *pc); +static int JimParseCmd(struct JimParserCtx *pc); +static int JimParseVar(struct JimParserCtx *pc); +static int JimParseBrace(struct JimParserCtx *pc); +static int JimParseStr(struct JimParserCtx *pc); +static int JimParseComment(struct JimParserCtx *pc); +static char *JimParserGetToken(struct JimParserCtx *pc, + int *lenPtr, int *typePtr, int *linePtr); + +/* Initialize a parser context. + * 'prg' is a pointer to the program text, linenr is the line + * number of the first line contained in the program. */ +void JimParserInit(struct JimParserCtx *pc, const char *prg, + int len, int linenr) +{ + pc->prg = prg; + pc->p = prg; + pc->len = len; + pc->tstart = NULL; + pc->tend = NULL; + pc->tline = 0; + pc->tt = JIM_TT_NONE; + pc->eof = 0; + pc->state = JIM_PS_DEF; + pc->linenr = linenr; + pc->comment = 1; +} + +int JimParseScript(struct JimParserCtx *pc) +{ + while (1) { /* the while is used to reiterate with continue if needed */ + if (!pc->len) { + pc->tstart = pc->p; + pc->tend = pc->p-1; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EOL; + pc->eof = 1; + return JIM_OK; + } + switch (*(pc->p)) { + case '\\': + if (*(pc->p + 1) == '\n') + return JimParseSep(pc); + else { + pc->comment = 0; + return JimParseStr(pc); + } + break; + case ' ': + case '\t': + case '\r': + if (pc->state == JIM_PS_DEF) + return JimParseSep(pc); + else { + pc->comment = 0; + return JimParseStr(pc); + } + break; + case '\n': + case ';': + pc->comment = 1; + if (pc->state == JIM_PS_DEF) + return JimParseEol(pc); + else + return JimParseStr(pc); + break; + case '[': + pc->comment = 0; + return JimParseCmd(pc); + break; + case '$': + pc->comment = 0; + if (JimParseVar(pc) == JIM_ERR) { + pc->tstart = pc->tend = pc->p++; pc->len--; + pc->tline = pc->linenr; + pc->tt = JIM_TT_STR; + return JIM_OK; + } else + return JIM_OK; + break; + case '#': + if (pc->comment) { + JimParseComment(pc); + continue; + } else { + return JimParseStr(pc); + } + default: + pc->comment = 0; + return JimParseStr(pc); + break; + } + return JIM_OK; + } +} + +int JimParseSep(struct JimParserCtx *pc) +{ + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (*pc->p == ' ' || *pc->p == '\t' || *pc->p == '\r' || + (*pc->p == '\\' && *(pc->p + 1) == '\n')) { + if (*pc->p == '\\') { + pc->p++; pc->len--; + pc->linenr++; + } + pc->p++; pc->len--; + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_SEP; + return JIM_OK; +} + +int JimParseEol(struct JimParserCtx *pc) +{ + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (*pc->p == ' ' || *pc->p == '\n' || + *pc->p == '\t' || *pc->p == '\r' || *pc->p == ';') { + if (*pc->p == '\n') + pc->linenr++; + pc->p++; pc->len--; + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_EOL; + return JIM_OK; +} + +/* Todo. Don't stop if ']' appears inside {} or quoted. + * Also should handle the case of puts [string length "]"] */ +int JimParseCmd(struct JimParserCtx *pc) +{ + int level = 1; + int blevel = 0; + + pc->tstart = ++pc->p; pc->len--; + pc->tline = pc->linenr; + while (1) { + if (pc->len == 0) { + break; + } else if (*pc->p == '[' && blevel == 0) { + level++; + } else if (*pc->p == ']' && blevel == 0) { + level--; + if (!level) break; + } else if (*pc->p == '\\') { + pc->p++; pc->len--; + } else if (*pc->p == '{') { + blevel++; + } else if (*pc->p == '}') { + if (blevel != 0) + blevel--; + } else if (*pc->p == '\n') + pc->linenr++; + pc->p++; pc->len--; + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_CMD; + if (*pc->p == ']') { + pc->p++; pc->len--; + } + return JIM_OK; +} + +int JimParseVar(struct JimParserCtx *pc) +{ + int brace = 0, stop = 0, ttype = JIM_TT_VAR; + + pc->tstart = ++pc->p; pc->len--; /* skip the $ */ + pc->tline = pc->linenr; + if (*pc->p == '{') { + pc->tstart = ++pc->p; pc->len--; + brace = 1; + } + if (brace) { + while (!stop) { + if (*pc->p == '}' || pc->len == 0) { + pc->tend = pc->p-1; + stop = 1; + if (pc->len == 0) + break; + } + else if (*pc->p == '\n') + pc->linenr++; + pc->p++; pc->len--; + } + } else { + /* Include leading colons */ + while (*pc->p == ':') { + pc->p++; + pc->len--; + } + while (!stop) { + if (!((*pc->p >= 'a' && *pc->p <= 'z') || + (*pc->p >= 'A' && *pc->p <= 'Z') || + (*pc->p >= '0' && *pc->p <= '9') || *pc->p == '_')) + stop = 1; + else { + pc->p++; pc->len--; + } + } + /* Parse [dict get] syntax sugar. */ + if (*pc->p == '(') { + while (*pc->p != ')' && pc->len) { + pc->p++; pc->len--; + if (*pc->p == '\\' && pc->len >= 2) { + pc->p += 2; pc->len -= 2; + } + } + if (*pc->p != '\0') { + pc->p++; pc->len--; + } + ttype = JIM_TT_DICTSUGAR; + } + pc->tend = pc->p-1; + } + /* Check if we parsed just the '$' character. + * That's not a variable so an error is returned + * to tell the state machine to consider this '$' just + * a string. */ + if (pc->tstart == pc->p) { + pc->p--; pc->len++; + return JIM_ERR; + } + pc->tt = ttype; + return JIM_OK; +} + +int JimParseBrace(struct JimParserCtx *pc) +{ + int level = 1; + + pc->tstart = ++pc->p; pc->len--; + pc->tline = pc->linenr; + while (1) { + if (*pc->p == '\\' && pc->len >= 2) { + pc->p++; pc->len--; + if (*pc->p == '\n') + pc->linenr++; + } else if (*pc->p == '{') { + level++; + } else if (pc->len == 0 || *pc->p == '}') { + level--; + if (pc->len == 0 || level == 0) { + pc->tend = pc->p-1; + if (pc->len != 0) { + pc->p++; pc->len--; + } + pc->tt = JIM_TT_STR; + return JIM_OK; + } + } else if (*pc->p == '\n') { + pc->linenr++; + } + pc->p++; pc->len--; + } + return JIM_OK; /* unreached */ +} + +int JimParseStr(struct JimParserCtx *pc) +{ + int newword = (pc->tt == JIM_TT_SEP || pc->tt == JIM_TT_EOL || + pc->tt == JIM_TT_NONE || pc->tt == JIM_TT_STR); + if (newword && *pc->p == '{') { + return JimParseBrace(pc); + } else if (newword && *pc->p == '"') { + pc->state = JIM_PS_QUOTE; + pc->p++; pc->len--; + } + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (1) { + if (pc->len == 0) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + } + switch (*pc->p) { + case '\\': + if (pc->state == JIM_PS_DEF && + *(pc->p + 1) == '\n') { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + } + if (pc->len >= 2) { + pc->p++; pc->len--; + } + break; + case '$': + case '[': + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + case ' ': + case '\t': + case '\n': + case '\r': + case ';': + if (pc->state == JIM_PS_DEF) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + } else if (*pc->p == '\n') { + pc->linenr++; + } + break; + case '"': + if (pc->state == JIM_PS_QUOTE) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + pc->p++; pc->len--; + pc->state = JIM_PS_DEF; + return JIM_OK; + } + break; + } + pc->p++; pc->len--; + } + return JIM_OK; /* unreached */ +} + +int JimParseComment(struct JimParserCtx *pc) +{ + while (*pc->p) { + if (*pc->p == '\n') { + pc->linenr++; + if (*(pc->p-1) != '\\') { + pc->p++; pc->len--; + return JIM_OK; + } + } + pc->p++; pc->len--; + } + return JIM_OK; +} + +/* xdigitval and odigitval are helper functions for JimParserGetToken() */ +static int xdigitval(int c) +{ + if (c >= '0' && c <= '9') return c-'0'; + if (c >= 'a' && c <= 'f') return c-'a'+10; + if (c >= 'A' && c <= 'F') return c-'A'+10; + return -1; +} + +static int odigitval(int c) +{ + if (c >= '0' && c <= '7') return c-'0'; + return -1; +} + +/* Perform Tcl escape substitution of 's', storing the result + * string into 'dest'. The escaped string is guaranteed to + * be the same length or shorted than the source string. + * Slen is the length of the string at 's', if it's -1 the string + * length will be calculated by the function. + * + * The function returns the length of the resulting string. */ +static int JimEscape(char *dest, const char *s, int slen) +{ + char *p = dest; + int i, len; + + if (slen == -1) + slen = strlen(s); + + for (i = 0; i < slen; i++) { + switch (s[i]) { + case '\\': + switch (s[i + 1]) { + case 'a': *p++ = 0x7; i++; break; + case 'b': *p++ = 0x8; i++; break; + case 'f': *p++ = 0xc; i++; break; + case 'n': *p++ = 0xa; i++; break; + case 'r': *p++ = 0xd; i++; break; + case 't': *p++ = 0x9; i++; break; + case 'v': *p++ = 0xb; i++; break; + case '\0': *p++ = '\\'; i++; break; + case '\n': *p++ = ' '; i++; break; + default: + if (s[i + 1] == 'x') { + int val = 0; + int c = xdigitval(s[i + 2]); + if (c == -1) { + *p++ = 'x'; + i++; + break; + } + val = c; + c = xdigitval(s[i + 3]); + if (c == -1) { + *p++ = val; + i += 2; + break; + } + val = (val*16) + c; + *p++ = val; + i += 3; + break; + } else if (s[i + 1] >= '0' && s[i + 1] <= '7') + { + int val = 0; + int c = odigitval(s[i + 1]); + val = c; + c = odigitval(s[i + 2]); + if (c == -1) { + *p++ = val; + i ++; + break; + } + val = (val*8) + c; + c = odigitval(s[i + 3]); + if (c == -1) { + *p++ = val; + i += 2; + break; + } + val = (val*8) + c; + *p++ = val; + i += 3; + } else { + *p++ = s[i + 1]; + i++; + } + break; + } + break; + default: + *p++ = s[i]; + break; + } + } + len = p-dest; + *p++ = '\0'; + return len; +} + +/* Returns a dynamically allocated copy of the current token in the + * parser context. The function perform conversion of escapes if + * the token is of type JIM_TT_ESC. + * + * Note that after the conversion, tokens that are grouped with + * braces in the source code, are always recognizable from the + * identical string obtained in a different way from the type. + * + * For exmple the string: + * + * {expand}$a + * + * will return as first token "expand", of type JIM_TT_STR + * + * While the string: + * + * expand$a + * + * will return as first token "expand", of type JIM_TT_ESC + */ +char *JimParserGetToken(struct JimParserCtx *pc, + int *lenPtr, int *typePtr, int *linePtr) +{ + const char *start, *end; + char *token; + int len; + + start = JimParserTstart(pc); + end = JimParserTend(pc); + if (start > end) { + if (lenPtr) *lenPtr = 0; + if (typePtr) *typePtr = JimParserTtype(pc); + if (linePtr) *linePtr = JimParserTline(pc); + token = Jim_Alloc(1); + token[0] = '\0'; + return token; + } + len = (end-start) + 1; + token = Jim_Alloc(len + 1); + if (JimParserTtype(pc) != JIM_TT_ESC) { + /* No escape conversion needed? Just copy it. */ + memcpy(token, start, len); + token[len] = '\0'; + } else { + /* Else convert the escape chars. */ + len = JimEscape(token, start, len); + } + if (lenPtr) *lenPtr = len; + if (typePtr) *typePtr = JimParserTtype(pc); + if (linePtr) *linePtr = JimParserTline(pc); + return token; +} + +/* The following functin is not really part of the parsing engine of Jim, + * but it somewhat related. Given an string and its length, it tries + * to guess if the script is complete or there are instead " " or { } + * open and not completed. This is useful for interactive shells + * implementation and for [info complete]. + * + * If 'stateCharPtr' != NULL, the function stores ' ' on complete script, + * '{' on scripts incomplete missing one or more '}' to be balanced. + * '"' on scripts incomplete missing a '"' char. + * + * If the script is complete, 1 is returned, otherwise 0. */ +int Jim_ScriptIsComplete(const char *s, int len, char *stateCharPtr) +{ + int level = 0; + int state = ' '; + + while (len) { + switch (*s) { + case '\\': + if (len > 1) + s++; + break; + case '"': + if (state == ' ') { + state = '"'; + } else if (state == '"') { + state = ' '; + } + break; + case '{': + if (state == '{') { + level++; + } else if (state == ' ') { + state = '{'; + level++; + } + break; + case '}': + if (state == '{') { + level--; + if (level == 0) + state = ' '; + } + break; + } + s++; + len--; + } + if (stateCharPtr) + *stateCharPtr = state; + return state == ' '; +} + +/* ----------------------------------------------------------------------------- + * Tcl Lists parsing + * ---------------------------------------------------------------------------*/ +static int JimParseListSep(struct JimParserCtx *pc); +static int JimParseListStr(struct JimParserCtx *pc); + +int JimParseList(struct JimParserCtx *pc) +{ + if (pc->len == 0) { + pc->tstart = pc->tend = pc->p; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EOL; + pc->eof = 1; + return JIM_OK; + } + switch (*pc->p) { + case ' ': + case '\n': + case '\t': + case '\r': + if (pc->state == JIM_PS_DEF) + return JimParseListSep(pc); + else + return JimParseListStr(pc); + break; + default: + return JimParseListStr(pc); + break; + } + return JIM_OK; +} + +int JimParseListSep(struct JimParserCtx *pc) +{ + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (*pc->p == ' ' || *pc->p == '\t' || *pc->p == '\r' || *pc->p == '\n') + { + pc->p++; pc->len--; + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_SEP; + return JIM_OK; +} + +int JimParseListStr(struct JimParserCtx *pc) +{ + int newword = (pc->tt == JIM_TT_SEP || pc->tt == JIM_TT_EOL || + pc->tt == JIM_TT_NONE); + if (newword && *pc->p == '{') { + return JimParseBrace(pc); + } else if (newword && *pc->p == '"') { + pc->state = JIM_PS_QUOTE; + pc->p++; pc->len--; + } + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (1) { + if (pc->len == 0) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + } + switch (*pc->p) { + case '\\': + pc->p++; pc->len--; + break; + case ' ': + case '\t': + case '\n': + case '\r': + if (pc->state == JIM_PS_DEF) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; + } else if (*pc->p == '\n') { + pc->linenr++; + } + break; + case '"': + if (pc->state == JIM_PS_QUOTE) { + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + pc->p++; pc->len--; + pc->state = JIM_PS_DEF; + return JIM_OK; + } + break; + } + pc->p++; pc->len--; + } + return JIM_OK; /* unreached */ +} + +/* ----------------------------------------------------------------------------- + * Jim_Obj related functions + * ---------------------------------------------------------------------------*/ + +/* Return a new initialized object. */ +Jim_Obj *Jim_NewObj(Jim_Interp *interp) +{ + Jim_Obj *objPtr; + + /* -- Check if there are objects in the free list -- */ + if (interp->freeList != NULL) { + /* -- Unlink the object from the free list -- */ + objPtr = interp->freeList; + interp->freeList = objPtr->nextObjPtr; + } else { + /* -- No ready to use objects: allocate a new one -- */ + objPtr = Jim_Alloc(sizeof(*objPtr)); + } + + /* Object is returned with refCount of 0. Every + * kind of GC implemented should take care to don't try + * to scan objects with refCount == 0. */ + objPtr->refCount = 0; + /* All the other fields are left not initialized to save time. + * The caller will probably want set they to the right + * value anyway. */ + + /* -- Put the object into the live list -- */ + objPtr->prevObjPtr = NULL; + objPtr->nextObjPtr = interp->liveList; + if (interp->liveList) + interp->liveList->prevObjPtr = objPtr; + interp->liveList = objPtr; + + return objPtr; +} + +/* Free an object. Actually objects are never freed, but + * just moved to the free objects list, where they will be + * reused by Jim_NewObj(). */ +void Jim_FreeObj(Jim_Interp *interp, Jim_Obj *objPtr) +{ + /* Check if the object was already freed, panic. */ + if (objPtr->refCount != 0) { + Jim_Panic(interp,"!!!Object %p freed with bad refcount %d", objPtr, + objPtr->refCount); + } + /* Free the internal representation */ + Jim_FreeIntRep(interp, objPtr); + /* Free the string representation */ + if (objPtr->bytes != NULL) { + if (objPtr->bytes != JimEmptyStringRep) + Jim_Free(objPtr->bytes); + } + /* Unlink the object from the live objects list */ + if (objPtr->prevObjPtr) + objPtr->prevObjPtr->nextObjPtr = objPtr->nextObjPtr; + if (objPtr->nextObjPtr) + objPtr->nextObjPtr->prevObjPtr = objPtr->prevObjPtr; + if (interp->liveList == objPtr) + interp->liveList = objPtr->nextObjPtr; + /* Link the object into the free objects list */ + objPtr->prevObjPtr = NULL; + objPtr->nextObjPtr = interp->freeList; + if (interp->freeList) + interp->freeList->prevObjPtr = objPtr; + interp->freeList = objPtr; + objPtr->refCount = -1; +} + +/* Invalidate the string representation of an object. */ +void Jim_InvalidateStringRep(Jim_Obj *objPtr) +{ + if (objPtr->bytes != NULL) { + if (objPtr->bytes != JimEmptyStringRep) + Jim_Free(objPtr->bytes); + } + objPtr->bytes = NULL; +} + +#define Jim_SetStringRep(o, b, l) \ + do { (o)->bytes = b; (o)->length = l; } while (0) + +/* Set the initial string representation for an object. + * Does not try to free an old one. */ +void Jim_InitStringRep(Jim_Obj *objPtr, const char *bytes, int length) +{ + if (length == 0) { + objPtr->bytes = JimEmptyStringRep; + objPtr->length = 0; + } else { + objPtr->bytes = Jim_Alloc(length + 1); + objPtr->length = length; + memcpy(objPtr->bytes, bytes, length); + objPtr->bytes[length] = '\0'; + } +} + +/* Duplicate an object. The returned object has refcount = 0. */ +Jim_Obj *Jim_DuplicateObj(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_Obj *dupPtr; + + dupPtr = Jim_NewObj(interp); + if (objPtr->bytes == NULL) { + /* Object does not have a valid string representation. */ + dupPtr->bytes = NULL; + } else { + Jim_InitStringRep(dupPtr, objPtr->bytes, objPtr->length); + } + if (objPtr->typePtr != NULL) { + if (objPtr->typePtr->dupIntRepProc == NULL) { + dupPtr->internalRep = objPtr->internalRep; + } else { + objPtr->typePtr->dupIntRepProc(interp, objPtr, dupPtr); + } + dupPtr->typePtr = objPtr->typePtr; + } else { + dupPtr->typePtr = NULL; + } + return dupPtr; +} + +/* Return the string representation for objPtr. If the object + * string representation is invalid, calls the method to create + * a new one starting from the internal representation of the object. */ +const char *Jim_GetString(Jim_Obj *objPtr, int *lenPtr) +{ + if (objPtr->bytes == NULL) { + /* Invalid string repr. Generate it. */ + if (objPtr->typePtr->updateStringProc == NULL) { + Jim_Panic(NULL,"UpdataStringProc called against '%s' type.", + objPtr->typePtr->name); + } + objPtr->typePtr->updateStringProc(objPtr); + } + if (lenPtr) + *lenPtr = objPtr->length; + return objPtr->bytes; +} + +/* Just returns the length of the object's string rep */ +int Jim_Length(Jim_Obj *objPtr) +{ + int len; + + Jim_GetString(objPtr, &len); + return len; +} + +/* ----------------------------------------------------------------------------- + * String Object + * ---------------------------------------------------------------------------*/ +static void DupStringInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static int SetStringFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType stringObjType = { + "string", + NULL, + DupStringInternalRep, + NULL, + JIM_TYPE_REFERENCES, +}; + +void DupStringInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + JIM_NOTUSED(interp); + + /* This is a bit subtle: the only caller of this function + * should be Jim_DuplicateObj(), that will copy the + * string representaion. After the copy, the duplicated + * object will not have more room in teh buffer than + * srcPtr->length bytes. So we just set it to length. */ + dupPtr->internalRep.strValue.maxLength = srcPtr->length; +} + +int SetStringFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + /* Get a fresh string representation. */ + (void) Jim_GetString(objPtr, NULL); + /* Free any other internal representation. */ + Jim_FreeIntRep(interp, objPtr); + /* Set it as string, i.e. just set the maxLength field. */ + objPtr->typePtr = &stringObjType; + objPtr->internalRep.strValue.maxLength = objPtr->length; + return JIM_OK; +} + +Jim_Obj *Jim_NewStringObj(Jim_Interp *interp, const char *s, int len) +{ + Jim_Obj *objPtr = Jim_NewObj(interp); + + if (len == -1) + len = strlen(s); + /* Alloc/Set the string rep. */ + if (len == 0) { + objPtr->bytes = JimEmptyStringRep; + objPtr->length = 0; + } else { + objPtr->bytes = Jim_Alloc(len + 1); + objPtr->length = len; + memcpy(objPtr->bytes, s, len); + objPtr->bytes[len] = '\0'; + } + + /* No typePtr field for the vanilla string object. */ + objPtr->typePtr = NULL; + return objPtr; +} + +/* This version does not try to duplicate the 's' pointer, but + * use it directly. */ +Jim_Obj *Jim_NewStringObjNoAlloc(Jim_Interp *interp, char *s, int len) +{ + Jim_Obj *objPtr = Jim_NewObj(interp); + + if (len == -1) + len = strlen(s); + Jim_SetStringRep(objPtr, s, len); + objPtr->typePtr = NULL; + return objPtr; +} + +/* Low-level string append. Use it only against objects + * of type "string". */ +void StringAppendString(Jim_Obj *objPtr, const char *str, int len) +{ + int needlen; + + if (len == -1) + len = strlen(str); + needlen = objPtr->length + len; + if (objPtr->internalRep.strValue.maxLength < needlen || + objPtr->internalRep.strValue.maxLength == 0) { + if (objPtr->bytes == JimEmptyStringRep) { + objPtr->bytes = Jim_Alloc((needlen*2) + 1); + } else { + objPtr->bytes = Jim_Realloc(objPtr->bytes, (needlen*2) + 1); + } + objPtr->internalRep.strValue.maxLength = needlen*2; + } + memcpy(objPtr->bytes + objPtr->length, str, len); + objPtr->bytes[objPtr->length + len] = '\0'; + objPtr->length += len; +} + +/* Low-level wrapper to append an object. */ +void StringAppendObj(Jim_Obj *objPtr, Jim_Obj *appendObjPtr) +{ + int len; + const char *str; + + str = Jim_GetString(appendObjPtr, &len); + StringAppendString(objPtr, str, len); +} + +/* Higher level API to append strings to objects. */ +void Jim_AppendString(Jim_Interp *interp, Jim_Obj *objPtr, const char *str, + int len) +{ + if (Jim_IsShared(objPtr)) + Jim_Panic(interp,"Jim_AppendString called with shared object"); + if (objPtr->typePtr != &stringObjType) + SetStringFromAny(interp, objPtr); + StringAppendString(objPtr, str, len); +} + +void Jim_AppendString_sprintf(Jim_Interp *interp, Jim_Obj *objPtr, const char *fmt, ...) +{ + char *buf; + va_list ap; + + va_start(ap, fmt); + buf = jim_vasprintf(fmt, ap); + va_end(ap); + + if (buf) { + Jim_AppendString(interp, objPtr, buf, -1); + jim_vasprintf_done(buf); + } +} + + +void Jim_AppendObj(Jim_Interp *interp, Jim_Obj *objPtr, + Jim_Obj *appendObjPtr) +{ + int len; + const char *str; + + str = Jim_GetString(appendObjPtr, &len); + Jim_AppendString(interp, objPtr, str, len); +} + +void Jim_AppendStrings(Jim_Interp *interp, Jim_Obj *objPtr, ...) +{ + va_list ap; + + if (objPtr->typePtr != &stringObjType) + SetStringFromAny(interp, objPtr); + va_start(ap, objPtr); + while (1) { + char *s = va_arg(ap, char*); + + if (s == NULL) break; + Jim_AppendString(interp, objPtr, s, -1); + } + va_end(ap); +} + +int Jim_StringEqObj(Jim_Obj *aObjPtr, Jim_Obj *bObjPtr, int nocase) +{ + const char *aStr, *bStr; + int aLen, bLen, i; + + if (aObjPtr == bObjPtr) return 1; + aStr = Jim_GetString(aObjPtr, &aLen); + bStr = Jim_GetString(bObjPtr, &bLen); + if (aLen != bLen) return 0; + if (nocase == 0) + return memcmp(aStr, bStr, aLen) == 0; + for (i = 0; i < aLen; i++) { + if (tolower((int)aStr[i]) != tolower((int)bStr[i])) + return 0; + } + return 1; +} + +int Jim_StringMatchObj(Jim_Obj *patternObjPtr, Jim_Obj *objPtr, + int nocase) +{ + const char *pattern, *string; + int patternLen, stringLen; + + pattern = Jim_GetString(patternObjPtr, &patternLen); + string = Jim_GetString(objPtr, &stringLen); + return JimStringMatch(pattern, patternLen, string, stringLen, nocase); +} + +int Jim_StringCompareObj(Jim_Obj *firstObjPtr, + Jim_Obj *secondObjPtr, int nocase) +{ + const char *s1, *s2; + int l1, l2; + + s1 = Jim_GetString(firstObjPtr, &l1); + s2 = Jim_GetString(secondObjPtr, &l2); + return JimStringCompare(s1, l1, s2, l2, nocase); +} + +/* Convert a range, as returned by Jim_GetRange(), into + * an absolute index into an object of the specified length. + * This function may return negative values, or values + * bigger or equal to the length of the list if the index + * is out of range. */ +static int JimRelToAbsIndex(int len, int index_t) +{ + if (index_t < 0) + return len + index_t; + return index_t; +} + +/* Convert a pair of index as normalize by JimRelToAbsIndex(), + * into a range stored in *firstPtr, *lastPtr, *rangeLenPtr, suitable + * for implementation of commands like [string range] and [lrange]. + * + * The resulting range is guaranteed to address valid elements of + * the structure. */ +static void JimRelToAbsRange(int len, int first, int last, + int *firstPtr, int *lastPtr, int *rangeLenPtr) +{ + int rangeLen; + + if (first > last) { + rangeLen = 0; + } else { + rangeLen = last-first + 1; + if (rangeLen) { + if (first < 0) { + rangeLen += first; + first = 0; + } + if (last >= len) { + rangeLen -= (last-(len-1)); + last = len-1; + } + } + } + if (rangeLen < 0) rangeLen = 0; + + *firstPtr = first; + *lastPtr = last; + *rangeLenPtr = rangeLen; +} + +Jim_Obj *Jim_StringRangeObj(Jim_Interp *interp, + Jim_Obj *strObjPtr, Jim_Obj *firstObjPtr, Jim_Obj *lastObjPtr) +{ + int first, last; + const char *str; + int len, rangeLen; + + if (Jim_GetIndex(interp, firstObjPtr, &first) != JIM_OK || + Jim_GetIndex(interp, lastObjPtr, &last) != JIM_OK) + return NULL; + str = Jim_GetString(strObjPtr, &len); + first = JimRelToAbsIndex(len, first); + last = JimRelToAbsIndex(len, last); + JimRelToAbsRange(len, first, last, &first, &last, &rangeLen); + return Jim_NewStringObj(interp, str + first, rangeLen); +} + +static Jim_Obj *JimStringToLower(Jim_Interp *interp, Jim_Obj *strObjPtr) +{ + char *buf; + int i; + if (strObjPtr->typePtr != &stringObjType) { + SetStringFromAny(interp, strObjPtr); + } + + buf = Jim_Alloc(strObjPtr->length + 1); + + memcpy(buf, strObjPtr->bytes, strObjPtr->length + 1); + for (i = 0; i < strObjPtr->length; i++) + buf[i] = tolower((unsigned)buf[i]); + return Jim_NewStringObjNoAlloc(interp, buf, strObjPtr->length); +} + +static Jim_Obj *JimStringToUpper(Jim_Interp *interp, Jim_Obj *strObjPtr) +{ + char *buf; + int i; + if (strObjPtr->typePtr != &stringObjType) { + SetStringFromAny(interp, strObjPtr); + } + + buf = Jim_Alloc(strObjPtr->length + 1); + + memcpy(buf, strObjPtr->bytes, strObjPtr->length + 1); + for (i = 0; i < strObjPtr->length; i++) + buf[i] = toupper((unsigned)buf[i]); + return Jim_NewStringObjNoAlloc(interp, buf, strObjPtr->length); +} + +/* This is the core of the [format] command. + * TODO: Lots of things work - via a hack + * However, no format item can be >= JIM_MAX_FMT + */ +#define JIM_MAX_FMT 2048 +static Jim_Obj *Jim_FormatString_Inner(Jim_Interp *interp, Jim_Obj *fmtObjPtr, + int objc, Jim_Obj *const *objv, char *sprintf_buf) +{ + const char *fmt, *_fmt; + int fmtLen; + Jim_Obj *resObjPtr; + + + fmt = Jim_GetString(fmtObjPtr, &fmtLen); + _fmt = fmt; + resObjPtr = Jim_NewStringObj(interp, "", 0); + while (fmtLen) { + const char *p = fmt; + char spec[2], c; + jim_wide wideValue; + double doubleValue; + /* we cheat and use Sprintf()! */ + char fmt_str[100]; + char *cp; + int width; + int ljust; + int zpad; + int spad; + int altfm; + int forceplus; + int prec; + int inprec; + int haveprec; + int accum; + + while (*fmt != '%' && fmtLen) { + fmt++; fmtLen--; + } + Jim_AppendString(interp, resObjPtr, p, fmt-p); + if (fmtLen == 0) + break; + fmt++; fmtLen--; /* skip '%' */ + zpad = 0; + spad = 0; + width = -1; + ljust = 0; + altfm = 0; + forceplus = 0; + inprec = 0; + haveprec = 0; + prec = -1; /* not found yet */ + next_fmt: + if (fmtLen <= 0) { + break; + } + switch (*fmt) { + /* terminals */ + case 'b': /* binary - not all printfs() do this */ + case 's': /* string */ + case 'i': /* integer */ + case 'd': /* decimal */ + case 'x': /* hex */ + case 'X': /* CAP hex */ + case 'c': /* char */ + case 'o': /* octal */ + case 'u': /* unsigned */ + case 'f': /* float */ + break; + + /* non-terminals */ + case '0': /* zero pad */ + zpad = 1; + fmt++; fmtLen--; + goto next_fmt; + break; + case '+': + forceplus = 1; + fmt++; fmtLen--; + goto next_fmt; + break; + case ' ': /* sign space */ + spad = 1; + fmt++; fmtLen--; + goto next_fmt; + break; + case '-': + ljust = 1; + fmt++; fmtLen--; + goto next_fmt; + break; + case '#': + altfm = 1; + fmt++; fmtLen--; + goto next_fmt; + + case '.': + inprec = 1; + fmt++; fmtLen--; + goto next_fmt; + break; + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + accum = 0; + while (isdigit((unsigned)*fmt) && (fmtLen > 0)) { + accum = (accum * 10) + (*fmt - '0'); + fmt++; fmtLen--; + } + if (inprec) { + haveprec = 1; + prec = accum; + } else { + width = accum; + } + goto next_fmt; + case '*': + /* suck up the next item as an integer */ + fmt++; fmtLen--; + objc--; + if (objc <= 0) { + goto not_enough_args; + } + if (Jim_GetWide(interp,objv[0],&wideValue)== JIM_ERR) { + Jim_FreeNewObj(interp, resObjPtr); + return NULL; + } + if (inprec) { + haveprec = 1; + prec = wideValue; + if (prec < 0) { + /* man 3 printf says */ + /* if prec is negative, it is zero */ + prec = 0; + } + } else { + width = wideValue; + if (width < 0) { + ljust = 1; + width = -width; + } + } + objv++; + goto next_fmt; + break; + } + + + if (*fmt != '%') { + if (objc == 0) { + not_enough_args: + Jim_FreeNewObj(interp, resObjPtr); + Jim_SetResultString(interp, + "not enough arguments for all format specifiers", -1); + return NULL; + } else { + objc--; + } + } + + /* + * Create the formatter + * cause we cheat and use sprintf() + */ + cp = fmt_str; + *cp++ = '%'; + if (altfm) { + *cp++ = '#'; + } + if (forceplus) { + *cp++ = '+'; + } else if (spad) { + /* PLUS overrides */ + *cp++ = ' '; + } + if (ljust) { + *cp++ = '-'; + } + if (zpad) { + *cp++ = '0'; + } + if (width > 0) { + sprintf(cp, "%d", width); + /* skip ahead */ + cp = strchr(cp,0); + } + /* did we find a period? */ + if (inprec) { + /* then add it */ + *cp++ = '.'; + /* did something occur after the period? */ + if (haveprec) { + sprintf(cp, "%d", prec); + } + cp = strchr(cp,0); + } + *cp = 0; + + /* here we do the work */ + /* actually - we make sprintf() do it for us */ + switch (*fmt) { + case 's': + *cp++ = 's'; + *cp = 0; + /* BUG: we do not handled embeded NULLs */ + snprintf(sprintf_buf, JIM_MAX_FMT, fmt_str, Jim_GetString(objv[0], NULL)); + break; + case 'c': + *cp++ = 'c'; + *cp = 0; + if (Jim_GetWide(interp, objv[0], &wideValue) == JIM_ERR) { + Jim_FreeNewObj(interp, resObjPtr); + return NULL; + } + c = (char) wideValue; + snprintf(sprintf_buf, JIM_MAX_FMT, fmt_str, c); + break; + case 'f': + case 'F': + case 'g': + case 'G': + case 'e': + case 'E': + *cp++ = *fmt; + *cp = 0; + if (Jim_GetDouble(interp, objv[0], &doubleValue) == JIM_ERR) { + Jim_FreeNewObj(interp, resObjPtr); + return NULL; + } + snprintf(sprintf_buf, JIM_MAX_FMT, fmt_str, doubleValue); + break; + case 'b': + case 'd': + case 'o': + case 'i': + case 'u': + case 'x': + case 'X': + /* jim widevaluse are 64bit */ + if (sizeof(jim_wide) == sizeof(long long)) { + *cp++ = 'l'; + *cp++ = 'l'; + } else { + *cp++ = 'l'; + } + *cp++ = *fmt; + *cp = 0; + if (Jim_GetWide(interp, objv[0], &wideValue) == JIM_ERR) { + Jim_FreeNewObj(interp, resObjPtr); + return NULL; + } + snprintf(sprintf_buf, JIM_MAX_FMT, fmt_str, wideValue); + break; + case '%': + sprintf_buf[0] = '%'; + sprintf_buf[1] = 0; + objv--; /* undo the objv++ below */ + break; + default: + spec[0] = *fmt; spec[1] = '\0'; + Jim_FreeNewObj(interp, resObjPtr); + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad field specifier \"", spec, "\"", NULL); + return NULL; + } + /* force terminate */ +#if 0 + printf("FMT was: %s\n", fmt_str); + printf("RES was: |%s|\n", sprintf_buf); +#endif + + sprintf_buf[ JIM_MAX_FMT - 1] = 0; + Jim_AppendString(interp, resObjPtr, sprintf_buf, strlen(sprintf_buf)); + /* next obj */ + objv++; + fmt++; + fmtLen--; + } + return resObjPtr; +} + +Jim_Obj *Jim_FormatString(Jim_Interp *interp, Jim_Obj *fmtObjPtr, + int objc, Jim_Obj *const *objv) +{ + char *sprintf_buf = malloc(JIM_MAX_FMT); + Jim_Obj *t = Jim_FormatString_Inner(interp, fmtObjPtr, objc, objv, sprintf_buf); + free(sprintf_buf); + return t; +} + +/* ----------------------------------------------------------------------------- + * Compared String Object + * ---------------------------------------------------------------------------*/ + +/* This is strange object that allows to compare a C literal string + * with a Jim object in very short time if the same comparison is done + * multiple times. For example every time the [if] command is executed, + * Jim has to check if a given argument is "else". This comparions if + * the code has no errors are true most of the times, so we can cache + * inside the object the pointer of the string of the last matching + * comparison. Because most C compilers perform literal sharing, + * so that: char *x = "foo", char *y = "foo", will lead to x == y, + * this works pretty well even if comparisons are at different places + * inside the C code. */ + +static Jim_ObjType comparedStringObjType = { + "compared-string", + NULL, + NULL, + NULL, + JIM_TYPE_REFERENCES, +}; + +/* The only way this object is exposed to the API is via the following + * function. Returns true if the string and the object string repr. + * are the same, otherwise zero is returned. + * + * Note: this isn't binary safe, but it hardly needs to be.*/ +int Jim_CompareStringImmediate(Jim_Interp *interp, Jim_Obj *objPtr, + const char *str) +{ + if (objPtr->typePtr == &comparedStringObjType && + objPtr->internalRep.ptr == str) + return 1; + else { + const char *objStr = Jim_GetString(objPtr, NULL); + if (strcmp(str, objStr) != 0) return 0; + if (objPtr->typePtr != &comparedStringObjType) { + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &comparedStringObjType; + } + objPtr->internalRep.ptr = (char*)str; /*ATTENTION: const cast */ + return 1; + } +} + +int qsortCompareStringPointers(const void *a, const void *b) +{ + char * const *sa = (char * const *)a; + char * const *sb = (char * const *)b; + return strcmp(*sa, *sb); +} + +int Jim_GetEnum(Jim_Interp *interp, Jim_Obj *objPtr, + const char * const *tablePtr, int *indexPtr, const char *name, int flags) +{ + const char * const *entryPtr = NULL; + char **tablePtrSorted; + int i, count = 0; + + *indexPtr = -1; + for (entryPtr = tablePtr, i = 0; *entryPtr != NULL; entryPtr++, i++) { + if (Jim_CompareStringImmediate(interp, objPtr, *entryPtr)) { + *indexPtr = i; + return JIM_OK; + } + count++; /* If nothing matches, this will reach the len of tablePtr */ + } + if (flags & JIM_ERRMSG) { + if (name == NULL) + name = "option"; + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad ", name, " \"", Jim_GetString(objPtr, NULL), "\": must be one of ", + NULL); + tablePtrSorted = Jim_Alloc(sizeof(char*)*count); + memcpy(tablePtrSorted, tablePtr, sizeof(char*)*count); + qsort(tablePtrSorted, count, sizeof(char*), qsortCompareStringPointers); + for (i = 0; i < count; i++) { + if (i + 1 == count && count > 1) + Jim_AppendString(interp, Jim_GetResult(interp), "or ", -1); + Jim_AppendString(interp, Jim_GetResult(interp), + tablePtrSorted[i], -1); + if (i + 1 != count) + Jim_AppendString(interp, Jim_GetResult(interp), ", ", -1); + } + Jim_Free(tablePtrSorted); + } + return JIM_ERR; +} + +int Jim_GetNvp(Jim_Interp *interp, + Jim_Obj *objPtr, + const Jim_Nvp *nvp_table, + const Jim_Nvp ** result) +{ + Jim_Nvp *n; + int e; + + e = Jim_Nvp_name2value_obj(interp, nvp_table, objPtr, &n); + if (e == JIM_ERR) { + return e; + } + + /* Success? found? */ + if (n->name) { + /* remove const */ + *result = (Jim_Nvp *)n; + return JIM_OK; + } else { + return JIM_ERR; + } +} + +/* ----------------------------------------------------------------------------- + * Source Object + * + * This object is just a string from the language point of view, but + * in the internal representation it contains the filename and line number + * where this given token was read. This information is used by + * Jim_EvalObj() if the object passed happens to be of type "source". + * + * This allows to propagate the information about line numbers and file + * names and give error messages with absolute line numbers. + * + * Note that this object uses shared strings for filenames, and the + * pointer to the filename together with the line number is taken into + * the space for the "inline" internal represenation of the Jim_Object, + * so there is almost memory zero-overhead. + * + * Also the object will be converted to something else if the given + * token it represents in the source file is not something to be + * evaluated (not a script), and will be specialized in some other way, + * so the time overhead is alzo null. + * ---------------------------------------------------------------------------*/ + +static void FreeSourceInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupSourceInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); + +static Jim_ObjType sourceObjType = { + "source", + FreeSourceInternalRep, + DupSourceInternalRep, + NULL, + JIM_TYPE_REFERENCES, +}; + +void FreeSourceInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_ReleaseSharedString(interp, + objPtr->internalRep.sourceValue.fileName); +} + +void DupSourceInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + dupPtr->internalRep.sourceValue.fileName = + Jim_GetSharedString(interp, + srcPtr->internalRep.sourceValue.fileName); + dupPtr->internalRep.sourceValue.lineNumber = + dupPtr->internalRep.sourceValue.lineNumber; + dupPtr->typePtr = &sourceObjType; +} + +static void JimSetSourceInfo(Jim_Interp *interp, Jim_Obj *objPtr, + const char *fileName, int lineNumber) +{ + if (Jim_IsShared(objPtr)) + Jim_Panic(interp,"JimSetSourceInfo called with shared object"); + if (objPtr->typePtr != NULL) + Jim_Panic(interp,"JimSetSourceInfo called with typePtr != NULL"); + objPtr->internalRep.sourceValue.fileName = + Jim_GetSharedString(interp, fileName); + objPtr->internalRep.sourceValue.lineNumber = lineNumber; + objPtr->typePtr = &sourceObjType; +} + +/* ----------------------------------------------------------------------------- + * Script Object + * ---------------------------------------------------------------------------*/ + +#define JIM_CMDSTRUCT_EXPAND -1 + +static void FreeScriptInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupScriptInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static int SetScriptFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType scriptObjType = { + "script", + FreeScriptInternalRep, + DupScriptInternalRep, + NULL, + JIM_TYPE_REFERENCES, +}; + +/* The ScriptToken structure represents every token into a scriptObj. + * Every token contains an associated Jim_Obj that can be specialized + * by commands operating on it. */ +typedef struct ScriptToken { + int type; + Jim_Obj *objPtr; + int linenr; +} ScriptToken; + +/* This is the script object internal representation. An array of + * ScriptToken structures, with an associated command structure array. + * The command structure is a pre-computed representation of the + * command length and arguments structure as a simple liner array + * of integers. + * + * For example the script: + * + * puts hello + * set $i $x$y [foo]BAR + * + * will produce a ScriptObj with the following Tokens: + * + * ESC puts + * SEP + * ESC hello + * EOL + * ESC set + * EOL + * VAR i + * SEP + * VAR x + * VAR y + * SEP + * CMD foo + * ESC BAR + * EOL + * + * This is a description of the tokens, separators, and of lines. + * The command structure instead represents the number of arguments + * of every command, followed by the tokens of which every argument + * is composed. So for the example script, the cmdstruct array will + * contain: + * + * 2 1 1 4 1 1 2 2 + * + * Because "puts hello" has two args (2), composed of single tokens (1 1) + * While "set $i $x$y [foo]BAR" has four (4) args, the first two + * composed of single tokens (1 1) and the last two of double tokens + * (2 2). + * + * The precomputation of the command structure makes Jim_Eval() faster, + * and simpler because there aren't dynamic lengths / allocations. + * + * -- {expand} handling -- + * + * Expand is handled in a special way. When a command + * contains at least an argument with the {expand} prefix, + * the command structure presents a -1 before the integer + * describing the number of arguments. This is used in order + * to send the command exection to a different path in case + * of {expand} and guarantee a fast path for the more common + * case. Also, the integers describing the number of tokens + * are expressed with negative sign, to allow for fast check + * of what's an {expand}-prefixed argument and what not. + * + * For example the command: + * + * list {expand}{1 2} + * + * Will produce the following cmdstruct array: + * + * -1 2 1 -2 + * + * -- the substFlags field of the structure -- + * + * The scriptObj structure is used to represent both "script" objects + * and "subst" objects. In the second case, the cmdStruct related + * fields are not used at all, but there is an additional field used + * that is 'substFlags': this represents the flags used to turn + * the string into the intenral representation used to perform the + * substitution. If this flags are not what the application requires + * the scriptObj is created again. For example the script: + * + * subst -nocommands $string + * subst -novariables $string + * + * Will recreate the internal representation of the $string object + * two times. + */ +typedef struct ScriptObj { + int len; /* Length as number of tokens. */ + int commands; /* number of top-level commands in script. */ + ScriptToken *token; /* Tokens array. */ + int *cmdStruct; /* commands structure */ + int csLen; /* length of the cmdStruct array. */ + int substFlags; /* flags used for the compilation of "subst" objects */ + int inUse; /* Used to share a ScriptObj. Currently + only used by Jim_EvalObj() as protection against + shimmering of the currently evaluated object. */ + char *fileName; +} ScriptObj; + +void FreeScriptInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + int i; + struct ScriptObj *script = (void*) objPtr->internalRep.ptr; + + if (!script) + return; + + script->inUse--; + if (script->inUse != 0) return; + for (i = 0; i < script->len; i++) { + if (script->token[i].objPtr != NULL) + Jim_DecrRefCount(interp, script->token[i].objPtr); + } + Jim_Free(script->token); + Jim_Free(script->cmdStruct); + Jim_Free(script->fileName); + Jim_Free(script); +} + +void DupScriptInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + JIM_NOTUSED(interp); + JIM_NOTUSED(srcPtr); + + /* Just returns an simple string. */ + dupPtr->typePtr = NULL; +} + +/* Add a new token to the internal repr of a script object */ +static void ScriptObjAddToken(Jim_Interp *interp, struct ScriptObj *script, + char *strtoken, int len, int type, char *filename, int linenr) +{ + int prevtype; + struct ScriptToken *token; + + prevtype = (script->len == 0) ? JIM_TT_EOL : \ + script->token[script->len-1].type; + /* Skip tokens without meaning, like words separators + * following a word separator or an end of command and + * so on. */ + if (prevtype == JIM_TT_EOL) { + if (type == JIM_TT_EOL || type == JIM_TT_SEP) { + Jim_Free(strtoken); + return; + } + } else if (prevtype == JIM_TT_SEP) { + if (type == JIM_TT_SEP) { + Jim_Free(strtoken); + return; + } else if (type == JIM_TT_EOL) { + /* If an EOL is following by a SEP, drop the previous + * separator. */ + script->len--; + Jim_DecrRefCount(interp, script->token[script->len].objPtr); + } + } else if (prevtype != JIM_TT_EOL && prevtype != JIM_TT_SEP && + type == JIM_TT_ESC && len == 0) + { + /* Don't add empty tokens used in interpolation */ + Jim_Free(strtoken); + return; + } + /* Make space for a new istruction */ + script->len++; + script->token = Jim_Realloc(script->token, + sizeof(ScriptToken)*script->len); + /* Initialize the new token */ + token = script->token + (script->len-1); + token->type = type; + /* Every object is intially as a string, but the + * internal type may be specialized during execution of the + * script. */ + token->objPtr = Jim_NewStringObjNoAlloc(interp, strtoken, len); + /* To add source info to SEP and EOL tokens is useless because + * they will never by called as arguments of Jim_EvalObj(). */ + if (filename && type != JIM_TT_SEP && type != JIM_TT_EOL) + JimSetSourceInfo(interp, token->objPtr, filename, linenr); + Jim_IncrRefCount(token->objPtr); + token->linenr = linenr; +} + +/* Add an integer into the command structure field of the script object. */ +static void ScriptObjAddInt(struct ScriptObj *script, int val) +{ + script->csLen++; + script->cmdStruct = Jim_Realloc(script->cmdStruct, + sizeof(int)*script->csLen); + script->cmdStruct[script->csLen-1] = val; +} + +/* Search a Jim_Obj contained in 'script' with the same stinrg repr. + * of objPtr. Search nested script objects recursively. */ +static Jim_Obj *ScriptSearchLiteral(Jim_Interp *interp, ScriptObj *script, + ScriptObj *scriptBarrier, Jim_Obj *objPtr) +{ + int i; + + for (i = 0; i < script->len; i++) { + if (script->token[i].objPtr != objPtr && + Jim_StringEqObj(script->token[i].objPtr, objPtr, 0)) { + return script->token[i].objPtr; + } + /* Enter recursively on scripts only if the object + * is not the same as the one we are searching for + * shared occurrences. */ + if (script->token[i].objPtr->typePtr == &scriptObjType && + script->token[i].objPtr != objPtr) { + Jim_Obj *foundObjPtr; + + ScriptObj *subScript = + script->token[i].objPtr->internalRep.ptr; + /* Don't recursively enter the script we are trying + * to make shared to avoid circular references. */ + if (subScript == scriptBarrier) continue; + if (subScript != script) { + foundObjPtr = + ScriptSearchLiteral(interp, subScript, + scriptBarrier, objPtr); + if (foundObjPtr != NULL) + return foundObjPtr; + } + } + } + return NULL; +} + +/* Share literals of a script recursively sharing sub-scripts literals. */ +static void ScriptShareLiterals(Jim_Interp *interp, ScriptObj *script, + ScriptObj *topLevelScript) +{ + int i, j; + + return; + /* Try to share with toplevel object. */ + if (topLevelScript != NULL) { + for (i = 0; i < script->len; i++) { + Jim_Obj *foundObjPtr; + char *str = script->token[i].objPtr->bytes; + + if (script->token[i].objPtr->refCount != 1) continue; + if (script->token[i].objPtr->typePtr == &scriptObjType) continue; + if (strchr(str, ' ') || strchr(str, '\n')) continue; + foundObjPtr = ScriptSearchLiteral(interp, + topLevelScript, + script, /* barrier */ + script->token[i].objPtr); + if (foundObjPtr != NULL) { + Jim_IncrRefCount(foundObjPtr); + Jim_DecrRefCount(interp, + script->token[i].objPtr); + script->token[i].objPtr = foundObjPtr; + } + } + } + /* Try to share locally */ + for (i = 0; i < script->len; i++) { + char *str = script->token[i].objPtr->bytes; + + if (script->token[i].objPtr->refCount != 1) continue; + if (strchr(str, ' ') || strchr(str, '\n')) continue; + for (j = 0; j < script->len; j++) { + if (script->token[i].objPtr != + script->token[j].objPtr && + Jim_StringEqObj(script->token[i].objPtr, + script->token[j].objPtr, 0)) + { + Jim_IncrRefCount(script->token[j].objPtr); + Jim_DecrRefCount(interp, + script->token[i].objPtr); + script->token[i].objPtr = + script->token[j].objPtr; + } + } + } +} + +/* This method takes the string representation of an object + * as a Tcl script, and generates the pre-parsed internal representation + * of the script. */ +int SetScriptFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr) +{ + int scriptTextLen; + const char *scriptText = Jim_GetString(objPtr, &scriptTextLen); + struct JimParserCtx parser; + struct ScriptObj *script = Jim_Alloc(sizeof(*script)); + ScriptToken *token; + int args, tokens, start, end, i; + int initialLineNumber; + int propagateSourceInfo = 0; + + script->len = 0; + script->csLen = 0; + script->commands = 0; + script->token = NULL; + script->cmdStruct = NULL; + script->inUse = 1; + /* Try to get information about filename / line number */ + if (objPtr->typePtr == &sourceObjType) { + script->fileName = + Jim_StrDup(objPtr->internalRep.sourceValue.fileName); + initialLineNumber = objPtr->internalRep.sourceValue.lineNumber; + propagateSourceInfo = 1; + } else { + script->fileName = Jim_StrDup(""); + initialLineNumber = 1; + } + + JimParserInit(&parser, scriptText, scriptTextLen, initialLineNumber); + while (!JimParserEof(&parser)) { + char *token_t; + int len, type, linenr; + + JimParseScript(&parser); + token_t = JimParserGetToken(&parser, &len, &type, &linenr); + ScriptObjAddToken(interp, script, token_t, len, type, + propagateSourceInfo ? script->fileName : NULL, + linenr); + } + token = script->token; + + /* Compute the command structure array + * (see the ScriptObj struct definition for more info) */ + start = 0; /* Current command start token index */ + end = -1; /* Current command end token index */ + while (1) { + int expand = 0; /* expand flag. set to 1 on {expand} form. */ + int interpolation = 0; /* set to 1 if there is at least one + argument of the command obtained via + interpolation of more tokens. */ + /* Search for the end of command, while + * count the number of args. */ + start = ++end; + if (start >= script->len) break; + args = 1; /* Number of args in current command */ + while (token[end].type != JIM_TT_EOL) { + if (end == 0 || token[end-1].type == JIM_TT_SEP || + token[end-1].type == JIM_TT_EOL) + { + if (token[end].type == JIM_TT_STR && + token[end + 1].type != JIM_TT_SEP && + token[end + 1].type != JIM_TT_EOL && + (!strcmp(token[end].objPtr->bytes, "expand") || + !strcmp(token[end].objPtr->bytes, "*"))) + expand++; + } + if (token[end].type == JIM_TT_SEP) + args++; + end++; + } + interpolation = !((end-start + 1) == args*2); + /* Add the 'number of arguments' info into cmdstruct. + * Negative value if there is list expansion involved. */ + if (expand) + ScriptObjAddInt(script, -1); + ScriptObjAddInt(script, args); + /* Now add info about the number of tokens. */ + tokens = 0; /* Number of tokens in current argument. */ + expand = 0; + for (i = start; i <= end; i++) { + if (token[i].type == JIM_TT_SEP || + token[i].type == JIM_TT_EOL) + { + if (tokens == 1 && expand) + expand = 0; + ScriptObjAddInt(script, + expand ? -tokens : tokens); + + expand = 0; + tokens = 0; + continue; + } else if (tokens == 0 && token[i].type == JIM_TT_STR && + (!strcmp(token[i].objPtr->bytes, "expand") || + !strcmp(token[i].objPtr->bytes, "*"))) + { + expand++; + } + tokens++; + } + } + /* Perform literal sharing, but only for objects that appear + * to be scripts written as literals inside the source code, + * and not computed at runtime. Literal sharing is a costly + * operation that should be done only against objects that + * are likely to require compilation only the first time, and + * then are executed multiple times. */ + if (propagateSourceInfo && interp->framePtr->procBodyObjPtr) { + Jim_Obj *bodyObjPtr = interp->framePtr->procBodyObjPtr; + if (bodyObjPtr->typePtr == &scriptObjType) { + ScriptObj *bodyScript = + bodyObjPtr->internalRep.ptr; + ScriptShareLiterals(interp, script, bodyScript); + } + } else if (propagateSourceInfo) { + ScriptShareLiterals(interp, script, NULL); + } + /* Free the old internal rep and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + Jim_SetIntRepPtr(objPtr, script); + objPtr->typePtr = &scriptObjType; + return JIM_OK; +} + +ScriptObj *Jim_GetScript(Jim_Interp *interp, Jim_Obj *objPtr) +{ + if (objPtr->typePtr != &scriptObjType) { + SetScriptFromAny(interp, objPtr); + } + return (ScriptObj*) Jim_GetIntRepPtr(objPtr); +} + +/* ----------------------------------------------------------------------------- + * Commands + * ---------------------------------------------------------------------------*/ + +/* Commands HashTable Type. + * + * Keys are dynamic allocated strings, Values are Jim_Cmd structures. */ +static void Jim_CommandsHT_ValDestructor(void *interp, void *val) +{ + Jim_Cmd *cmdPtr = (void*) val; + + if (cmdPtr->cmdProc == NULL) { + Jim_DecrRefCount(interp, cmdPtr->argListObjPtr); + Jim_DecrRefCount(interp, cmdPtr->bodyObjPtr); + if (cmdPtr->staticVars) { + Jim_FreeHashTable(cmdPtr->staticVars); + Jim_Free(cmdPtr->staticVars); + } + } else if (cmdPtr->delProc != NULL) { + /* If it was a C coded command, call the delProc if any */ + cmdPtr->delProc(interp, cmdPtr->privData); + } + Jim_Free(val); +} + +static Jim_HashTableType JimCommandsHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + JimStringCopyHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + Jim_CommandsHT_ValDestructor /* val destructor */ +}; + +/* ------------------------- Commands related functions --------------------- */ + +int Jim_CreateCommand(Jim_Interp *interp, const char *cmdName, + Jim_CmdProc cmdProc, void *privData, Jim_DelCmdProc delProc) +{ + Jim_HashEntry *he; + Jim_Cmd *cmdPtr; + + he = Jim_FindHashEntry(&interp->commands, cmdName); + if (he == NULL) { /* New command to create */ + cmdPtr = Jim_Alloc(sizeof(*cmdPtr)); + Jim_AddHashEntry(&interp->commands, cmdName, cmdPtr); + } else { + Jim_InterpIncrProcEpoch(interp); + /* Free the arglist/body objects if it was a Tcl procedure */ + cmdPtr = he->val; + if (cmdPtr->cmdProc == NULL) { + Jim_DecrRefCount(interp, cmdPtr->argListObjPtr); + Jim_DecrRefCount(interp, cmdPtr->bodyObjPtr); + if (cmdPtr->staticVars) { + Jim_FreeHashTable(cmdPtr->staticVars); + Jim_Free(cmdPtr->staticVars); + } + cmdPtr->staticVars = NULL; + } else if (cmdPtr->delProc != NULL) { + /* If it was a C coded command, call the delProc if any */ + cmdPtr->delProc(interp, cmdPtr->privData); + } + } + + /* Store the new details for this proc */ + cmdPtr->delProc = delProc; + cmdPtr->cmdProc = cmdProc; + cmdPtr->privData = privData; + + /* There is no need to increment the 'proc epoch' because + * creation of a new procedure can never affect existing + * cached commands. We don't do negative caching. */ + return JIM_OK; +} + +int Jim_CreateProcedure(Jim_Interp *interp, const char *cmdName, + Jim_Obj *argListObjPtr, Jim_Obj *staticsListObjPtr, Jim_Obj *bodyObjPtr, + int arityMin, int arityMax) +{ + Jim_Cmd *cmdPtr; + + cmdPtr = Jim_Alloc(sizeof(*cmdPtr)); + cmdPtr->cmdProc = NULL; /* Not a C coded command */ + cmdPtr->argListObjPtr = argListObjPtr; + cmdPtr->bodyObjPtr = bodyObjPtr; + Jim_IncrRefCount(argListObjPtr); + Jim_IncrRefCount(bodyObjPtr); + cmdPtr->arityMin = arityMin; + cmdPtr->arityMax = arityMax; + cmdPtr->staticVars = NULL; + + /* Create the statics hash table. */ + if (staticsListObjPtr) { + int len, i; + + Jim_ListLength(interp, staticsListObjPtr, &len); + if (len != 0) { + cmdPtr->staticVars = Jim_Alloc(sizeof(Jim_HashTable)); + Jim_InitHashTable(cmdPtr->staticVars, getJimVariablesHashTableType(), + interp); + for (i = 0; i < len; i++) { + Jim_Obj *objPtr=NULL, *initObjPtr=NULL, *nameObjPtr=NULL; + Jim_Var *varPtr; + int subLen; + + Jim_ListIndex(interp, staticsListObjPtr, i, &objPtr, JIM_NONE); + /* Check if it's composed of two elements. */ + Jim_ListLength(interp, objPtr, &subLen); + if (subLen == 1 || subLen == 2) { + /* Try to get the variable value from the current + * environment. */ + Jim_ListIndex(interp, objPtr, 0, &nameObjPtr, JIM_NONE); + if (subLen == 1) { + initObjPtr = Jim_GetVariable(interp, nameObjPtr, + JIM_NONE); + if (initObjPtr == NULL) { + Jim_SetResult(interp, + Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "variable for initialization of static \"", + Jim_GetString(nameObjPtr, NULL), + "\" not found in the local context", + NULL); + goto err; + } + } else { + Jim_ListIndex(interp, objPtr, 1, &initObjPtr, JIM_NONE); + } + varPtr = Jim_Alloc(sizeof(*varPtr)); + varPtr->objPtr = initObjPtr; + Jim_IncrRefCount(initObjPtr); + varPtr->linkFramePtr = NULL; + if (Jim_AddHashEntry(cmdPtr->staticVars, + Jim_GetString(nameObjPtr, NULL), + varPtr) != JIM_OK) + { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "static variable name \"", + Jim_GetString(objPtr, NULL), "\"", + " duplicated in statics list", NULL); + Jim_DecrRefCount(interp, initObjPtr); + Jim_Free(varPtr); + goto err; + } + } else { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "too many fields in static specifier \"", + objPtr, "\"", NULL); + goto err; + } + } + } + } + + /* Add the new command */ + + /* it may already exist, so we try to delete the old one */ + if (Jim_DeleteHashEntry(&interp->commands, cmdName) != JIM_ERR) { + /* There was an old procedure with the same name, this requires + * a 'proc epoch' update. */ + Jim_InterpIncrProcEpoch(interp); + } + /* If a procedure with the same name didn't existed there is no need + * to increment the 'proc epoch' because creation of a new procedure + * can never affect existing cached commands. We don't do + * negative caching. */ + Jim_AddHashEntry(&interp->commands, cmdName, cmdPtr); + return JIM_OK; + +err: + Jim_FreeHashTable(cmdPtr->staticVars); + Jim_Free(cmdPtr->staticVars); + Jim_DecrRefCount(interp, argListObjPtr); + Jim_DecrRefCount(interp, bodyObjPtr); + Jim_Free(cmdPtr); + return JIM_ERR; +} + +int Jim_DeleteCommand(Jim_Interp *interp, const char *cmdName) +{ + if (Jim_DeleteHashEntry(&interp->commands, cmdName) == JIM_ERR) + return JIM_ERR; + Jim_InterpIncrProcEpoch(interp); + return JIM_OK; +} + +int Jim_RenameCommand(Jim_Interp *interp, const char *oldName, + const char *newName) +{ + Jim_Cmd *cmdPtr; + Jim_HashEntry *he; + Jim_Cmd *copyCmdPtr; + + if (newName[0] == '\0') /* Delete! */ + return Jim_DeleteCommand(interp, oldName); + /* Rename */ + he = Jim_FindHashEntry(&interp->commands, oldName); + if (he == NULL) + return JIM_ERR; /* Invalid command name */ + cmdPtr = he->val; + copyCmdPtr = Jim_Alloc(sizeof(Jim_Cmd)); + *copyCmdPtr = *cmdPtr; + /* In order to avoid that a procedure will get arglist/body/statics + * freed by the hash table methods, fake a C-coded command + * setting cmdPtr->cmdProc as not NULL */ + cmdPtr->cmdProc = (void*)1; + /* Also make sure delProc is NULL. */ + cmdPtr->delProc = NULL; + /* Destroy the old command, and make sure the new is freed + * as well. */ + Jim_DeleteHashEntry(&interp->commands, oldName); + Jim_DeleteHashEntry(&interp->commands, newName); + /* Now the new command. We are sure it can't fail because + * the target name was already freed. */ + Jim_AddHashEntry(&interp->commands, newName, copyCmdPtr); + /* Increment the epoch */ + Jim_InterpIncrProcEpoch(interp); + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * Command object + * ---------------------------------------------------------------------------*/ + +static int SetCommandFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType commandObjType = { + "command", + NULL, + NULL, + NULL, + JIM_TYPE_REFERENCES, +}; + +int SetCommandFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_HashEntry *he; + const char *cmdName; + + /* Get the string representation */ + cmdName = Jim_GetString(objPtr, NULL); + /* Lookup this name into the commands hash table */ + he = Jim_FindHashEntry(&interp->commands, cmdName); + if (he == NULL) + return JIM_ERR; + + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &commandObjType; + objPtr->internalRep.cmdValue.procEpoch = interp->procEpoch; + objPtr->internalRep.cmdValue.cmdPtr = (void*)he->val; + return JIM_OK; +} + +/* This function returns the command structure for the command name + * stored in objPtr. It tries to specialize the objPtr to contain + * a cached info instead to perform the lookup into the hash table + * every time. The information cached may not be uptodate, in such + * a case the lookup is performed and the cache updated. */ +Jim_Cmd *Jim_GetCommand(Jim_Interp *interp, Jim_Obj *objPtr, int flags) +{ + if ((objPtr->typePtr != &commandObjType || + objPtr->internalRep.cmdValue.procEpoch != interp->procEpoch) && + SetCommandFromAny(interp, objPtr) == JIM_ERR) { + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "invalid command name \"", objPtr->bytes, "\"", + NULL); + } + return NULL; + } + return objPtr->internalRep.cmdValue.cmdPtr; +} + +/* ----------------------------------------------------------------------------- + * Variables + * ---------------------------------------------------------------------------*/ + +/* Variables HashTable Type. + * + * Keys are dynamic allocated strings, Values are Jim_Var structures. */ +static void JimVariablesHTValDestructor(void *interp, void *val) +{ + Jim_Var *varPtr = (void*) val; + + Jim_DecrRefCount(interp, varPtr->objPtr); + Jim_Free(val); +} + +static Jim_HashTableType JimVariablesHashTableType = { + JimStringCopyHTHashFunction, /* hash function */ + JimStringCopyHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimStringCopyHTKeyCompare, /* key compare */ + JimStringCopyHTKeyDestructor, /* key destructor */ + JimVariablesHTValDestructor /* val destructor */ +}; + +static Jim_HashTableType *getJimVariablesHashTableType(void) +{ + return &JimVariablesHashTableType; +} + +/* ----------------------------------------------------------------------------- + * Variable object + * ---------------------------------------------------------------------------*/ + +#define JIM_DICT_SUGAR 100 /* Only returned by SetVariableFromAny() */ + +static int SetVariableFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType variableObjType = { + "variable", + NULL, + NULL, + NULL, + JIM_TYPE_REFERENCES, +}; + +/* Return true if the string "str" looks like syntax sugar for [dict]. I.e. + * is in the form "varname(key)". */ +static int Jim_NameIsDictSugar(const char *str, int len) +{ + if (len == -1) + len = strlen(str); + if (len && str[len-1] == ')' && strchr(str, '(') != NULL) + return 1; + return 0; +} + +/* This method should be called only by the variable API. + * It returns JIM_OK on success (variable already exists), + * JIM_ERR if it does not exists, JIM_DICT_GLUE if it's not + * a variable name, but syntax glue for [dict] i.e. the last + * character is ')' */ +int SetVariableFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr) +{ + Jim_HashEntry *he; + const char *varName; + int len; + + /* Check if the object is already an uptodate variable */ + if (objPtr->typePtr == &variableObjType && + objPtr->internalRep.varValue.callFrameId == interp->framePtr->id) + return JIM_OK; /* nothing to do */ + /* Get the string representation */ + varName = Jim_GetString(objPtr, &len); + /* Make sure it's not syntax glue to get/set dict. */ + if (Jim_NameIsDictSugar(varName, len)) + return JIM_DICT_SUGAR; + if (varName[0] == ':' && varName[1] == ':') { + he = Jim_FindHashEntry(&interp->topFramePtr->vars, varName + 2); + if (he == NULL) { + return JIM_ERR; + } + } + else { + /* Lookup this name into the variables hash table */ + he = Jim_FindHashEntry(&interp->framePtr->vars, varName); + if (he == NULL) { + /* Try with static vars. */ + if (interp->framePtr->staticVars == NULL) + return JIM_ERR; + if (!(he = Jim_FindHashEntry(interp->framePtr->staticVars, varName))) + return JIM_ERR; + } + } + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &variableObjType; + objPtr->internalRep.varValue.callFrameId = interp->framePtr->id; + objPtr->internalRep.varValue.varPtr = (void*)he->val; + return JIM_OK; +} + +/* -------------------- Variables related functions ------------------------- */ +static int JimDictSugarSet(Jim_Interp *interp, Jim_Obj *ObjPtr, + Jim_Obj *valObjPtr); +static Jim_Obj *JimDictSugarGet(Jim_Interp *interp, Jim_Obj *ObjPtr); + +/* For now that's dummy. Variables lookup should be optimized + * in many ways, with caching of lookups, and possibly with + * a table of pre-allocated vars in every CallFrame for local vars. + * All the caching should also have an 'epoch' mechanism similar + * to the one used by Tcl for procedures lookup caching. */ + +int Jim_SetVariable(Jim_Interp *interp, Jim_Obj *nameObjPtr, Jim_Obj *valObjPtr) +{ + const char *name; + Jim_Var *var; + int err; + + if ((err = SetVariableFromAny(interp, nameObjPtr)) != JIM_OK) { + /* Check for [dict] syntax sugar. */ + if (err == JIM_DICT_SUGAR) + return JimDictSugarSet(interp, nameObjPtr, valObjPtr); + /* New variable to create */ + name = Jim_GetString(nameObjPtr, NULL); + + var = Jim_Alloc(sizeof(*var)); + var->objPtr = valObjPtr; + Jim_IncrRefCount(valObjPtr); + var->linkFramePtr = NULL; + /* Insert the new variable */ + if (name[0] == ':' && name[1] == ':') { + /* Into to the top evel frame */ + Jim_AddHashEntry(&interp->topFramePtr->vars, name + 2, var); + } + else { + Jim_AddHashEntry(&interp->framePtr->vars, name, var); + } + /* Make the object int rep a variable */ + Jim_FreeIntRep(interp, nameObjPtr); + nameObjPtr->typePtr = &variableObjType; + nameObjPtr->internalRep.varValue.callFrameId = + interp->framePtr->id; + nameObjPtr->internalRep.varValue.varPtr = var; + } else { + var = nameObjPtr->internalRep.varValue.varPtr; + if (var->linkFramePtr == NULL) { + Jim_IncrRefCount(valObjPtr); + Jim_DecrRefCount(interp, var->objPtr); + var->objPtr = valObjPtr; + } else { /* Else handle the link */ + Jim_CallFrame *savedCallFrame; + + savedCallFrame = interp->framePtr; + interp->framePtr = var->linkFramePtr; + err = Jim_SetVariable(interp, var->objPtr, valObjPtr); + interp->framePtr = savedCallFrame; + if (err != JIM_OK) + return err; + } + } + return JIM_OK; +} + +int Jim_SetVariableStr(Jim_Interp *interp, const char *name, Jim_Obj *objPtr) +{ + Jim_Obj *nameObjPtr; + int result; + + nameObjPtr = Jim_NewStringObj(interp, name, -1); + Jim_IncrRefCount(nameObjPtr); + result = Jim_SetVariable(interp, nameObjPtr, objPtr); + Jim_DecrRefCount(interp, nameObjPtr); + return result; +} + +int Jim_SetGlobalVariableStr(Jim_Interp *interp, const char *name, Jim_Obj *objPtr) +{ + Jim_CallFrame *savedFramePtr; + int result; + + savedFramePtr = interp->framePtr; + interp->framePtr = interp->topFramePtr; + result = Jim_SetVariableStr(interp, name, objPtr); + interp->framePtr = savedFramePtr; + return result; +} + +int Jim_SetVariableStrWithStr(Jim_Interp *interp, const char *name, const char *val) +{ + Jim_Obj *nameObjPtr, *valObjPtr; + int result; + + nameObjPtr = Jim_NewStringObj(interp, name, -1); + valObjPtr = Jim_NewStringObj(interp, val, -1); + Jim_IncrRefCount(nameObjPtr); + Jim_IncrRefCount(valObjPtr); + result = Jim_SetVariable(interp, nameObjPtr, valObjPtr); + Jim_DecrRefCount(interp, nameObjPtr); + Jim_DecrRefCount(interp, valObjPtr); + return result; +} + +int Jim_SetVariableLink(Jim_Interp *interp, Jim_Obj *nameObjPtr, + Jim_Obj *targetNameObjPtr, Jim_CallFrame *targetCallFrame) +{ + const char *varName; + int len; + + /* Check for cycles. */ + if (interp->framePtr == targetCallFrame) { + Jim_Obj *objPtr = targetNameObjPtr; + Jim_Var *varPtr; + /* Cycles are only possible with 'uplevel 0' */ + while (1) { + if (Jim_StringEqObj(objPtr, nameObjPtr, 0)) { + Jim_SetResultString(interp, + "can't upvar from variable to itself", -1); + return JIM_ERR; + } + if (SetVariableFromAny(interp, objPtr) != JIM_OK) + break; + varPtr = objPtr->internalRep.varValue.varPtr; + if (varPtr->linkFramePtr != targetCallFrame) break; + objPtr = varPtr->objPtr; + } + } + varName = Jim_GetString(nameObjPtr, &len); + if (Jim_NameIsDictSugar(varName, len)) { + Jim_SetResultString(interp, + "Dict key syntax invalid as link source", -1); + return JIM_ERR; + } + /* Perform the binding */ + Jim_SetVariable(interp, nameObjPtr, targetNameObjPtr); + /* We are now sure 'nameObjPtr' type is variableObjType */ + nameObjPtr->internalRep.varValue.varPtr->linkFramePtr = targetCallFrame; + return JIM_OK; +} + +/* Return the Jim_Obj pointer associated with a variable name, + * or NULL if the variable was not found in the current context. + * The same optimization discussed in the comment to the + * 'SetVariable' function should apply here. */ +Jim_Obj *Jim_GetVariable(Jim_Interp *interp, Jim_Obj *nameObjPtr, int flags) +{ + int err; + + /* All the rest is handled here */ + if ((err = SetVariableFromAny(interp, nameObjPtr)) != JIM_OK) { + /* Check for [dict] syntax sugar. */ + if (err == JIM_DICT_SUGAR) + return JimDictSugarGet(interp, nameObjPtr); + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "can't read \"", nameObjPtr->bytes, + "\": no such variable", NULL); + } + return NULL; + } else { + Jim_Var *varPtr; + Jim_Obj *objPtr; + Jim_CallFrame *savedCallFrame; + + varPtr = nameObjPtr->internalRep.varValue.varPtr; + if (varPtr->linkFramePtr == NULL) + return varPtr->objPtr; + /* The variable is a link? Resolve it. */ + savedCallFrame = interp->framePtr; + interp->framePtr = varPtr->linkFramePtr; + objPtr = Jim_GetVariable(interp, varPtr->objPtr, JIM_NONE); + if (objPtr == NULL && flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "can't read \"", nameObjPtr->bytes, + "\": no such variable", NULL); + } + interp->framePtr = savedCallFrame; + return objPtr; + } +} + +Jim_Obj *Jim_GetGlobalVariable(Jim_Interp *interp, Jim_Obj *nameObjPtr, + int flags) +{ + Jim_CallFrame *savedFramePtr; + Jim_Obj *objPtr; + + savedFramePtr = interp->framePtr; + interp->framePtr = interp->topFramePtr; + objPtr = Jim_GetVariable(interp, nameObjPtr, flags); + interp->framePtr = savedFramePtr; + + return objPtr; +} + +Jim_Obj *Jim_GetVariableStr(Jim_Interp *interp, const char *name, int flags) +{ + Jim_Obj *nameObjPtr, *varObjPtr; + + nameObjPtr = Jim_NewStringObj(interp, name, -1); + Jim_IncrRefCount(nameObjPtr); + varObjPtr = Jim_GetVariable(interp, nameObjPtr, flags); + Jim_DecrRefCount(interp, nameObjPtr); + return varObjPtr; +} + +Jim_Obj *Jim_GetGlobalVariableStr(Jim_Interp *interp, const char *name, + int flags) +{ + Jim_CallFrame *savedFramePtr; + Jim_Obj *objPtr; + + savedFramePtr = interp->framePtr; + interp->framePtr = interp->topFramePtr; + objPtr = Jim_GetVariableStr(interp, name, flags); + interp->framePtr = savedFramePtr; + + return objPtr; +} + +/* Unset a variable. + * Note: On success unset invalidates all the variable objects created + * in the current call frame incrementing. */ +int Jim_UnsetVariable(Jim_Interp *interp, Jim_Obj *nameObjPtr, int flags) +{ + const char *name; + Jim_Var *varPtr; + int err; + + if ((err = SetVariableFromAny(interp, nameObjPtr)) != JIM_OK) { + /* Check for [dict] syntax sugar. */ + if (err == JIM_DICT_SUGAR) + return JimDictSugarSet(interp, nameObjPtr, NULL); + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "can't unset \"", nameObjPtr->bytes, + "\": no such variable", NULL); + return JIM_ERR; /* var not found */ + } + varPtr = nameObjPtr->internalRep.varValue.varPtr; + /* If it's a link call UnsetVariable recursively */ + if (varPtr->linkFramePtr) { + int retval; + + Jim_CallFrame *savedCallFrame; + + savedCallFrame = interp->framePtr; + interp->framePtr = varPtr->linkFramePtr; + retval = Jim_UnsetVariable(interp, varPtr->objPtr, JIM_NONE); + interp->framePtr = savedCallFrame; + if (retval != JIM_OK && flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "can't unset \"", nameObjPtr->bytes, + "\": no such variable", NULL); + } + return retval; + } else { + name = Jim_GetString(nameObjPtr, NULL); + if (Jim_DeleteHashEntry(&interp->framePtr->vars, name) + != JIM_OK) return JIM_ERR; + /* Change the callframe id, invalidating var lookup caching */ + JimChangeCallFrameId(interp, interp->framePtr); + return JIM_OK; + } +} + +/* ---------- Dict syntax sugar (similar to array Tcl syntax) -------------- */ + +/* Given a variable name for [dict] operation syntax sugar, + * this function returns two objects, the first with the name + * of the variable to set, and the second with the rispective key. + * For example "foo(bar)" will return objects with string repr. of + * "foo" and "bar". + * + * The returned objects have refcount = 1. The function can't fail. */ +static void JimDictSugarParseVarKey(Jim_Interp *interp, Jim_Obj *objPtr, + Jim_Obj **varPtrPtr, Jim_Obj **keyPtrPtr) +{ + const char *str, *p; + char *t; + int len, keyLen, nameLen; + Jim_Obj *varObjPtr, *keyObjPtr; + + str = Jim_GetString(objPtr, &len); + p = strchr(str, '('); + p++; + keyLen = len-((p-str) + 1); + nameLen = (p-str)-1; + /* Create the objects with the variable name and key. */ + t = Jim_Alloc(nameLen + 1); + memcpy(t, str, nameLen); + t[nameLen] = '\0'; + varObjPtr = Jim_NewStringObjNoAlloc(interp, t, nameLen); + + t = Jim_Alloc(keyLen + 1); + memcpy(t, p, keyLen); + t[keyLen] = '\0'; + keyObjPtr = Jim_NewStringObjNoAlloc(interp, t, keyLen); + + Jim_IncrRefCount(varObjPtr); + Jim_IncrRefCount(keyObjPtr); + *varPtrPtr = varObjPtr; + *keyPtrPtr = keyObjPtr; +} + +/* Helper of Jim_SetVariable() to deal with dict-syntax variable names. + * Also used by Jim_UnsetVariable() with valObjPtr = NULL. */ +static int JimDictSugarSet(Jim_Interp *interp, Jim_Obj *objPtr, + Jim_Obj *valObjPtr) +{ + Jim_Obj *varObjPtr, *keyObjPtr; + int err = JIM_OK; + + JimDictSugarParseVarKey(interp, objPtr, &varObjPtr, &keyObjPtr); + err = Jim_SetDictKeysVector(interp, varObjPtr, &keyObjPtr, 1, + valObjPtr); + Jim_DecrRefCount(interp, varObjPtr); + Jim_DecrRefCount(interp, keyObjPtr); + return err; +} + +/* Helper of Jim_GetVariable() to deal with dict-syntax variable names */ +static Jim_Obj *JimDictSugarGet(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_Obj *varObjPtr, *keyObjPtr, *dictObjPtr, *resObjPtr; + + JimDictSugarParseVarKey(interp, objPtr, &varObjPtr, &keyObjPtr); + dictObjPtr = Jim_GetVariable(interp, varObjPtr, JIM_ERRMSG); + if (!dictObjPtr) { + resObjPtr = NULL; + goto err; + } + if (Jim_DictKey(interp, dictObjPtr, keyObjPtr, &resObjPtr, JIM_ERRMSG) + != JIM_OK) { + resObjPtr = NULL; + } +err: + Jim_DecrRefCount(interp, varObjPtr); + Jim_DecrRefCount(interp, keyObjPtr); + return resObjPtr; +} + +/* --------- $var(INDEX) substitution, using a specialized object ----------- */ + +static void FreeDictSubstInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupDictSubstInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, + Jim_Obj *dupPtr); + +static Jim_ObjType dictSubstObjType = { + "dict-substitution", + FreeDictSubstInternalRep, + DupDictSubstInternalRep, + NULL, + JIM_TYPE_NONE, +}; + +void FreeDictSubstInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_DecrRefCount(interp, objPtr->internalRep.dictSubstValue.varNameObjPtr); + Jim_DecrRefCount(interp, objPtr->internalRep.dictSubstValue.indexObjPtr); +} + +void DupDictSubstInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, + Jim_Obj *dupPtr) +{ + JIM_NOTUSED(interp); + + dupPtr->internalRep.dictSubstValue.varNameObjPtr = + srcPtr->internalRep.dictSubstValue.varNameObjPtr; + dupPtr->internalRep.dictSubstValue.indexObjPtr = + srcPtr->internalRep.dictSubstValue.indexObjPtr; + dupPtr->typePtr = &dictSubstObjType; +} + +/* This function is used to expand [dict get] sugar in the form + * of $var(INDEX). The function is mainly used by Jim_EvalObj() + * to deal with tokens of type JIM_TT_DICTSUGAR. objPtr points to an + * object that is *guaranteed* to be in the form VARNAME(INDEX). + * The 'index' part is [subst]ituted, and is used to lookup a key inside + * the [dict]ionary contained in variable VARNAME. */ +Jim_Obj *Jim_ExpandDictSugar(Jim_Interp *interp, Jim_Obj *objPtr) +{ + Jim_Obj *varObjPtr, *keyObjPtr, *dictObjPtr, *resObjPtr; + Jim_Obj *substKeyObjPtr = NULL; + + if (objPtr->typePtr != &dictSubstObjType) { + JimDictSugarParseVarKey(interp, objPtr, &varObjPtr, &keyObjPtr); + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &dictSubstObjType; + objPtr->internalRep.dictSubstValue.varNameObjPtr = varObjPtr; + objPtr->internalRep.dictSubstValue.indexObjPtr = keyObjPtr; + } + if (Jim_SubstObj(interp, objPtr->internalRep.dictSubstValue.indexObjPtr, + &substKeyObjPtr, JIM_NONE) + != JIM_OK) { + substKeyObjPtr = NULL; + goto err; + } + Jim_IncrRefCount(substKeyObjPtr); + dictObjPtr = Jim_GetVariable(interp, + objPtr->internalRep.dictSubstValue.varNameObjPtr, JIM_ERRMSG); + if (!dictObjPtr) { + resObjPtr = NULL; + goto err; + } + if (Jim_DictKey(interp, dictObjPtr, substKeyObjPtr, &resObjPtr, JIM_ERRMSG) + != JIM_OK) { + resObjPtr = NULL; + goto err; + } +err: + if (substKeyObjPtr) Jim_DecrRefCount(interp, substKeyObjPtr); + return resObjPtr; +} + +/* ----------------------------------------------------------------------------- + * CallFrame + * ---------------------------------------------------------------------------*/ + +static Jim_CallFrame *JimCreateCallFrame(Jim_Interp *interp) +{ + Jim_CallFrame *cf; + if (interp->freeFramesList) { + cf = interp->freeFramesList; + interp->freeFramesList = cf->nextFramePtr; + } else { + cf = Jim_Alloc(sizeof(*cf)); + cf->vars.table = NULL; + } + + cf->id = interp->callFrameEpoch++; + cf->parentCallFrame = NULL; + cf->argv = NULL; + cf->argc = 0; + cf->procArgsObjPtr = NULL; + cf->procBodyObjPtr = NULL; + cf->nextFramePtr = NULL; + cf->staticVars = NULL; + if (cf->vars.table == NULL) + Jim_InitHashTable(&cf->vars, &JimVariablesHashTableType, interp); + return cf; +} + +/* Used to invalidate every caching related to callframe stability. */ +static void JimChangeCallFrameId(Jim_Interp *interp, Jim_CallFrame *cf) +{ + cf->id = interp->callFrameEpoch++; +} + +#define JIM_FCF_NONE 0 /* no flags */ +#define JIM_FCF_NOHT 1 /* don't free the hash table */ +static void JimFreeCallFrame(Jim_Interp *interp, Jim_CallFrame *cf, + int flags) +{ + if (cf->procArgsObjPtr) Jim_DecrRefCount(interp, cf->procArgsObjPtr); + if (cf->procBodyObjPtr) Jim_DecrRefCount(interp, cf->procBodyObjPtr); + if (!(flags & JIM_FCF_NOHT)) + Jim_FreeHashTable(&cf->vars); + else { + int i; + Jim_HashEntry **table = cf->vars.table, *he; + + for (i = 0; i < JIM_HT_INITIAL_SIZE; i++) { + he = table[i]; + while (he != NULL) { + Jim_HashEntry *nextEntry = he->next; + Jim_Var *varPtr = (void*) he->val; + + Jim_DecrRefCount(interp, varPtr->objPtr); + Jim_Free(he->val); + Jim_Free((void*)he->key); /* ATTENTION: const cast */ + Jim_Free(he); + table[i] = NULL; + he = nextEntry; + } + } + cf->vars.used = 0; + } + cf->nextFramePtr = interp->freeFramesList; + interp->freeFramesList = cf; +} + +/* ----------------------------------------------------------------------------- + * References + * ---------------------------------------------------------------------------*/ + +/* References HashTable Type. + * + * Keys are jim_wide integers, dynamically allocated for now but in the + * future it's worth to cache this 8 bytes objects. Values are poitners + * to Jim_References. */ +static void JimReferencesHTValDestructor(void *interp, void *val) +{ + Jim_Reference *refPtr = (void*) val; + + Jim_DecrRefCount(interp, refPtr->objPtr); + if (refPtr->finalizerCmdNamePtr != NULL) { + Jim_DecrRefCount(interp, refPtr->finalizerCmdNamePtr); + } + Jim_Free(val); +} + +unsigned int JimReferencesHTHashFunction(const void *key) +{ + /* Only the least significant bits are used. */ + const jim_wide *widePtr = key; + unsigned int intValue = (unsigned int) *widePtr; + return Jim_IntHashFunction(intValue); +} + +unsigned int JimReferencesHTDoubleHashFunction(const void *key) +{ + /* Only the least significant bits are used. */ + const jim_wide *widePtr = key; + unsigned int intValue = (unsigned int) *widePtr; + return intValue; /* identity function. */ +} + +const void *JimReferencesHTKeyDup(void *privdata, const void *key) +{ + void *copy = Jim_Alloc(sizeof(jim_wide)); + JIM_NOTUSED(privdata); + + memcpy(copy, key, sizeof(jim_wide)); + return copy; +} + +int JimReferencesHTKeyCompare(void *privdata, const void *key1, + const void *key2) +{ + JIM_NOTUSED(privdata); + + return memcmp(key1, key2, sizeof(jim_wide)) == 0; +} + +void JimReferencesHTKeyDestructor(void *privdata, const void *key) +{ + JIM_NOTUSED(privdata); + + Jim_Free((void*)key); +} + +static Jim_HashTableType JimReferencesHashTableType = { + JimReferencesHTHashFunction, /* hash function */ + JimReferencesHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimReferencesHTKeyCompare, /* key compare */ + JimReferencesHTKeyDestructor, /* key destructor */ + JimReferencesHTValDestructor /* val destructor */ +}; + +/* ----------------------------------------------------------------------------- + * Reference object type and References API + * ---------------------------------------------------------------------------*/ + +static void UpdateStringOfReference(struct Jim_Obj *objPtr); + +static Jim_ObjType referenceObjType = { + "reference", + NULL, + NULL, + UpdateStringOfReference, + JIM_TYPE_REFERENCES, +}; + +void UpdateStringOfReference(struct Jim_Obj *objPtr) +{ + int len; + char buf[JIM_REFERENCE_SPACE + 1]; + Jim_Reference *refPtr; + + refPtr = objPtr->internalRep.refValue.refPtr; + len = JimFormatReference(buf, refPtr, objPtr->internalRep.refValue.id); + objPtr->bytes = Jim_Alloc(len + 1); + memcpy(objPtr->bytes, buf, len + 1); + objPtr->length = len; +} + +/* returns true if 'c' is a valid reference tag character. + * i.e. inside the range [_a-zA-Z0-9] */ +static int isrefchar(int c) +{ + if (c == '_' || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || + (c >= '0' && c <= '9')) return 1; + return 0; +} + +int SetReferenceFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + jim_wide wideValue; + int i, len; + const char *str, *start, *end; + char refId[21]; + Jim_Reference *refPtr; + Jim_HashEntry *he; + + /* Get the string representation */ + str = Jim_GetString(objPtr, &len); + /* Check if it looks like a reference */ + if (len < JIM_REFERENCE_SPACE) goto badformat; + /* Trim spaces */ + start = str; + end = str + len-1; + while (*start == ' ') start++; + while (*end == ' ' && end > start) end--; + if (end-start + 1 != JIM_REFERENCE_SPACE) goto badformat; + /* .%020> */ + if (memcmp(start, "references, &wideValue); + if (he == NULL) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "Invalid reference ID \"", str, "\"", NULL); + return JIM_ERR; + } + refPtr = he->val; + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &referenceObjType; + objPtr->internalRep.refValue.id = wideValue; + objPtr->internalRep.refValue.refPtr = refPtr; + return JIM_OK; + +badformat: + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "expected reference but got \"", str, "\"", NULL); + return JIM_ERR; +} + +/* Returns a new reference pointing to objPtr, having cmdNamePtr + * as finalizer command (or NULL if there is no finalizer). + * The returned reference object has refcount = 0. */ +Jim_Obj *Jim_NewReference(Jim_Interp *interp, Jim_Obj *objPtr, Jim_Obj *tagPtr, + Jim_Obj *cmdNamePtr) +{ + struct Jim_Reference *refPtr; + jim_wide wideValue = interp->referenceNextId; + Jim_Obj *refObjPtr; + const char *tag; + int tagLen, i; + + /* Perform the Garbage Collection if needed. */ + Jim_CollectIfNeeded(interp); + + refPtr = Jim_Alloc(sizeof(*refPtr)); + refPtr->objPtr = objPtr; + Jim_IncrRefCount(objPtr); + refPtr->finalizerCmdNamePtr = cmdNamePtr; + if (cmdNamePtr) + Jim_IncrRefCount(cmdNamePtr); + Jim_AddHashEntry(&interp->references, &wideValue, refPtr); + refObjPtr = Jim_NewObj(interp); + refObjPtr->typePtr = &referenceObjType; + refObjPtr->bytes = NULL; + refObjPtr->internalRep.refValue.id = interp->referenceNextId; + refObjPtr->internalRep.refValue.refPtr = refPtr; + interp->referenceNextId++; + /* Set the tag. Trimmered at JIM_REFERENCE_TAGLEN. Everything + * that does not pass the 'isrefchar' test is replaced with '_' */ + tag = Jim_GetString(tagPtr, &tagLen); + if (tagLen > JIM_REFERENCE_TAGLEN) + tagLen = JIM_REFERENCE_TAGLEN; + for (i = 0; i < JIM_REFERENCE_TAGLEN; i++) { + if (i < tagLen) + refPtr->tag[i] = tag[i]; + else + refPtr->tag[i] = '_'; + } + refPtr->tag[JIM_REFERENCE_TAGLEN] = '\0'; + return refObjPtr; +} + +Jim_Reference *Jim_GetReference(Jim_Interp *interp, Jim_Obj *objPtr) +{ + if (objPtr->typePtr != &referenceObjType && + SetReferenceFromAny(interp, objPtr) == JIM_ERR) + return NULL; + return objPtr->internalRep.refValue.refPtr; +} + +int Jim_SetFinalizer(Jim_Interp *interp, Jim_Obj *objPtr, Jim_Obj *cmdNamePtr) +{ + Jim_Reference *refPtr; + + if ((refPtr = Jim_GetReference(interp, objPtr)) == NULL) + return JIM_ERR; + Jim_IncrRefCount(cmdNamePtr); + if (refPtr->finalizerCmdNamePtr) + Jim_DecrRefCount(interp, refPtr->finalizerCmdNamePtr); + refPtr->finalizerCmdNamePtr = cmdNamePtr; + return JIM_OK; +} + +int Jim_GetFinalizer(Jim_Interp *interp, Jim_Obj *objPtr, Jim_Obj **cmdNamePtrPtr) +{ + Jim_Reference *refPtr; + + if ((refPtr = Jim_GetReference(interp, objPtr)) == NULL) + return JIM_ERR; + *cmdNamePtrPtr = refPtr->finalizerCmdNamePtr; + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * References Garbage Collection + * ---------------------------------------------------------------------------*/ + +/* This the hash table type for the "MARK" phase of the GC */ +static Jim_HashTableType JimRefMarkHashTableType = { + JimReferencesHTHashFunction, /* hash function */ + JimReferencesHTKeyDup, /* key dup */ + NULL, /* val dup */ + JimReferencesHTKeyCompare, /* key compare */ + JimReferencesHTKeyDestructor, /* key destructor */ + NULL /* val destructor */ +}; + +/* #define JIM_DEBUG_GC 1 */ + +/* Performs the garbage collection. */ +int Jim_Collect(Jim_Interp *interp) +{ + Jim_HashTable marks; + Jim_HashTableIterator *htiter; + Jim_HashEntry *he; + Jim_Obj *objPtr; + int collected = 0; + + /* Avoid recursive calls */ + if (interp->lastCollectId == -1) { + /* Jim_Collect() already running. Return just now. */ + return 0; + } + interp->lastCollectId = -1; + + /* Mark all the references found into the 'mark' hash table. + * The references are searched in every live object that + * is of a type that can contain references. */ + Jim_InitHashTable(&marks, &JimRefMarkHashTableType, NULL); + objPtr = interp->liveList; + while (objPtr) { + if (objPtr->typePtr == NULL || + objPtr->typePtr->flags & JIM_TYPE_REFERENCES) { + const char *str, *p; + int len; + + /* If the object is of type reference, to get the + * Id is simple... */ + if (objPtr->typePtr == &referenceObjType) { + Jim_AddHashEntry(&marks, + &objPtr->internalRep.refValue.id, NULL); +#ifdef JIM_DEBUG_GC + Jim_fprintf(interp,interp->cookie_stdout, + "MARK (reference): %d refcount: %d" JIM_NL, + (int) objPtr->internalRep.refValue.id, + objPtr->refCount); +#endif + objPtr = objPtr->nextObjPtr; + continue; + } + /* Get the string repr of the object we want + * to scan for references. */ + p = str = Jim_GetString(objPtr, &len); + /* Skip objects too little to contain references. */ + if (len < JIM_REFERENCE_SPACE) { + objPtr = objPtr->nextObjPtr; + continue; + } + /* Extract references from the object string repr. */ + while (1) { + int i; + jim_wide id; + char buf[21]; + + if ((p = strstr(p, "cookie_stdout,"MARK: %d" JIM_NL, (int)id); +#endif + p += JIM_REFERENCE_SPACE; + } + } + objPtr = objPtr->nextObjPtr; + } + + /* Run the references hash table to destroy every reference that + * is not referenced outside (not present in the mark HT). */ + htiter = Jim_GetHashTableIterator(&interp->references); + while ((he = Jim_NextHashEntry(htiter)) != NULL) { + const jim_wide *refId; + Jim_Reference *refPtr; + + refId = he->key; + /* Check if in the mark phase we encountered + * this reference. */ + if (Jim_FindHashEntry(&marks, refId) == NULL) { +#ifdef JIM_DEBUG_GC + Jim_fprintf(interp,interp->cookie_stdout,"COLLECTING %d" JIM_NL, (int)*refId); +#endif + collected++; + /* Drop the reference, but call the + * finalizer first if registered. */ + refPtr = he->val; + if (refPtr->finalizerCmdNamePtr) { + char *refstr = Jim_Alloc(JIM_REFERENCE_SPACE + 1); + Jim_Obj *objv[3], *oldResult; + + JimFormatReference(refstr, refPtr, *refId); + + objv[0] = refPtr->finalizerCmdNamePtr; + objv[1] = Jim_NewStringObjNoAlloc(interp, + refstr, 32); + objv[2] = refPtr->objPtr; + Jim_IncrRefCount(objv[0]); + Jim_IncrRefCount(objv[1]); + Jim_IncrRefCount(objv[2]); + + /* Drop the reference itself */ + Jim_DeleteHashEntry(&interp->references, refId); + + /* Call the finalizer. Errors ignored. */ + oldResult = interp->result; + Jim_IncrRefCount(oldResult); + Jim_EvalObjVector(interp, 3, objv); + Jim_SetResult(interp, oldResult); + Jim_DecrRefCount(interp, oldResult); + + Jim_DecrRefCount(interp, objv[0]); + Jim_DecrRefCount(interp, objv[1]); + Jim_DecrRefCount(interp, objv[2]); + } else { + Jim_DeleteHashEntry(&interp->references, refId); + } + } + } + Jim_FreeHashTableIterator(htiter); + Jim_FreeHashTable(&marks); + interp->lastCollectId = interp->referenceNextId; + interp->lastCollectTime = time(NULL); + return collected; +} + +#define JIM_COLLECT_ID_PERIOD 5000 +#define JIM_COLLECT_TIME_PERIOD 300 + +void Jim_CollectIfNeeded(Jim_Interp *interp) +{ + jim_wide elapsedId; + int elapsedTime; + + elapsedId = interp->referenceNextId - interp->lastCollectId; + elapsedTime = time(NULL) - interp->lastCollectTime; + + + if (elapsedId > JIM_COLLECT_ID_PERIOD || + elapsedTime > JIM_COLLECT_TIME_PERIOD) { + Jim_Collect(interp); + } +} + +/* ----------------------------------------------------------------------------- + * Interpreter related functions + * ---------------------------------------------------------------------------*/ + +Jim_Interp *Jim_CreateInterp(void) +{ + Jim_Interp *i = Jim_Alloc(sizeof(*i)); + Jim_Obj *pathPtr; + + i->errorLine = 0; + i->errorFileName = Jim_StrDup(""); + i->numLevels = 0; + i->maxNestingDepth = JIM_MAX_NESTING_DEPTH; + i->returnCode = JIM_OK; + i->exitCode = 0; + i->procEpoch = 0; + i->callFrameEpoch = 0; + i->liveList = i->freeList = NULL; + i->scriptFileName = Jim_StrDup(""); + i->referenceNextId = 0; + i->lastCollectId = 0; + i->lastCollectTime = time(NULL); + i->freeFramesList = NULL; + i->prngState = NULL; + i->evalRetcodeLevel = -1; + i->cookie_stdin = stdin; + i->cookie_stdout = stdout; + i->cookie_stderr = stderr; + i->cb_fwrite = ((size_t (*)(const void *, size_t, size_t, void *))(fwrite)); + i->cb_fread = ((size_t (*)(void *, size_t, size_t, void *))(fread)); + i->cb_vfprintf = ((int (*)(void *, const char *fmt, va_list))(vfprintf)); + i->cb_fflush = ((int (*)(void *))(fflush)); + i->cb_fgets = ((char * (*)(char *, int, void *))(fgets)); + + /* Note that we can create objects only after the + * interpreter liveList and freeList pointers are + * initialized to NULL. */ + Jim_InitHashTable(&i->commands, &JimCommandsHashTableType, i); + Jim_InitHashTable(&i->references, &JimReferencesHashTableType, i); + Jim_InitHashTable(&i->sharedStrings, &JimSharedStringsHashTableType, + NULL); + Jim_InitHashTable(&i->stub, &JimStringCopyHashTableType, NULL); + Jim_InitHashTable(&i->assocData, &JimAssocDataHashTableType, i); + Jim_InitHashTable(&i->packages, &JimStringKeyValCopyHashTableType, NULL); + i->framePtr = i->topFramePtr = JimCreateCallFrame(i); + i->emptyObj = Jim_NewEmptyStringObj(i); + i->result = i->emptyObj; + i->stackTrace = Jim_NewListObj(i, NULL, 0); + i->unknown = Jim_NewStringObj(i, "unknown", -1); + i->unknown_called = 0; + Jim_IncrRefCount(i->emptyObj); + Jim_IncrRefCount(i->result); + Jim_IncrRefCount(i->stackTrace); + Jim_IncrRefCount(i->unknown); + + /* Initialize key variables every interpreter should contain */ + pathPtr = Jim_NewStringObj(i, "./", -1); + Jim_SetVariableStr(i, "jim_libpath", pathPtr); + Jim_SetVariableStrWithStr(i, "jim_interactive", "0"); + + /* Export the core API to extensions */ + JimRegisterCoreApi(i); + return i; +} + +/* This is the only function Jim exports directly without + * to use the STUB system. It is only used by embedders + * in order to get an interpreter with the Jim API pointers + * registered. */ +Jim_Interp *ExportedJimCreateInterp(void) +{ + return Jim_CreateInterp(); +} + +void Jim_FreeInterp(Jim_Interp *i) +{ + Jim_CallFrame *cf = i->framePtr, *prevcf, *nextcf; + Jim_Obj *objPtr, *nextObjPtr; + + Jim_DecrRefCount(i, i->emptyObj); + Jim_DecrRefCount(i, i->result); + Jim_DecrRefCount(i, i->stackTrace); + Jim_DecrRefCount(i, i->unknown); + Jim_Free((void*)i->errorFileName); + Jim_Free((void*)i->scriptFileName); + Jim_FreeHashTable(&i->commands); + Jim_FreeHashTable(&i->references); + Jim_FreeHashTable(&i->stub); + Jim_FreeHashTable(&i->assocData); + Jim_FreeHashTable(&i->packages); + Jim_Free(i->prngState); + /* Free the call frames list */ + while (cf) { + prevcf = cf->parentCallFrame; + JimFreeCallFrame(i, cf, JIM_FCF_NONE); + cf = prevcf; + } + /* Check that the live object list is empty, otherwise + * there is a memory leak. */ + if (i->liveList != NULL) { + objPtr = i->liveList; + + Jim_fprintf(i, i->cookie_stdout,JIM_NL "-------------------------------------" JIM_NL); + Jim_fprintf(i, i->cookie_stdout,"Objects still in the free list:" JIM_NL); + while (objPtr) { + const char *type = objPtr->typePtr ? + objPtr->typePtr->name : ""; + Jim_fprintf(i, i->cookie_stdout,"%p \"%-10s\": '%.20s' (refCount: %d)" JIM_NL, + objPtr, type, + objPtr->bytes ? objPtr->bytes + : "(null)", objPtr->refCount); + if (objPtr->typePtr == &sourceObjType) { + Jim_fprintf(i, i->cookie_stdout, "FILE %s LINE %d" JIM_NL, + objPtr->internalRep.sourceValue.fileName, + objPtr->internalRep.sourceValue.lineNumber); + } + objPtr = objPtr->nextObjPtr; + } + Jim_fprintf(i, i->cookie_stdout, "-------------------------------------" JIM_NL JIM_NL); + Jim_Panic(i,"Live list non empty freeing the interpreter! Leak?"); + } + /* Free all the freed objects. */ + objPtr = i->freeList; + while (objPtr) { + nextObjPtr = objPtr->nextObjPtr; + Jim_Free(objPtr); + objPtr = nextObjPtr; + } + /* Free cached CallFrame structures */ + cf = i->freeFramesList; + while (cf) { + nextcf = cf->nextFramePtr; + if (cf->vars.table != NULL) + Jim_Free(cf->vars.table); + Jim_Free(cf); + cf = nextcf; + } + /* Free the sharedString hash table. Make sure to free it + * after every other Jim_Object was freed. */ + Jim_FreeHashTable(&i->sharedStrings); + /* Free the interpreter structure. */ + Jim_Free(i); +} + +/* Store the call frame relative to the level represented by + * levelObjPtr into *framePtrPtr. If levelObjPtr == NULL, the + * level is assumed to be '1'. + * + * If a newLevelptr int pointer is specified, the function stores + * the absolute level integer value of the new target callframe into + * *newLevelPtr. (this is used to adjust interp->numLevels + * in the implementation of [uplevel], so that [info level] will + * return a correct information). + * + * This function accepts the 'level' argument in the form + * of the commands [uplevel] and [upvar]. + * + * For a function accepting a relative integer as level suitable + * for implementation of [info level ?level?] check the + * GetCallFrameByInteger() function. */ +int Jim_GetCallFrameByLevel(Jim_Interp *interp, Jim_Obj *levelObjPtr, + Jim_CallFrame **framePtrPtr, int *newLevelPtr) +{ + long level; + const char *str; + Jim_CallFrame *framePtr; + + if (newLevelPtr) *newLevelPtr = interp->numLevels; + if (levelObjPtr) { + str = Jim_GetString(levelObjPtr, NULL); + if (str[0] == '#') { + char *endptr; + /* speedup for the toplevel (level #0) */ + if (str[1] == '0' && str[2] == '\0') { + if (newLevelPtr) *newLevelPtr = 0; + *framePtrPtr = interp->topFramePtr; + return JIM_OK; + } + + level = strtol(str + 1, &endptr, 0); + if (str[1] == '\0' || endptr[0] != '\0' || level < 0) + goto badlevel; + /* An 'absolute' level is converted into the + * 'number of levels to go back' format. */ + level = interp->numLevels - level; + if (level < 0) goto badlevel; + } else { + if (Jim_GetLong(interp, levelObjPtr, &level) != JIM_OK || level < 0) + goto badlevel; + } + } else { + str = "1"; /* Needed to format the error message. */ + level = 1; + } + /* Lookup */ + framePtr = interp->framePtr; + if (newLevelPtr) *newLevelPtr = (*newLevelPtr)-level; + while (level--) { + framePtr = framePtr->parentCallFrame; + if (framePtr == NULL) goto badlevel; + } + *framePtrPtr = framePtr; + return JIM_OK; +badlevel: + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad level \"", str, "\"", NULL); + return JIM_ERR; +} + +/* Similar to Jim_GetCallFrameByLevel() but the level is specified + * as a relative integer like in the [info level ?level?] command. */ +static int JimGetCallFrameByInteger(Jim_Interp *interp, Jim_Obj *levelObjPtr, + Jim_CallFrame **framePtrPtr) +{ + jim_wide level; + jim_wide relLevel; /* level relative to the current one. */ + Jim_CallFrame *framePtr; + + if (Jim_GetWide(interp, levelObjPtr, &level) != JIM_OK) + goto badlevel; + if (level > 0) { + /* An 'absolute' level is converted into the + * 'number of levels to go back' format. */ + relLevel = interp->numLevels - level; + } else { + relLevel = -level; + } + /* Lookup */ + framePtr = interp->framePtr; + while (relLevel--) { + framePtr = framePtr->parentCallFrame; + if (framePtr == NULL) goto badlevel; + } + *framePtrPtr = framePtr; + return JIM_OK; +badlevel: + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad level \"", Jim_GetString(levelObjPtr, NULL), "\"", NULL); + return JIM_ERR; +} + +static void JimSetErrorFileName(Jim_Interp *interp, char *filename) +{ + Jim_Free((void*)interp->errorFileName); + interp->errorFileName = Jim_StrDup(filename); +} + +static void JimSetErrorLineNumber(Jim_Interp *interp, int linenr) +{ + interp->errorLine = linenr; +} + +static void JimResetStackTrace(Jim_Interp *interp) +{ + Jim_DecrRefCount(interp, interp->stackTrace); + interp->stackTrace = Jim_NewListObj(interp, NULL, 0); + Jim_IncrRefCount(interp->stackTrace); +} + +static void JimAppendStackTrace(Jim_Interp *interp, const char *procname, + const char *filename, int linenr) +{ + /* No need to add this dummy entry to the stack trace */ + if (strcmp(procname, "unknown") == 0) { + return; + } + + if (Jim_IsShared(interp->stackTrace)) { + interp->stackTrace = + Jim_DuplicateObj(interp, interp->stackTrace); + Jim_IncrRefCount(interp->stackTrace); + } + Jim_ListAppendElement(interp, interp->stackTrace, + Jim_NewStringObj(interp, procname, -1)); + Jim_ListAppendElement(interp, interp->stackTrace, + Jim_NewStringObj(interp, filename, -1)); + Jim_ListAppendElement(interp, interp->stackTrace, + Jim_NewIntObj(interp, linenr)); +} + +int Jim_SetAssocData(Jim_Interp *interp, const char *key, Jim_InterpDeleteProc *delProc, void *data) +{ + AssocDataValue *assocEntryPtr = (AssocDataValue *)Jim_Alloc(sizeof(AssocDataValue)); + assocEntryPtr->delProc = delProc; + assocEntryPtr->data = data; + return Jim_AddHashEntry(&interp->assocData, key, assocEntryPtr); +} + +void *Jim_GetAssocData(Jim_Interp *interp, const char *key) +{ + Jim_HashEntry *entryPtr = Jim_FindHashEntry(&interp->assocData, key); + if (entryPtr != NULL) { + AssocDataValue *assocEntryPtr = (AssocDataValue *)entryPtr->val; + return assocEntryPtr->data; + } + return NULL; +} + +int Jim_DeleteAssocData(Jim_Interp *interp, const char *key) +{ + return Jim_DeleteHashEntry(&interp->assocData, key); +} + +int Jim_GetExitCode(Jim_Interp *interp) { + return interp->exitCode; +} + +void *Jim_SetStdin(Jim_Interp *interp, void *fp) +{ + if (fp != NULL) interp->cookie_stdin = fp; + return interp->cookie_stdin; +} + +void *Jim_SetStdout(Jim_Interp *interp, void *fp) +{ + if (fp != NULL) interp->cookie_stdout = fp; + return interp->cookie_stdout; +} + +void *Jim_SetStderr(Jim_Interp *interp, void *fp) +{ + if (fp != NULL) interp->cookie_stderr = fp; + return interp->cookie_stderr; +} + +/* ----------------------------------------------------------------------------- + * Shared strings. + * Every interpreter has an hash table where to put shared dynamically + * allocate strings that are likely to be used a lot of times. + * For example, in the 'source' object type, there is a pointer to + * the filename associated with that object. Every script has a lot + * of this objects with the identical file name, so it is wise to share + * this info. + * + * The API is trivial: Jim_GetSharedString(interp, "foobar") + * returns the pointer to the shared string. Every time a reference + * to the string is no longer used, the user should call + * Jim_ReleaseSharedString(interp, stringPointer). Once no one is using + * a given string, it is removed from the hash table. + * ---------------------------------------------------------------------------*/ +const char *Jim_GetSharedString(Jim_Interp *interp, const char *str) +{ + Jim_HashEntry *he = Jim_FindHashEntry(&interp->sharedStrings, str); + + if (he == NULL) { + char *strCopy = Jim_StrDup(str); + + Jim_AddHashEntry(&interp->sharedStrings, strCopy, (void*)1); + return strCopy; + } else { + intptr_t refCount = (intptr_t) he->val; + + refCount++; + he->val = (void*) refCount; + return he->key; + } +} + +void Jim_ReleaseSharedString(Jim_Interp *interp, const char *str) +{ + intptr_t refCount; + Jim_HashEntry *he = Jim_FindHashEntry(&interp->sharedStrings, str); + + if (he == NULL) + Jim_Panic(interp,"Jim_ReleaseSharedString called with " + "unknown shared string '%s'", str); + refCount = (intptr_t) he->val; + refCount--; + if (refCount == 0) { + Jim_DeleteHashEntry(&interp->sharedStrings, str); + } else { + he->val = (void*) refCount; + } +} + +/* ----------------------------------------------------------------------------- + * Integer object + * ---------------------------------------------------------------------------*/ +#define JIM_INTEGER_SPACE 24 + +static void UpdateStringOfInt(struct Jim_Obj *objPtr); +static int SetIntFromAny(Jim_Interp *interp, Jim_Obj *objPtr, int flags); + +static Jim_ObjType intObjType = { + "int", + NULL, + NULL, + UpdateStringOfInt, + JIM_TYPE_NONE, +}; + +void UpdateStringOfInt(struct Jim_Obj *objPtr) +{ + int len; + char buf[JIM_INTEGER_SPACE + 1]; + + len = Jim_WideToString(buf, objPtr->internalRep.wideValue); + objPtr->bytes = Jim_Alloc(len + 1); + memcpy(objPtr->bytes, buf, len + 1); + objPtr->length = len; +} + +int SetIntFromAny(Jim_Interp *interp, Jim_Obj *objPtr, int flags) +{ + jim_wide wideValue; + const char *str; + + /* Get the string representation */ + str = Jim_GetString(objPtr, NULL); + /* Try to convert into a jim_wide */ + if (Jim_StringToWide(str, &wideValue, 0) != JIM_OK) { + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "expected integer but got \"", str, "\"", NULL); + } + return JIM_ERR; + } + if ((wideValue == JIM_WIDE_MIN || wideValue == JIM_WIDE_MAX) && + errno == ERANGE) { + Jim_SetResultString(interp, + "Integer value too big to be represented", -1); + return JIM_ERR; + } + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &intObjType; + objPtr->internalRep.wideValue = wideValue; + return JIM_OK; +} + +int Jim_GetWide(Jim_Interp *interp, Jim_Obj *objPtr, jim_wide *widePtr) +{ + if (objPtr->typePtr != &intObjType && + SetIntFromAny(interp, objPtr, JIM_ERRMSG) == JIM_ERR) + return JIM_ERR; + *widePtr = objPtr->internalRep.wideValue; + return JIM_OK; +} + +/* Get a wide but does not set an error if the format is bad. */ +static int JimGetWideNoErr(Jim_Interp *interp, Jim_Obj *objPtr, + jim_wide *widePtr) +{ + if (objPtr->typePtr != &intObjType && + SetIntFromAny(interp, objPtr, JIM_NONE) == JIM_ERR) + return JIM_ERR; + *widePtr = objPtr->internalRep.wideValue; + return JIM_OK; +} + +int Jim_GetLong(Jim_Interp *interp, Jim_Obj *objPtr, long *longPtr) +{ + jim_wide wideValue; + int retval; + + retval = Jim_GetWide(interp, objPtr, &wideValue); + if (retval == JIM_OK) { + *longPtr = (long) wideValue; + return JIM_OK; + } + return JIM_ERR; +} + +void Jim_SetWide(Jim_Interp *interp, Jim_Obj *objPtr, jim_wide wideValue) +{ + if (Jim_IsShared(objPtr)) + Jim_Panic(interp,"Jim_SetWide called with shared object"); + if (objPtr->typePtr != &intObjType) { + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &intObjType; + } + Jim_InvalidateStringRep(objPtr); + objPtr->internalRep.wideValue = wideValue; +} + +Jim_Obj *Jim_NewIntObj(Jim_Interp *interp, jim_wide wideValue) +{ + Jim_Obj *objPtr; + + objPtr = Jim_NewObj(interp); + objPtr->typePtr = &intObjType; + objPtr->bytes = NULL; + objPtr->internalRep.wideValue = wideValue; + return objPtr; +} + +/* ----------------------------------------------------------------------------- + * Double object + * ---------------------------------------------------------------------------*/ +#define JIM_DOUBLE_SPACE 30 + +static void UpdateStringOfDouble(struct Jim_Obj *objPtr); +static int SetDoubleFromAny(Jim_Interp *interp, Jim_Obj *objPtr); + +static Jim_ObjType doubleObjType = { + "double", + NULL, + NULL, + UpdateStringOfDouble, + JIM_TYPE_NONE, +}; + +void UpdateStringOfDouble(struct Jim_Obj *objPtr) +{ + int len; + char buf[JIM_DOUBLE_SPACE + 1]; + + len = Jim_DoubleToString(buf, objPtr->internalRep.doubleValue); + objPtr->bytes = Jim_Alloc(len + 1); + memcpy(objPtr->bytes, buf, len + 1); + objPtr->length = len; +} + +int SetDoubleFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + double doubleValue; + const char *str; + + /* Get the string representation */ + str = Jim_GetString(objPtr, NULL); + /* Try to convert into a double */ + if (Jim_StringToDouble(str, &doubleValue) != JIM_OK) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "expected number but got '", str, "'", NULL); + return JIM_ERR; + } + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &doubleObjType; + objPtr->internalRep.doubleValue = doubleValue; + return JIM_OK; +} + +int Jim_GetDouble(Jim_Interp *interp, Jim_Obj *objPtr, double *doublePtr) +{ + if (objPtr->typePtr != &doubleObjType && + SetDoubleFromAny(interp, objPtr) == JIM_ERR) + return JIM_ERR; + *doublePtr = objPtr->internalRep.doubleValue; + return JIM_OK; +} + +void Jim_SetDouble(Jim_Interp *interp, Jim_Obj *objPtr, double doubleValue) +{ + if (Jim_IsShared(objPtr)) + Jim_Panic(interp,"Jim_SetDouble called with shared object"); + if (objPtr->typePtr != &doubleObjType) { + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &doubleObjType; + } + Jim_InvalidateStringRep(objPtr); + objPtr->internalRep.doubleValue = doubleValue; +} + +Jim_Obj *Jim_NewDoubleObj(Jim_Interp *interp, double doubleValue) +{ + Jim_Obj *objPtr; + + objPtr = Jim_NewObj(interp); + objPtr->typePtr = &doubleObjType; + objPtr->bytes = NULL; + objPtr->internalRep.doubleValue = doubleValue; + return objPtr; +} + +/* ----------------------------------------------------------------------------- + * List object + * ---------------------------------------------------------------------------*/ +static void ListAppendElement(Jim_Obj *listPtr, Jim_Obj *objPtr); +static void FreeListInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupListInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static void UpdateStringOfList(struct Jim_Obj *objPtr); +static int SetListFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +/* Note that while the elements of the list may contain references, + * the list object itself can't. This basically means that the + * list object string representation as a whole can't contain references + * that are not presents in the single elements. */ +static Jim_ObjType listObjType = { + "list", + FreeListInternalRep, + DupListInternalRep, + UpdateStringOfList, + JIM_TYPE_NONE, +}; + +void FreeListInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + int i; + + for (i = 0; i < objPtr->internalRep.listValue.len; i++) { + Jim_DecrRefCount(interp, objPtr->internalRep.listValue.ele[i]); + } + Jim_Free(objPtr->internalRep.listValue.ele); +} + +void DupListInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + int i; + JIM_NOTUSED(interp); + + dupPtr->internalRep.listValue.len = srcPtr->internalRep.listValue.len; + dupPtr->internalRep.listValue.maxLen = srcPtr->internalRep.listValue.maxLen; + dupPtr->internalRep.listValue.ele = + Jim_Alloc(sizeof(Jim_Obj*)*srcPtr->internalRep.listValue.maxLen); + memcpy(dupPtr->internalRep.listValue.ele, srcPtr->internalRep.listValue.ele, + sizeof(Jim_Obj*)*srcPtr->internalRep.listValue.len); + for (i = 0; i < dupPtr->internalRep.listValue.len; i++) { + Jim_IncrRefCount(dupPtr->internalRep.listValue.ele[i]); + } + dupPtr->typePtr = &listObjType; +} + +/* The following function checks if a given string can be encoded + * into a list element without any kind of quoting, surrounded by braces, + * or using escapes to quote. */ +#define JIM_ELESTR_SIMPLE 0 +#define JIM_ELESTR_BRACE 1 +#define JIM_ELESTR_QUOTE 2 +static int ListElementQuotingType(const char *s, int len) +{ + int i, level, trySimple = 1; + + /* Try with the SIMPLE case */ + if (len == 0) return JIM_ELESTR_BRACE; + if (s[0] == '"' || s[0] == '{') { + trySimple = 0; + goto testbrace; + } + for (i = 0; i < len; i++) { + switch (s[i]) { + case ' ': + case '$': + case '"': + case '[': + case ']': + case ';': + case '\\': + case '\r': + case '\n': + case '\t': + case '\f': + case '\v': + trySimple = 0; + case '{': + case '}': + goto testbrace; + } + } + return JIM_ELESTR_SIMPLE; + +testbrace: + /* Test if it's possible to do with braces */ + if (s[len-1] == '\\' || + s[len-1] == ']') return JIM_ELESTR_QUOTE; + level = 0; + for (i = 0; i < len; i++) { + switch (s[i]) { + case '{': level++; break; + case '}': level--; + if (level < 0) return JIM_ELESTR_QUOTE; + break; + case '\\': + if (s[i + 1] == '\n') + return JIM_ELESTR_QUOTE; + else + if (s[i + 1] != '\0') i++; + break; + } + } + if (level == 0) { + if (!trySimple) return JIM_ELESTR_BRACE; + for (i = 0; i < len; i++) { + switch (s[i]) { + case ' ': + case '$': + case '"': + case '[': + case ']': + case ';': + case '\\': + case '\r': + case '\n': + case '\t': + case '\f': + case '\v': + return JIM_ELESTR_BRACE; + break; + } + } + return JIM_ELESTR_SIMPLE; + } + return JIM_ELESTR_QUOTE; +} + +/* Returns the malloc-ed representation of a string + * using backslash to quote special chars. */ +char *BackslashQuoteString(const char *s, int len, int *qlenPtr) +{ + char *q = Jim_Alloc(len*2 + 1), *p; + + p = q; + while (*s) { + switch (*s) { + case ' ': + case '$': + case '"': + case '[': + case ']': + case '{': + case '}': + case ';': + case '\\': + *p++ = '\\'; + *p++ = *s++; + break; + case '\n': *p++ = '\\'; *p++ = 'n'; s++; break; + case '\r': *p++ = '\\'; *p++ = 'r'; s++; break; + case '\t': *p++ = '\\'; *p++ = 't'; s++; break; + case '\f': *p++ = '\\'; *p++ = 'f'; s++; break; + case '\v': *p++ = '\\'; *p++ = 'v'; s++; break; + default: + *p++ = *s++; + break; + } + } + *p = '\0'; + *qlenPtr = p-q; + return q; +} + +void UpdateStringOfList(struct Jim_Obj *objPtr) +{ + int i, bufLen, realLength; + const char *strRep; + char *p; + int *quotingType; + Jim_Obj **ele = objPtr->internalRep.listValue.ele; + + /* (Over) Estimate the space needed. */ + quotingType = Jim_Alloc(sizeof(int)*objPtr->internalRep.listValue.len + 1); + bufLen = 0; + for (i = 0; i < objPtr->internalRep.listValue.len; i++) { + int len; + + strRep = Jim_GetString(ele[i], &len); + quotingType[i] = ListElementQuotingType(strRep, len); + switch (quotingType[i]) { + case JIM_ELESTR_SIMPLE: bufLen += len; break; + case JIM_ELESTR_BRACE: bufLen += len + 2; break; + case JIM_ELESTR_QUOTE: bufLen += len*2; break; + } + bufLen++; /* elements separator. */ + } + bufLen++; + + /* Generate the string rep. */ + p = objPtr->bytes = Jim_Alloc(bufLen + 1); + realLength = 0; + for (i = 0; i < objPtr->internalRep.listValue.len; i++) { + int len, qlen; + strRep = Jim_GetString(ele[i], &len); + char *q; + + switch (quotingType[i]) { + case JIM_ELESTR_SIMPLE: + memcpy(p, strRep, len); + p += len; + realLength += len; + break; + case JIM_ELESTR_BRACE: + *p++ = '{'; + memcpy(p, strRep, len); + p += len; + *p++ = '}'; + realLength += len + 2; + break; + case JIM_ELESTR_QUOTE: + q = BackslashQuoteString(strRep, len, &qlen); + memcpy(p, q, qlen); + Jim_Free(q); + p += qlen; + realLength += qlen; + break; + } + /* Add a separating space */ + if (i + 1 != objPtr->internalRep.listValue.len) { + *p++ = ' '; + realLength ++; + } + } + *p = '\0'; /* nul term. */ + objPtr->length = realLength; + Jim_Free(quotingType); +} + +int SetListFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr) +{ + struct JimParserCtx parser; + const char *str; + int strLen; + + /* Get the string representation */ + str = Jim_GetString(objPtr, &strLen); + + /* Free the old internal repr just now and initialize the + * new one just now. The string->list conversion can't fail. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &listObjType; + objPtr->internalRep.listValue.len = 0; + objPtr->internalRep.listValue.maxLen = 0; + objPtr->internalRep.listValue.ele = NULL; + + /* Convert into a list */ + JimParserInit(&parser, str, strLen, 1); + while (!JimParserEof(&parser)) { + char *token; + int tokenLen, type; + Jim_Obj *elementPtr; + + JimParseList(&parser); + if (JimParserTtype(&parser) != JIM_TT_STR && + JimParserTtype(&parser) != JIM_TT_ESC) + continue; + token = JimParserGetToken(&parser, &tokenLen, &type, NULL); + elementPtr = Jim_NewStringObjNoAlloc(interp, token, tokenLen); + ListAppendElement(objPtr, elementPtr); + } + return JIM_OK; +} + +Jim_Obj *Jim_NewListObj(Jim_Interp *interp, Jim_Obj *const *elements, + int len) +{ + Jim_Obj *objPtr; + int i; + + objPtr = Jim_NewObj(interp); + objPtr->typePtr = &listObjType; + objPtr->bytes = NULL; + objPtr->internalRep.listValue.ele = NULL; + objPtr->internalRep.listValue.len = 0; + objPtr->internalRep.listValue.maxLen = 0; + for (i = 0; i < len; i++) { + ListAppendElement(objPtr, elements[i]); + } + return objPtr; +} + +/* Return a vector of Jim_Obj with the elements of a Jim list, and the + * length of the vector. Note that the user of this function should make + * sure that the list object can't shimmer while the vector returned + * is in use, this vector is the one stored inside the internal representation + * of the list object. This function is not exported, extensions should + * always access to the List object elements using Jim_ListIndex(). */ +static void JimListGetElements(Jim_Interp *interp, Jim_Obj *listObj, int *argc, + Jim_Obj ***listVec) +{ + Jim_ListLength(interp, listObj, argc); + assert(listObj->typePtr == &listObjType); + *listVec = listObj->internalRep.listValue.ele; +} + +/* ListSortElements type values */ +enum {JIM_LSORT_ASCII, JIM_LSORT_NOCASE, JIM_LSORT_ASCII_DECR, + JIM_LSORT_NOCASE_DECR}; + +/* Sort the internal rep of a list. */ +static int ListSortString(Jim_Obj **lhsObj, Jim_Obj **rhsObj) +{ + return Jim_StringCompareObj(*lhsObj, *rhsObj, 0); +} + +static int ListSortStringDecr(Jim_Obj **lhsObj, Jim_Obj **rhsObj) +{ + return Jim_StringCompareObj(*lhsObj, *rhsObj, 0) * -1; +} + +static int ListSortStringNoCase(Jim_Obj **lhsObj, Jim_Obj **rhsObj) +{ + return Jim_StringCompareObj(*lhsObj, *rhsObj, 1); +} + +static int ListSortStringNoCaseDecr(Jim_Obj **lhsObj, Jim_Obj **rhsObj) +{ + return Jim_StringCompareObj(*lhsObj, *rhsObj, 1) * -1; +} + +/* Sort a list *in place*. MUST be called with non-shared objects. */ +static void ListSortElements(Jim_Interp *interp, Jim_Obj *listObjPtr, int type) +{ + typedef int (qsort_comparator)(const void *, const void *); + int (*fn)(Jim_Obj**, Jim_Obj**); + Jim_Obj **vector; + int len; + + if (Jim_IsShared(listObjPtr)) + Jim_Panic(interp,"Jim_ListSortElements called with shared object"); + if (listObjPtr->typePtr != &listObjType) + SetListFromAny(interp, listObjPtr); + + vector = listObjPtr->internalRep.listValue.ele; + len = listObjPtr->internalRep.listValue.len; + switch (type) { + case JIM_LSORT_ASCII: fn = ListSortString; break; + case JIM_LSORT_NOCASE: fn = ListSortStringNoCase; break; + case JIM_LSORT_ASCII_DECR: fn = ListSortStringDecr; break; + case JIM_LSORT_NOCASE_DECR: fn = ListSortStringNoCaseDecr; break; + default: + fn = NULL; /* avoid warning */ + Jim_Panic(interp,"ListSort called with invalid sort type"); + } + qsort(vector, len, sizeof(Jim_Obj *), (qsort_comparator *)fn); + Jim_InvalidateStringRep(listObjPtr); +} + +/* This is the low-level function to append an element to a list. + * The higher-level Jim_ListAppendElement() performs shared object + * check and invalidate the string repr. This version is used + * in the internals of the List Object and is not exported. + * + * NOTE: this function can be called only against objects + * with internal type of List. */ +void ListAppendElement(Jim_Obj *listPtr, Jim_Obj *objPtr) +{ + int requiredLen = listPtr->internalRep.listValue.len + 1; + + if (requiredLen > listPtr->internalRep.listValue.maxLen) { + int maxLen = requiredLen * 2; + + listPtr->internalRep.listValue.ele = + Jim_Realloc(listPtr->internalRep.listValue.ele, + sizeof(Jim_Obj*)*maxLen); + listPtr->internalRep.listValue.maxLen = maxLen; + } + listPtr->internalRep.listValue.ele[listPtr->internalRep.listValue.len] = + objPtr; + listPtr->internalRep.listValue.len ++; + Jim_IncrRefCount(objPtr); +} + +/* This is the low-level function to insert elements into a list. + * The higher-level Jim_ListInsertElements() performs shared object + * check and invalidate the string repr. This version is used + * in the internals of the List Object and is not exported. + * + * NOTE: this function can be called only against objects + * with internal type of List. */ +void ListInsertElements(Jim_Obj *listPtr, int index_t, int elemc, + Jim_Obj *const *elemVec) +{ + int currentLen = listPtr->internalRep.listValue.len; + int requiredLen = currentLen + elemc; + int i; + Jim_Obj **point; + + if (requiredLen > listPtr->internalRep.listValue.maxLen) { + int maxLen = requiredLen * 2; + + listPtr->internalRep.listValue.ele = + Jim_Realloc(listPtr->internalRep.listValue.ele, + sizeof(Jim_Obj*)*maxLen); + listPtr->internalRep.listValue.maxLen = maxLen; + } + point = listPtr->internalRep.listValue.ele + index_t; + memmove(point + elemc, point, (currentLen-index_t) * sizeof(Jim_Obj*)); + for (i = 0; i < elemc; ++i) { + point[i] = elemVec[i]; + Jim_IncrRefCount(point[i]); + } + listPtr->internalRep.listValue.len += elemc; +} + +/* Appends every element of appendListPtr into listPtr. + * Both have to be of the list type. */ +void ListAppendList(Jim_Obj *listPtr, Jim_Obj *appendListPtr) +{ + int i, oldLen = listPtr->internalRep.listValue.len; + int appendLen = appendListPtr->internalRep.listValue.len; + int requiredLen = oldLen + appendLen; + + if (requiredLen > listPtr->internalRep.listValue.maxLen) { + int maxLen = requiredLen * 2; + + listPtr->internalRep.listValue.ele = + Jim_Realloc(listPtr->internalRep.listValue.ele, + sizeof(Jim_Obj*)*maxLen); + listPtr->internalRep.listValue.maxLen = maxLen; + } + for (i = 0; i < appendLen; i++) { + Jim_Obj *objPtr = appendListPtr->internalRep.listValue.ele[i]; + listPtr->internalRep.listValue.ele[oldLen + i] = objPtr; + Jim_IncrRefCount(objPtr); + } + listPtr->internalRep.listValue.len += appendLen; +} + +void Jim_ListAppendElement(Jim_Interp *interp, Jim_Obj *listPtr, Jim_Obj *objPtr) +{ + if (Jim_IsShared(listPtr)) + Jim_Panic(interp,"Jim_ListAppendElement called with shared object"); + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + Jim_InvalidateStringRep(listPtr); + ListAppendElement(listPtr, objPtr); +} + +void Jim_ListAppendList(Jim_Interp *interp, Jim_Obj *listPtr, Jim_Obj *appendListPtr) +{ + if (Jim_IsShared(listPtr)) + Jim_Panic(interp,"Jim_ListAppendList called with shared object"); + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + Jim_InvalidateStringRep(listPtr); + ListAppendList(listPtr, appendListPtr); +} + +void Jim_ListLength(Jim_Interp *interp, Jim_Obj *listPtr, int *intPtr) +{ + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + *intPtr = listPtr->internalRep.listValue.len; +} + +void Jim_ListInsertElements(Jim_Interp *interp, Jim_Obj *listPtr, int index_t, + int objc, Jim_Obj *const *objVec) +{ + if (Jim_IsShared(listPtr)) + Jim_Panic(interp,"Jim_ListInsertElement called with shared object"); + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + if (index_t >= 0 && index_t > listPtr->internalRep.listValue.len) + index_t = listPtr->internalRep.listValue.len; + else if (index_t < 0) + index_t = 0; + Jim_InvalidateStringRep(listPtr); + ListInsertElements(listPtr, index_t, objc, objVec); +} + +int Jim_ListIndex(Jim_Interp *interp, Jim_Obj *listPtr, int index_t, + Jim_Obj **objPtrPtr, int flags) +{ + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + if ((index_t >= 0 && index_t >= listPtr->internalRep.listValue.len) || + (index_t < 0 && (-index_t-1) >= listPtr->internalRep.listValue.len)) { + if (flags & JIM_ERRMSG) { + Jim_SetResultString(interp, + "list index out of range", -1); + } + return JIM_ERR; + } + if (index_t < 0) + index_t = listPtr->internalRep.listValue.len + index_t; + *objPtrPtr = listPtr->internalRep.listValue.ele[index_t]; + return JIM_OK; +} + +static int ListSetIndex(Jim_Interp *interp, Jim_Obj *listPtr, int index_t, + Jim_Obj *newObjPtr, int flags) +{ + if (listPtr->typePtr != &listObjType) + SetListFromAny(interp, listPtr); + if ((index_t >= 0 && index_t >= listPtr->internalRep.listValue.len) || + (index_t < 0 && (-index_t-1) >= listPtr->internalRep.listValue.len)) { + if (flags & JIM_ERRMSG) { + Jim_SetResultString(interp, + "list index_t out of range", -1); + } + return JIM_ERR; + } + if (index_t < 0) + index_t = listPtr->internalRep.listValue.len + index_t; + Jim_DecrRefCount(interp, listPtr->internalRep.listValue.ele[index_t]); + listPtr->internalRep.listValue.ele[index_t] = newObjPtr; + Jim_IncrRefCount(newObjPtr); + return JIM_OK; +} + +/* Modify the list stored into the variable named 'varNamePtr' + * setting the element specified by the 'indexc' indexes objects in 'indexv', + * with the new element 'newObjptr'. */ +int Jim_SetListIndex(Jim_Interp *interp, Jim_Obj *varNamePtr, + Jim_Obj *const *indexv, int indexc, Jim_Obj *newObjPtr) +{ + Jim_Obj *varObjPtr, *objPtr, *listObjPtr; + int shared, i, index_t; + + varObjPtr = objPtr = Jim_GetVariable(interp, varNamePtr, JIM_ERRMSG); + if (objPtr == NULL) + return JIM_ERR; + if ((shared = Jim_IsShared(objPtr))) + varObjPtr = objPtr = Jim_DuplicateObj(interp, objPtr); + for (i = 0; i < indexc-1; i++) { + listObjPtr = objPtr; + if (Jim_GetIndex(interp, indexv[i], &index_t) != JIM_OK) + goto err; + if (Jim_ListIndex(interp, listObjPtr, index_t, &objPtr, + JIM_ERRMSG) != JIM_OK) { + goto err; + } + if (Jim_IsShared(objPtr)) { + objPtr = Jim_DuplicateObj(interp, objPtr); + ListSetIndex(interp, listObjPtr, index_t, objPtr, JIM_NONE); + } + Jim_InvalidateStringRep(listObjPtr); + } + if (Jim_GetIndex(interp, indexv[indexc-1], &index_t) != JIM_OK) + goto err; + if (ListSetIndex(interp, objPtr, index_t, newObjPtr, JIM_ERRMSG) == JIM_ERR) + goto err; + Jim_InvalidateStringRep(objPtr); + Jim_InvalidateStringRep(varObjPtr); + if (Jim_SetVariable(interp, varNamePtr, varObjPtr) != JIM_OK) + goto err; + Jim_SetResult(interp, varObjPtr); + return JIM_OK; +err: + if (shared) { + Jim_FreeNewObj(interp, varObjPtr); + } + return JIM_ERR; +} + +Jim_Obj *Jim_ConcatObj(Jim_Interp *interp, int objc, Jim_Obj *const *objv) +{ + int i; + + /* If all the objects in objv are lists without string rep. + * it's possible to return a list as result, that's the + * concatenation of all the lists. */ + for (i = 0; i < objc; i++) { + if (objv[i]->typePtr != &listObjType || objv[i]->bytes) + break; + } + if (i == objc) { + Jim_Obj *objPtr = Jim_NewListObj(interp, NULL, 0); + for (i = 0; i < objc; i++) + Jim_ListAppendList(interp, objPtr, objv[i]); + return objPtr; + } else { + /* Else... we have to glue strings together */ + int len = 0, objLen; + char *bytes, *p; + + /* Compute the length */ + for (i = 0; i < objc; i++) { + Jim_GetString(objv[i], &objLen); + len += objLen; + } + if (objc) len += objc-1; + /* Create the string rep, and a stinrg object holding it. */ + p = bytes = Jim_Alloc(len + 1); + for (i = 0; i < objc; i++) { + const char *s = Jim_GetString(objv[i], &objLen); + while (objLen && (*s == ' ' || *s == '\t' || *s == '\n')) + { + s++; objLen--; len--; + } + while (objLen && (s[objLen-1] == ' ' || + s[objLen-1] == '\n' || s[objLen-1] == '\t')) { + objLen--; len--; + } + memcpy(p, s, objLen); + p += objLen; + if (objLen && i + 1 != objc) { + *p++ = ' '; + } else if (i + 1 != objc) { + /* Drop the space calcuated for this + * element that is instead null. */ + len--; + } + } + *p = '\0'; + return Jim_NewStringObjNoAlloc(interp, bytes, len); + } +} + +/* Returns a list composed of the elements in the specified range. + * first and start are directly accepted as Jim_Objects and + * processed for the end?-index? case. */ +Jim_Obj *Jim_ListRange(Jim_Interp *interp, Jim_Obj *listObjPtr, Jim_Obj *firstObjPtr, Jim_Obj *lastObjPtr) +{ + int first, last; + int len, rangeLen; + + if (Jim_GetIndex(interp, firstObjPtr, &first) != JIM_OK || + Jim_GetIndex(interp, lastObjPtr, &last) != JIM_OK) + return NULL; + Jim_ListLength(interp, listObjPtr, &len); /* will convert into list */ + first = JimRelToAbsIndex(len, first); + last = JimRelToAbsIndex(len, last); + JimRelToAbsRange(len, first, last, &first, &last, &rangeLen); + return Jim_NewListObj(interp, + listObjPtr->internalRep.listValue.ele + first, rangeLen); +} + +/* ----------------------------------------------------------------------------- + * Dict object + * ---------------------------------------------------------------------------*/ +static void FreeDictInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupDictInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static void UpdateStringOfDict(struct Jim_Obj *objPtr); +static int SetDictFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +/* Dict HashTable Type. + * + * Keys and Values are Jim objects. */ + +unsigned int JimObjectHTHashFunction(const void *key) +{ + const char *str; + Jim_Obj *objPtr = (Jim_Obj*) key; + int len, h; + + str = Jim_GetString(objPtr, &len); + h = Jim_GenHashFunction((unsigned char*)str, len); + return h; +} + +int JimObjectHTKeyCompare(void *privdata, const void *key1, const void *key2) +{ + JIM_NOTUSED(privdata); + + return Jim_StringEqObj((Jim_Obj*)key1, (Jim_Obj*)key2, 0); +} + +static void JimObjectHTKeyValDestructor(void *interp, void *val) +{ + Jim_Obj *objPtr = val; + + Jim_DecrRefCount(interp, objPtr); +} + +static Jim_HashTableType JimDictHashTableType = { + JimObjectHTHashFunction, /* hash function */ + NULL, /* key dup */ + NULL, /* val dup */ + JimObjectHTKeyCompare, /* key compare */ + (void(*)(void*, const void*)) /* ATTENTION: const cast */ + JimObjectHTKeyValDestructor, /* key destructor */ + JimObjectHTKeyValDestructor /* val destructor */ +}; + +/* Note that while the elements of the dict may contain references, + * the list object itself can't. This basically means that the + * dict object string representation as a whole can't contain references + * that are not presents in the single elements. */ +static Jim_ObjType dictObjType = { + "dict", + FreeDictInternalRep, + DupDictInternalRep, + UpdateStringOfDict, + JIM_TYPE_NONE, +}; + +void FreeDictInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + JIM_NOTUSED(interp); + + Jim_FreeHashTable(objPtr->internalRep.ptr); + Jim_Free(objPtr->internalRep.ptr); +} + +void DupDictInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + Jim_HashTable *ht, *dupHt; + Jim_HashTableIterator *htiter; + Jim_HashEntry *he; + + /* Create a new hash table */ + ht = srcPtr->internalRep.ptr; + dupHt = Jim_Alloc(sizeof(*dupHt)); + Jim_InitHashTable(dupHt, &JimDictHashTableType, interp); + if (ht->size != 0) + Jim_ExpandHashTable(dupHt, ht->size); + /* Copy every element from the source to the dup hash table */ + htiter = Jim_GetHashTableIterator(ht); + while ((he = Jim_NextHashEntry(htiter)) != NULL) { + const Jim_Obj *keyObjPtr = he->key; + Jim_Obj *valObjPtr = he->val; + + Jim_IncrRefCount((Jim_Obj*)keyObjPtr); /* ATTENTION: const cast */ + Jim_IncrRefCount(valObjPtr); + Jim_AddHashEntry(dupHt, keyObjPtr, valObjPtr); + } + Jim_FreeHashTableIterator(htiter); + + dupPtr->internalRep.ptr = dupHt; + dupPtr->typePtr = &dictObjType; +} + +void UpdateStringOfDict(struct Jim_Obj *objPtr) +{ + int i, bufLen, realLength; + const char *strRep; + char *p; + int *quotingType, objc; + Jim_HashTable *ht; + Jim_HashTableIterator *htiter; + Jim_HashEntry *he; + Jim_Obj **objv; + + /* Trun the hash table into a flat vector of Jim_Objects. */ + ht = objPtr->internalRep.ptr; + objc = ht->used*2; + objv = Jim_Alloc(objc*sizeof(Jim_Obj*)); + htiter = Jim_GetHashTableIterator(ht); + i = 0; + while ((he = Jim_NextHashEntry(htiter)) != NULL) { + objv[i++] = (Jim_Obj*)he->key; /* ATTENTION: const cast */ + objv[i++] = he->val; + } + Jim_FreeHashTableIterator(htiter); + /* (Over) Estimate the space needed. */ + quotingType = Jim_Alloc(sizeof(int)*objc); + bufLen = 0; + for (i = 0; i < objc; i++) { + int len; + + strRep = Jim_GetString(objv[i], &len); + quotingType[i] = ListElementQuotingType(strRep, len); + switch (quotingType[i]) { + case JIM_ELESTR_SIMPLE: bufLen += len; break; + case JIM_ELESTR_BRACE: bufLen += len + 2; break; + case JIM_ELESTR_QUOTE: bufLen += len*2; break; + } + bufLen++; /* elements separator. */ + } + bufLen++; + + /* Generate the string rep. */ + p = objPtr->bytes = Jim_Alloc(bufLen + 1); + realLength = 0; + for (i = 0; i < objc; i++) { + int len, qlen; + strRep = Jim_GetString(objv[i], &len); + char *q; + + switch (quotingType[i]) { + case JIM_ELESTR_SIMPLE: + memcpy(p, strRep, len); + p += len; + realLength += len; + break; + case JIM_ELESTR_BRACE: + *p++ = '{'; + memcpy(p, strRep, len); + p += len; + *p++ = '}'; + realLength += len + 2; + break; + case JIM_ELESTR_QUOTE: + q = BackslashQuoteString(strRep, len, &qlen); + memcpy(p, q, qlen); + Jim_Free(q); + p += qlen; + realLength += qlen; + break; + } + /* Add a separating space */ + if (i + 1 != objc) { + *p++ = ' '; + realLength ++; + } + } + *p = '\0'; /* nul term. */ + objPtr->length = realLength; + Jim_Free(quotingType); + Jim_Free(objv); +} + +int SetDictFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr) +{ + struct JimParserCtx parser; + Jim_HashTable *ht; + Jim_Obj *objv[2]; + const char *str; + int i, strLen; + + /* Get the string representation */ + str = Jim_GetString(objPtr, &strLen); + + /* Free the old internal repr just now and initialize the + * new one just now. The string->list conversion can't fail. */ + Jim_FreeIntRep(interp, objPtr); + ht = Jim_Alloc(sizeof(*ht)); + Jim_InitHashTable(ht, &JimDictHashTableType, interp); + objPtr->typePtr = &dictObjType; + objPtr->internalRep.ptr = ht; + + /* Convert into a dict */ + JimParserInit(&parser, str, strLen, 1); + i = 0; + while (!JimParserEof(&parser)) { + char *token; + int tokenLen, type; + + JimParseList(&parser); + if (JimParserTtype(&parser) != JIM_TT_STR && + JimParserTtype(&parser) != JIM_TT_ESC) + continue; + token = JimParserGetToken(&parser, &tokenLen, &type, NULL); + objv[i++] = Jim_NewStringObjNoAlloc(interp, token, tokenLen); + if (i == 2) { + i = 0; + Jim_IncrRefCount(objv[0]); + Jim_IncrRefCount(objv[1]); + if (Jim_AddHashEntry(ht, objv[0], objv[1]) != JIM_OK) { + Jim_HashEntry *he; + he = Jim_FindHashEntry(ht, objv[0]); + Jim_DecrRefCount(interp, objv[0]); + /* ATTENTION: const cast */ + Jim_DecrRefCount(interp, (Jim_Obj*)he->val); + he->val = objv[1]; + } + } + } + if (i) { + Jim_FreeNewObj(interp, objv[0]); + objPtr->typePtr = NULL; + Jim_FreeHashTable(ht); + Jim_SetResultString(interp, "invalid dictionary value: must be a list with an even number of elements", -1); + return JIM_ERR; + } + return JIM_OK; +} + +/* Dict object API */ + +/* Add an element to a dict. objPtr must be of the "dict" type. + * The higer-level exported function is Jim_DictAddElement(). + * If an element with the specified key already exists, the value + * associated is replaced with the new one. + * + * if valueObjPtr == NULL, the key is instead removed if it exists. */ +static void DictAddElement(Jim_Interp *interp, Jim_Obj *objPtr, + Jim_Obj *keyObjPtr, Jim_Obj *valueObjPtr) +{ + Jim_HashTable *ht = objPtr->internalRep.ptr; + + if (valueObjPtr == NULL) { /* unset */ + Jim_DeleteHashEntry(ht, keyObjPtr); + return; + } + Jim_IncrRefCount(keyObjPtr); + Jim_IncrRefCount(valueObjPtr); + if (Jim_AddHashEntry(ht, keyObjPtr, valueObjPtr) != JIM_OK) { + Jim_HashEntry *he = Jim_FindHashEntry(ht, keyObjPtr); + Jim_DecrRefCount(interp, keyObjPtr); + /* ATTENTION: const cast */ + Jim_DecrRefCount(interp, (Jim_Obj*)he->val); + he->val = valueObjPtr; + } +} + +/* Add an element, higher-level interface for DictAddElement(). + * If valueObjPtr == NULL, the key is removed if it exists. */ +int Jim_DictAddElement(Jim_Interp *interp, Jim_Obj *objPtr, + Jim_Obj *keyObjPtr, Jim_Obj *valueObjPtr) +{ + if (Jim_IsShared(objPtr)) + Jim_Panic(interp,"Jim_DictAddElement called with shared object"); + if (objPtr->typePtr != &dictObjType) { + if (SetDictFromAny(interp, objPtr) != JIM_OK) + return JIM_ERR; + } + DictAddElement(interp, objPtr, keyObjPtr, valueObjPtr); + Jim_InvalidateStringRep(objPtr); + return JIM_OK; +} + +Jim_Obj *Jim_NewDictObj(Jim_Interp *interp, Jim_Obj *const *elements, int len) +{ + Jim_Obj *objPtr; + int i; + + if (len % 2) + Jim_Panic(interp,"Jim_NewDicObj() 'len' argument must be even"); + + objPtr = Jim_NewObj(interp); + objPtr->typePtr = &dictObjType; + objPtr->bytes = NULL; + objPtr->internalRep.ptr = Jim_Alloc(sizeof(Jim_HashTable)); + Jim_InitHashTable(objPtr->internalRep.ptr, &JimDictHashTableType, interp); + for (i = 0; i < len; i += 2) + DictAddElement(interp, objPtr, elements[i], elements[i + 1]); + return objPtr; +} + +/* Return the value associated to the specified dict key */ +int Jim_DictKey(Jim_Interp *interp, Jim_Obj *dictPtr, Jim_Obj *keyPtr, + Jim_Obj **objPtrPtr, int flags) +{ + Jim_HashEntry *he; + Jim_HashTable *ht; + + if (dictPtr->typePtr != &dictObjType) { + if (SetDictFromAny(interp, dictPtr) != JIM_OK) + return JIM_ERR; + } + ht = dictPtr->internalRep.ptr; + if ((he = Jim_FindHashEntry(ht, keyPtr)) == NULL) { + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "key \"", Jim_GetString(keyPtr, NULL), + "\" not found in dictionary", NULL); + } + return JIM_ERR; + } + *objPtrPtr = he->val; + return JIM_OK; +} + +/* Return the value associated to the specified dict keys */ +int Jim_DictKeysVector(Jim_Interp *interp, Jim_Obj *dictPtr, + Jim_Obj *const *keyv, int keyc, Jim_Obj **objPtrPtr, int flags) +{ + Jim_Obj *objPtr = NULL; + int i; + + if (keyc == 0) { + *objPtrPtr = dictPtr; + return JIM_OK; + } + + for (i = 0; i < keyc; i++) { + if (Jim_DictKey(interp, dictPtr, keyv[i], &objPtr, flags) + != JIM_OK) + return JIM_ERR; + dictPtr = objPtr; + } + *objPtrPtr = objPtr; + return JIM_OK; +} + +/* Modify the dict stored into the variable named 'varNamePtr' + * setting the element specified by the 'keyc' keys objects in 'keyv', + * with the new value of the element 'newObjPtr'. + * + * If newObjPtr == NULL the operation is to remove the given key + * from the dictionary. */ +int Jim_SetDictKeysVector(Jim_Interp *interp, Jim_Obj *varNamePtr, + Jim_Obj *const *keyv, int keyc, Jim_Obj *newObjPtr) +{ + Jim_Obj *varObjPtr, *objPtr, *dictObjPtr; + int shared, i; + + varObjPtr = objPtr = Jim_GetVariable(interp, varNamePtr, JIM_ERRMSG); + if (objPtr == NULL) { + if (newObjPtr == NULL) /* Cannot remove a key from non existing var */ + return JIM_ERR; + varObjPtr = objPtr = Jim_NewDictObj(interp, NULL, 0); + if (Jim_SetVariable(interp, varNamePtr, objPtr) != JIM_OK) { + Jim_FreeNewObj(interp, varObjPtr); + return JIM_ERR; + } + } + if ((shared = Jim_IsShared(objPtr))) + varObjPtr = objPtr = Jim_DuplicateObj(interp, objPtr); + for (i = 0; i < keyc-1; i++) { + dictObjPtr = objPtr; + + /* Check if it's a valid dictionary */ + if (dictObjPtr->typePtr != &dictObjType) { + if (SetDictFromAny(interp, dictObjPtr) != JIM_OK) + goto err; + } + /* Check if the given key exists. */ + Jim_InvalidateStringRep(dictObjPtr); + if (Jim_DictKey(interp, dictObjPtr, keyv[i], &objPtr, + newObjPtr ? JIM_NONE : JIM_ERRMSG) == JIM_OK) + { + /* This key exists at the current level. + * Make sure it's not shared!. */ + if (Jim_IsShared(objPtr)) { + objPtr = Jim_DuplicateObj(interp, objPtr); + DictAddElement(interp, dictObjPtr, keyv[i], objPtr); + } + } else { + /* Key not found. If it's an [unset] operation + * this is an error. Only the last key may not + * exist. */ + if (newObjPtr == NULL) + goto err; + /* Otherwise set an empty dictionary + * as key's value. */ + objPtr = Jim_NewDictObj(interp, NULL, 0); + DictAddElement(interp, dictObjPtr, keyv[i], objPtr); + } + } + if (Jim_DictAddElement(interp, objPtr, keyv[keyc-1], newObjPtr) + != JIM_OK) + goto err; + Jim_InvalidateStringRep(objPtr); + Jim_InvalidateStringRep(varObjPtr); + if (Jim_SetVariable(interp, varNamePtr, varObjPtr) != JIM_OK) + goto err; + Jim_SetResult(interp, varObjPtr); + return JIM_OK; +err: + if (shared) { + Jim_FreeNewObj(interp, varObjPtr); + } + return JIM_ERR; +} + +/* ----------------------------------------------------------------------------- + * Index object + * ---------------------------------------------------------------------------*/ +static void UpdateStringOfIndex(struct Jim_Obj *objPtr); +static int SetIndexFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType indexObjType = { + "index", + NULL, + NULL, + UpdateStringOfIndex, + JIM_TYPE_NONE, +}; + +void UpdateStringOfIndex(struct Jim_Obj *objPtr) +{ + int len; + char buf[JIM_INTEGER_SPACE + 1]; + + if (objPtr->internalRep.indexValue >= 0) + len = sprintf(buf, "%d", objPtr->internalRep.indexValue); + else if (objPtr->internalRep.indexValue == -1) + len = sprintf(buf, "end"); + else { + len = sprintf(buf, "end%d", objPtr->internalRep.indexValue + 1); + } + objPtr->bytes = Jim_Alloc(len + 1); + memcpy(objPtr->bytes, buf, len + 1); + objPtr->length = len; +} + +int SetIndexFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + int index_t, end = 0; + const char *str; + + /* Get the string representation */ + str = Jim_GetString(objPtr, NULL); + /* Try to convert into an index */ + if (!strcmp(str, "end")) { + index_t = 0; + end = 1; + } else { + if (!strncmp(str, "end-", 4)) { + str += 4; + end = 1; + } + if (Jim_StringToIndex(str, &index_t) != JIM_OK) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad index \"", Jim_GetString(objPtr, NULL), "\": " + "must be integer or end?-integer?", NULL); + return JIM_ERR; + } + } + if (end) { + if (index_t < 0) + index_t = INT_MAX; + else + index_t = -(index_t + 1); + } else if (index_t < 0) + index_t = -INT_MAX; + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &indexObjType; + objPtr->internalRep.indexValue = index_t; + return JIM_OK; +} + +int Jim_GetIndex(Jim_Interp *interp, Jim_Obj *objPtr, int *indexPtr) +{ + /* Avoid shimmering if the object is an integer. */ + if (objPtr->typePtr == &intObjType) { + jim_wide val = objPtr->internalRep.wideValue; + if (!(val < LONG_MIN) && !(val > LONG_MAX)) { + *indexPtr = (val < 0) ? -INT_MAX : (long)val;; + return JIM_OK; + } + } + if (objPtr->typePtr != &indexObjType && + SetIndexFromAny(interp, objPtr) == JIM_ERR) + return JIM_ERR; + *indexPtr = objPtr->internalRep.indexValue; + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * Return Code Object. + * ---------------------------------------------------------------------------*/ + +static int SetReturnCodeFromAny(Jim_Interp *interp, Jim_Obj *objPtr); + +static Jim_ObjType returnCodeObjType = { + "return-code", + NULL, + NULL, + NULL, + JIM_TYPE_NONE, +}; + +int SetReturnCodeFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + const char *str; + int strLen, returnCode; + jim_wide wideValue; + + /* Get the string representation */ + str = Jim_GetString(objPtr, &strLen); + /* Try to convert into an integer */ + if (JimGetWideNoErr(interp, objPtr, &wideValue) != JIM_ERR) + returnCode = (int) wideValue; + else if (!JimStringCompare(str, strLen, "ok", 2, JIM_NOCASE)) + returnCode = JIM_OK; + else if (!JimStringCompare(str, strLen, "error", 5, JIM_NOCASE)) + returnCode = JIM_ERR; + else if (!JimStringCompare(str, strLen, "return", 6, JIM_NOCASE)) + returnCode = JIM_RETURN; + else if (!JimStringCompare(str, strLen, "break", 5, JIM_NOCASE)) + returnCode = JIM_BREAK; + else if (!JimStringCompare(str, strLen, "continue", 8, JIM_NOCASE)) + returnCode = JIM_CONTINUE; + else if (!JimStringCompare(str, strLen, "eval", 4, JIM_NOCASE)) + returnCode = JIM_EVAL; + else if (!JimStringCompare(str, strLen, "exit", 4, JIM_NOCASE)) + returnCode = JIM_EXIT; + else { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "expected return code but got '", str, "'", + NULL); + return JIM_ERR; + } + /* Free the old internal repr and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + objPtr->typePtr = &returnCodeObjType; + objPtr->internalRep.returnCode = returnCode; + return JIM_OK; +} + +int Jim_GetReturnCode(Jim_Interp *interp, Jim_Obj *objPtr, int *intPtr) +{ + if (objPtr->typePtr != &returnCodeObjType && + SetReturnCodeFromAny(interp, objPtr) == JIM_ERR) + return JIM_ERR; + *intPtr = objPtr->internalRep.returnCode; + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * Expression Parsing + * ---------------------------------------------------------------------------*/ +static int JimParseExprOperator(struct JimParserCtx *pc); +static int JimParseExprNumber(struct JimParserCtx *pc); +static int JimParseExprIrrational(struct JimParserCtx *pc); + +/* Exrp's Stack machine operators opcodes. */ + +/* Binary operators (numbers) */ +#define JIM_EXPROP_BINARY_NUM_FIRST 0 /* first */ +#define JIM_EXPROP_MUL 0 +#define JIM_EXPROP_DIV 1 +#define JIM_EXPROP_MOD 2 +#define JIM_EXPROP_SUB 3 +#define JIM_EXPROP_ADD 4 +#define JIM_EXPROP_LSHIFT 5 +#define JIM_EXPROP_RSHIFT 6 +#define JIM_EXPROP_ROTL 7 +#define JIM_EXPROP_ROTR 8 +#define JIM_EXPROP_LT 9 +#define JIM_EXPROP_GT 10 +#define JIM_EXPROP_LTE 11 +#define JIM_EXPROP_GTE 12 +#define JIM_EXPROP_NUMEQ 13 +#define JIM_EXPROP_NUMNE 14 +#define JIM_EXPROP_BITAND 15 +#define JIM_EXPROP_BITXOR 16 +#define JIM_EXPROP_BITOR 17 +#define JIM_EXPROP_LOGICAND 18 +#define JIM_EXPROP_LOGICOR 19 +#define JIM_EXPROP_LOGICAND_LEFT 20 +#define JIM_EXPROP_LOGICOR_LEFT 21 +#define JIM_EXPROP_POW 22 +#define JIM_EXPROP_BINARY_NUM_LAST 22 /* last */ + +/* Binary operators (strings) */ +#define JIM_EXPROP_STREQ 23 +#define JIM_EXPROP_STRNE 24 + +/* Unary operators (numbers) */ +#define JIM_EXPROP_NOT 25 +#define JIM_EXPROP_BITNOT 26 +#define JIM_EXPROP_UNARYMINUS 27 +#define JIM_EXPROP_UNARYPLUS 28 +#define JIM_EXPROP_LOGICAND_RIGHT 29 +#define JIM_EXPROP_LOGICOR_RIGHT 30 + +/* Ternary operators */ +#define JIM_EXPROP_TERNARY 31 + +/* Operands */ +#define JIM_EXPROP_NUMBER 32 +#define JIM_EXPROP_COMMAND 33 +#define JIM_EXPROP_VARIABLE 34 +#define JIM_EXPROP_DICTSUGAR 35 +#define JIM_EXPROP_SUBST 36 +#define JIM_EXPROP_STRING 37 + +/* Operators table */ +typedef struct Jim_ExprOperator { + const char *name; + int precedence; + int arity; + int opcode; +} Jim_ExprOperator; + +/* name - precedence - arity - opcode */ +static struct Jim_ExprOperator Jim_ExprOperators[] = { + {"!", 300, 1, JIM_EXPROP_NOT}, + {"~", 300, 1, JIM_EXPROP_BITNOT}, + {"unarymin", 300, 1, JIM_EXPROP_UNARYMINUS}, + {"unaryplus", 300, 1, JIM_EXPROP_UNARYPLUS}, + + {"**", 250, 2, JIM_EXPROP_POW}, + + {"*", 200, 2, JIM_EXPROP_MUL}, + {"/", 200, 2, JIM_EXPROP_DIV}, + {"%", 200, 2, JIM_EXPROP_MOD}, + + {"-", 100, 2, JIM_EXPROP_SUB}, + {"+", 100, 2, JIM_EXPROP_ADD}, + + {"<<<", 90, 3, JIM_EXPROP_ROTL}, + {">>>", 90, 3, JIM_EXPROP_ROTR}, + {"<<", 90, 2, JIM_EXPROP_LSHIFT}, + {">>", 90, 2, JIM_EXPROP_RSHIFT}, + + {"<", 80, 2, JIM_EXPROP_LT}, + {">", 80, 2, JIM_EXPROP_GT}, + {"<=", 80, 2, JIM_EXPROP_LTE}, + {">=", 80, 2, JIM_EXPROP_GTE}, + + {"==", 70, 2, JIM_EXPROP_NUMEQ}, + {"!=", 70, 2, JIM_EXPROP_NUMNE}, + + {"eq", 60, 2, JIM_EXPROP_STREQ}, + {"ne", 60, 2, JIM_EXPROP_STRNE}, + + {"&", 50, 2, JIM_EXPROP_BITAND}, + {"^", 49, 2, JIM_EXPROP_BITXOR}, + {"|", 48, 2, JIM_EXPROP_BITOR}, + + {"&&", 10, 2, JIM_EXPROP_LOGICAND}, + {"||", 10, 2, JIM_EXPROP_LOGICOR}, + + {"?", 5, 3, JIM_EXPROP_TERNARY}, + /* private operators */ + {NULL, 10, 2, JIM_EXPROP_LOGICAND_LEFT}, + {NULL, 10, 1, JIM_EXPROP_LOGICAND_RIGHT}, + {NULL, 10, 2, JIM_EXPROP_LOGICOR_LEFT}, + {NULL, 10, 1, JIM_EXPROP_LOGICOR_RIGHT}, +}; + +#define JIM_EXPR_OPERATORS_NUM \ + (sizeof(Jim_ExprOperators)/sizeof(struct Jim_ExprOperator)) + +int JimParseExpression(struct JimParserCtx *pc) +{ + /* Discard spaces and quoted newline */ + while (*(pc->p) == ' ' || + *(pc->p) == '\t' || + *(pc->p) == '\r' || + *(pc->p) == '\n' || + (*(pc->p) == '\\' && *(pc->p + 1) == '\n')) { + pc->p++; pc->len--; + } + + if (pc->len == 0) { + pc->tstart = pc->tend = pc->p; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EOL; + pc->eof = 1; + return JIM_OK; + } + switch (*(pc->p)) { + case '(': + pc->tstart = pc->tend = pc->p; + pc->tline = pc->linenr; + pc->tt = JIM_TT_SUBEXPR_START; + pc->p++; pc->len--; + break; + case ')': + pc->tstart = pc->tend = pc->p; + pc->tline = pc->linenr; + pc->tt = JIM_TT_SUBEXPR_END; + pc->p++; pc->len--; + break; + case '[': + return JimParseCmd(pc); + break; + case '$': + if (JimParseVar(pc) == JIM_ERR) + return JimParseExprOperator(pc); + else + return JIM_OK; + break; + case '-': + if ((pc->tt == JIM_TT_NONE || pc->tt == JIM_TT_EXPR_OPERATOR) && + isdigit((int)*(pc->p + 1))) + return JimParseExprNumber(pc); + else + return JimParseExprOperator(pc); + break; + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': case '.': + return JimParseExprNumber(pc); + break; + case '"': + case '{': + /* Here it's possible to reuse the List String parsing. */ + pc->tt = JIM_TT_NONE; /* Make sure it's sensed as a new word. */ + return JimParseListStr(pc); + break; + case 'N': case 'I': + case 'n': case 'i': + if (JimParseExprIrrational(pc) == JIM_ERR) + return JimParseExprOperator(pc); + break; + default: + return JimParseExprOperator(pc); + break; + } + return JIM_OK; +} + +int JimParseExprNumber(struct JimParserCtx *pc) +{ + int allowdot = 1; + int allowhex = 0; + + pc->tstart = pc->p; + pc->tline = pc->linenr; + if (*pc->p == '-') { + pc->p++; pc->len--; + } + while (isdigit((int)*pc->p) + || (allowhex && isxdigit((int)*pc->p)) + || (allowdot && *pc->p == '.') + || (pc->p-pc->tstart == 1 && *pc->tstart == '0' && + (*pc->p == 'x' || *pc->p == 'X')) +) + { + if ((*pc->p == 'x') || (*pc->p == 'X')) { + allowhex = 1; + allowdot = 0; + } + if (*pc->p == '.') + allowdot = 0; + pc->p++; pc->len--; + if (!allowdot && *pc->p == 'e' && *(pc->p + 1) == '-') { + pc->p += 2; pc->len -= 2; + } + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_EXPR_NUMBER; + return JIM_OK; +} + +int JimParseExprIrrational(struct JimParserCtx *pc) +{ + const char *Tokens[] = {"NaN", "nan", "NAN", "Inf", "inf", "INF", NULL}; + const char **token; + for (token = Tokens; *token != NULL; token++) { + int len = strlen(*token); + if (strncmp(*token, pc->p, len) == 0) { + pc->tstart = pc->p; + pc->tend = pc->p + len - 1; + pc->p += len; pc->len -= len; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EXPR_NUMBER; + return JIM_OK; + } + } + return JIM_ERR; +} + +int JimParseExprOperator(struct JimParserCtx *pc) +{ + int i; + int bestIdx = -1, bestLen = 0; + + /* Try to get the longest match. */ + for (i = 0; i < (signed)JIM_EXPR_OPERATORS_NUM; i++) { + const char *opname; + int oplen; + + opname = Jim_ExprOperators[i].name; + if (opname == NULL) continue; + oplen = strlen(opname); + + if (strncmp(opname, pc->p, oplen) == 0 && oplen > bestLen) { + bestIdx = i; + bestLen = oplen; + } + } + if (bestIdx == -1) return JIM_ERR; + pc->tstart = pc->p; + pc->tend = pc->p + bestLen - 1; + pc->p += bestLen; pc->len -= bestLen; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EXPR_OPERATOR; + return JIM_OK; +} + +struct Jim_ExprOperator *JimExprOperatorInfo(const char *opname) +{ + int i; + for (i = 0; i < (signed)JIM_EXPR_OPERATORS_NUM; i++) + if (Jim_ExprOperators[i].name && + strcmp(opname, Jim_ExprOperators[i].name) == 0) + return &Jim_ExprOperators[i]; + return NULL; +} + +struct Jim_ExprOperator *JimExprOperatorInfoByOpcode(int opcode) +{ + int i; + for (i = 0; i < (signed)JIM_EXPR_OPERATORS_NUM; i++) + if (Jim_ExprOperators[i].opcode == opcode) + return &Jim_ExprOperators[i]; + return NULL; +} + +/* ----------------------------------------------------------------------------- + * Expression Object + * ---------------------------------------------------------------------------*/ +static void FreeExprInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupExprInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static int SetExprFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr); + +static Jim_ObjType exprObjType = { + "expression", + FreeExprInternalRep, + DupExprInternalRep, + NULL, + JIM_TYPE_REFERENCES, +}; + +/* Expr bytecode structure */ +typedef struct ExprByteCode { + int *opcode; /* Integer array of opcodes. */ + Jim_Obj **obj; /* Array of associated Jim Objects. */ + int len; /* Bytecode length */ + int inUse; /* Used for sharing. */ +} ExprByteCode; + +void FreeExprInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + int i; + ExprByteCode *expr = (void*) objPtr->internalRep.ptr; + + expr->inUse--; + if (expr->inUse != 0) return; + for (i = 0; i < expr->len; i++) + Jim_DecrRefCount(interp, expr->obj[i]); + Jim_Free(expr->opcode); + Jim_Free(expr->obj); + Jim_Free(expr); +} + +void DupExprInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + JIM_NOTUSED(interp); + JIM_NOTUSED(srcPtr); + + /* Just returns an simple string. */ + dupPtr->typePtr = NULL; +} + +/* Add a new instruction to an expression bytecode structure. */ +static void ExprObjAddInstr(Jim_Interp *interp, ExprByteCode *expr, + int opcode, char *str, int len) +{ + expr->opcode = Jim_Realloc(expr->opcode, sizeof(int)*(expr->len + 1)); + expr->obj = Jim_Realloc(expr->obj, sizeof(Jim_Obj*)*(expr->len + 1)); + expr->opcode[expr->len] = opcode; + expr->obj[expr->len] = Jim_NewStringObjNoAlloc(interp, str, len); + Jim_IncrRefCount(expr->obj[expr->len]); + expr->len++; +} + +/* Check if an expr program looks correct. */ +static int ExprCheckCorrectness(ExprByteCode *expr) +{ + int i; + int stacklen = 0; + + /* Try to check if there are stack underflows, + * and make sure at the end of the program there is + * a single result on the stack. */ + for (i = 0; i < expr->len; i++) { + switch (expr->opcode[i]) { + case JIM_EXPROP_NUMBER: + case JIM_EXPROP_STRING: + case JIM_EXPROP_SUBST: + case JIM_EXPROP_VARIABLE: + case JIM_EXPROP_DICTSUGAR: + case JIM_EXPROP_COMMAND: + stacklen++; + break; + case JIM_EXPROP_NOT: + case JIM_EXPROP_BITNOT: + case JIM_EXPROP_UNARYMINUS: + case JIM_EXPROP_UNARYPLUS: + /* Unary operations */ + if (stacklen < 1) return JIM_ERR; + break; + case JIM_EXPROP_ADD: + case JIM_EXPROP_SUB: + case JIM_EXPROP_MUL: + case JIM_EXPROP_DIV: + case JIM_EXPROP_MOD: + case JIM_EXPROP_LT: + case JIM_EXPROP_GT: + case JIM_EXPROP_LTE: + case JIM_EXPROP_GTE: + case JIM_EXPROP_ROTL: + case JIM_EXPROP_ROTR: + case JIM_EXPROP_LSHIFT: + case JIM_EXPROP_RSHIFT: + case JIM_EXPROP_NUMEQ: + case JIM_EXPROP_NUMNE: + case JIM_EXPROP_STREQ: + case JIM_EXPROP_STRNE: + case JIM_EXPROP_BITAND: + case JIM_EXPROP_BITXOR: + case JIM_EXPROP_BITOR: + case JIM_EXPROP_LOGICAND: + case JIM_EXPROP_LOGICOR: + case JIM_EXPROP_POW: + /* binary operations */ + if (stacklen < 2) return JIM_ERR; + stacklen--; + break; + default: + Jim_Panic(NULL,"Default opcode reached ExprCheckCorrectness"); + break; + } + } + if (stacklen != 1) return JIM_ERR; + return JIM_OK; +} + +static void ExprShareLiterals(Jim_Interp *interp, ExprByteCode *expr, + ScriptObj *topLevelScript) +{ + int i; + + return; + for (i = 0; i < expr->len; i++) { + Jim_Obj *foundObjPtr; + + if (expr->obj[i] == NULL) continue; + foundObjPtr = ScriptSearchLiteral(interp, topLevelScript, + NULL, expr->obj[i]); + if (foundObjPtr != NULL) { + Jim_IncrRefCount(foundObjPtr); + Jim_DecrRefCount(interp, expr->obj[i]); + expr->obj[i] = foundObjPtr; + } + } +} + +/* This procedure converts every occurrence of || and && opereators + * in lazy unary versions. + * + * a b || is converted into: + * + * a |L b |R + * + * a b && is converted into: + * + * a &L b &R + * + * "|L" checks if 'a' is true: + * 1) if it is true pushes 1 and skips istructions to reach + * the opcode just after |R. + * 2) if it is false does nothing. + * "|R" checks if 'b' is true: + * 1) if it is true pushes 1, otherwise pushes 0. + * + * "&L" checks if 'a' is true: + * 1) if it is true does nothing. + * 2) If it is false pushes 0 and skips istructions to reach + * the opcode just after &R + * "&R" checks if 'a' is true: + * if it is true pushes 1, otherwise pushes 0. + */ +static void ExprMakeLazy(Jim_Interp *interp, ExprByteCode *expr) +{ + while (1) { + int index_t = -1, leftindex, arity, i, offset; + Jim_ExprOperator *op; + + /* Search for || or && */ + for (i = 0; i < expr->len; i++) { + if (expr->opcode[i] == JIM_EXPROP_LOGICAND || + expr->opcode[i] == JIM_EXPROP_LOGICOR) { + index_t = i; + break; + } + } + if (index_t == -1) return; + /* Search for the end of the first operator */ + leftindex = index_t-1; + arity = 1; + while (arity) { + switch (expr->opcode[leftindex]) { + case JIM_EXPROP_NUMBER: + case JIM_EXPROP_COMMAND: + case JIM_EXPROP_VARIABLE: + case JIM_EXPROP_DICTSUGAR: + case JIM_EXPROP_SUBST: + case JIM_EXPROP_STRING: + break; + default: + op = JimExprOperatorInfoByOpcode(expr->opcode[leftindex]); + if (op == NULL) { + Jim_Panic(interp,"Default reached in ExprMakeLazy()"); + } + arity += op->arity; + break; + } + arity--; + leftindex--; + } + leftindex++; + expr->opcode = Jim_Realloc(expr->opcode, sizeof(int)*(expr->len + 2)); + expr->obj = Jim_Realloc(expr->obj, sizeof(Jim_Obj*)*(expr->len + 2)); + memmove(&expr->opcode[leftindex + 2], &expr->opcode[leftindex], + sizeof(int)*(expr->len-leftindex)); + memmove(&expr->obj[leftindex + 2], &expr->obj[leftindex], + sizeof(Jim_Obj*)*(expr->len-leftindex)); + expr->len += 2; + index_t += 2; + offset = (index_t-leftindex)-1; + Jim_DecrRefCount(interp, expr->obj[index_t]); + if (expr->opcode[index_t] == JIM_EXPROP_LOGICAND) { + expr->opcode[leftindex + 1] = JIM_EXPROP_LOGICAND_LEFT; + expr->opcode[index_t] = JIM_EXPROP_LOGICAND_RIGHT; + expr->obj[leftindex + 1] = Jim_NewStringObj(interp, "&L", -1); + expr->obj[index_t] = Jim_NewStringObj(interp, "&R", -1); + } else { + expr->opcode[leftindex + 1] = JIM_EXPROP_LOGICOR_LEFT; + expr->opcode[index_t] = JIM_EXPROP_LOGICOR_RIGHT; + expr->obj[leftindex + 1] = Jim_NewStringObj(interp, "|L", -1); + expr->obj[index_t] = Jim_NewStringObj(interp, "|R", -1); + } + expr->opcode[leftindex] = JIM_EXPROP_NUMBER; + expr->obj[leftindex] = Jim_NewIntObj(interp, offset); + Jim_IncrRefCount(expr->obj[index_t]); + Jim_IncrRefCount(expr->obj[leftindex]); + Jim_IncrRefCount(expr->obj[leftindex + 1]); + } +} + +/* This method takes the string representation of an expression + * and generates a program for the Expr's stack-based VM. */ +int SetExprFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr) +{ + int exprTextLen; + const char *exprText = Jim_GetString(objPtr, &exprTextLen); + struct JimParserCtx parser; + int i, shareLiterals; + ExprByteCode *expr = Jim_Alloc(sizeof(*expr)); + Jim_Stack stack; + Jim_ExprOperator *op; + + /* Perform literal sharing with the current procedure + * running only if this expression appears to be not generated + * at runtime. */ + shareLiterals = objPtr->typePtr == &sourceObjType; + + expr->opcode = NULL; + expr->obj = NULL; + expr->len = 0; + expr->inUse = 1; + + Jim_InitStack(&stack); + JimParserInit(&parser, exprText, exprTextLen, 1); + while (!JimParserEof(&parser)) { + char *token; + int len, type; + + if (JimParseExpression(&parser) != JIM_OK) { + Jim_SetResultString(interp, "Syntax error in expression", -1); + goto err; + } + token = JimParserGetToken(&parser, &len, &type, NULL); + if (type == JIM_TT_EOL) { + Jim_Free(token); + break; + } + switch (type) { + case JIM_TT_STR: + ExprObjAddInstr(interp, expr, JIM_EXPROP_STRING, token, len); + break; + case JIM_TT_ESC: + ExprObjAddInstr(interp, expr, JIM_EXPROP_SUBST, token, len); + break; + case JIM_TT_VAR: + ExprObjAddInstr(interp, expr, JIM_EXPROP_VARIABLE, token, len); + break; + case JIM_TT_DICTSUGAR: + ExprObjAddInstr(interp, expr, JIM_EXPROP_DICTSUGAR, token, len); + break; + case JIM_TT_CMD: + ExprObjAddInstr(interp, expr, JIM_EXPROP_COMMAND, token, len); + break; + case JIM_TT_EXPR_NUMBER: + ExprObjAddInstr(interp, expr, JIM_EXPROP_NUMBER, token, len); + break; + case JIM_TT_EXPR_OPERATOR: + op = JimExprOperatorInfo(token); + while (1) { + Jim_ExprOperator *stackTopOp; + + if (Jim_StackPeek(&stack) != NULL) { + stackTopOp = JimExprOperatorInfo(Jim_StackPeek(&stack)); + } else { + stackTopOp = NULL; + } + if (Jim_StackLen(&stack) && op->arity != 1 && + stackTopOp && stackTopOp->precedence >= op->precedence) + { + ExprObjAddInstr(interp, expr, stackTopOp->opcode, + Jim_StackPeek(&stack), -1); + Jim_StackPop(&stack); + } else { + break; + } + } + Jim_StackPush(&stack, token); + break; + case JIM_TT_SUBEXPR_START: + Jim_StackPush(&stack, Jim_StrDup("(")); + Jim_Free(token); + break; + case JIM_TT_SUBEXPR_END: + { + int found = 0; + while (Jim_StackLen(&stack)) { + char *opstr = Jim_StackPop(&stack); + if (!strcmp(opstr, "(")) { + Jim_Free(opstr); + found = 1; + break; + } + op = JimExprOperatorInfo(opstr); + ExprObjAddInstr(interp, expr, op->opcode, opstr, -1); + } + if (!found) { + Jim_SetResultString(interp, + "Unexpected close parenthesis", -1); + goto err; + } + } + Jim_Free(token); + break; + default: + Jim_Panic(interp,"Default reached in SetExprFromAny()"); + break; + } + } + while (Jim_StackLen(&stack)) { + char *opstr = Jim_StackPop(&stack); + op = JimExprOperatorInfo(opstr); + if (op == NULL && !strcmp(opstr, "(")) { + Jim_Free(opstr); + Jim_SetResultString(interp, "Missing close parenthesis", -1); + goto err; + } + ExprObjAddInstr(interp, expr, op->opcode, opstr, -1); + } + /* Check program correctness. */ + if (ExprCheckCorrectness(expr) != JIM_OK) { + Jim_SetResultString(interp, "Invalid expression", -1); + goto err; + } + + /* Free the stack used for the compilation. */ + Jim_FreeStackElements(&stack, Jim_Free); + Jim_FreeStack(&stack); + + /* Convert || and && operators in unary |L |R and &L &R for lazyness */ + ExprMakeLazy(interp, expr); + + /* Perform literal sharing */ + if (shareLiterals && interp->framePtr->procBodyObjPtr) { + Jim_Obj *bodyObjPtr = interp->framePtr->procBodyObjPtr; + if (bodyObjPtr->typePtr == &scriptObjType) { + ScriptObj *bodyScript = bodyObjPtr->internalRep.ptr; + ExprShareLiterals(interp, expr, bodyScript); + } + } + + /* Free the old internal rep and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + Jim_SetIntRepPtr(objPtr, expr); + objPtr->typePtr = &exprObjType; + return JIM_OK; + +err: /* we jump here on syntax/compile errors. */ + Jim_FreeStackElements(&stack, Jim_Free); + Jim_FreeStack(&stack); + Jim_Free(expr->opcode); + for (i = 0; i < expr->len; i++) { + Jim_DecrRefCount(interp,expr->obj[i]); + } + Jim_Free(expr->obj); + Jim_Free(expr); + return JIM_ERR; +} + +ExprByteCode *Jim_GetExpression(Jim_Interp *interp, Jim_Obj *objPtr) +{ + if (objPtr->typePtr != &exprObjType) { + if (SetExprFromAny(interp, objPtr) != JIM_OK) + return NULL; + } + return (ExprByteCode*) Jim_GetIntRepPtr(objPtr); +} + +/* ----------------------------------------------------------------------------- + * Expressions evaluation. + * Jim uses a specialized stack-based virtual machine for expressions, + * that takes advantage of the fact that expr's operators + * can't be redefined. + * + * Jim_EvalExpression() uses the bytecode compiled by + * SetExprFromAny() method of the "expression" object. + * + * On success a Tcl Object containing the result of the evaluation + * is stored into expResultPtrPtr (having refcount of 1), and JIM_OK is + * returned. + * On error the function returns a retcode != to JIM_OK and set a suitable + * error on the interp. + * ---------------------------------------------------------------------------*/ +#define JIM_EE_STATICSTACK_LEN 10 + +int Jim_EvalExpression(Jim_Interp *interp, Jim_Obj *exprObjPtr, + Jim_Obj **exprResultPtrPtr) +{ + ExprByteCode *expr; + Jim_Obj **stack, *staticStack[JIM_EE_STATICSTACK_LEN]; + int stacklen = 0, i, error = 0, errRetCode = JIM_ERR; + + Jim_IncrRefCount(exprObjPtr); + expr = Jim_GetExpression(interp, exprObjPtr); + if (!expr) { + Jim_DecrRefCount(interp, exprObjPtr); + return JIM_ERR; /* error in expression. */ + } + /* In order to avoid that the internal repr gets freed due to + * shimmering of the exprObjPtr's object, we make the internal rep + * shared. */ + expr->inUse++; + + /* The stack-based expr VM itself */ + + /* Stack allocation. Expr programs have the feature that + * a program of length N can't require a stack longer than + * N. */ + if (expr->len > JIM_EE_STATICSTACK_LEN) + stack = Jim_Alloc(sizeof(Jim_Obj*)*expr->len); + else + stack = staticStack; + + /* Execute every istruction */ + for (i = 0; i < expr->len; i++) { + Jim_Obj *A, *B, *objPtr; + jim_wide wA, wB, wC; + double dA, dB, dC; + const char *sA, *sB; + int Alen, Blen, retcode; + int opcode = expr->opcode[i]; + + if (opcode == JIM_EXPROP_NUMBER || opcode == JIM_EXPROP_STRING) { + stack[stacklen++] = expr->obj[i]; + Jim_IncrRefCount(expr->obj[i]); + } else if (opcode == JIM_EXPROP_VARIABLE) { + objPtr = Jim_GetVariable(interp, expr->obj[i], JIM_ERRMSG); + if (objPtr == NULL) { + error = 1; + goto err; + } + stack[stacklen++] = objPtr; + Jim_IncrRefCount(objPtr); + } else if (opcode == JIM_EXPROP_SUBST) { + if ((retcode = Jim_SubstObj(interp, expr->obj[i], + &objPtr, JIM_NONE)) != JIM_OK) + { + error = 1; + errRetCode = retcode; + goto err; + } + stack[stacklen++] = objPtr; + Jim_IncrRefCount(objPtr); + } else if (opcode == JIM_EXPROP_DICTSUGAR) { + objPtr = Jim_ExpandDictSugar(interp, expr->obj[i]); + if (objPtr == NULL) { + error = 1; + goto err; + } + stack[stacklen++] = objPtr; + Jim_IncrRefCount(objPtr); + } else if (opcode == JIM_EXPROP_COMMAND) { + if ((retcode = Jim_EvalObj(interp, expr->obj[i])) != JIM_OK) { + error = 1; + errRetCode = retcode; + goto err; + } + stack[stacklen++] = interp->result; + Jim_IncrRefCount(interp->result); + } else if (opcode >= JIM_EXPROP_BINARY_NUM_FIRST && + opcode <= JIM_EXPROP_BINARY_NUM_LAST) + { + /* Note that there isn't to increment the + * refcount of objects. the references are moved + * from stack to A and B. */ + B = stack[--stacklen]; + A = stack[--stacklen]; + + /* --- Integer --- */ + if ((A->typePtr == &doubleObjType && !A->bytes) || + (B->typePtr == &doubleObjType && !B->bytes) || + JimGetWideNoErr(interp, A, &wA) != JIM_OK || + JimGetWideNoErr(interp, B, &wB) != JIM_OK) { + goto trydouble; + } + Jim_DecrRefCount(interp, A); + Jim_DecrRefCount(interp, B); + switch (expr->opcode[i]) { + case JIM_EXPROP_ADD: wC = wA + wB; break; + case JIM_EXPROP_SUB: wC = wA-wB; break; + case JIM_EXPROP_MUL: wC = wA*wB; break; + case JIM_EXPROP_LT: wC = wA < wB; break; + case JIM_EXPROP_GT: wC = wA > wB; break; + case JIM_EXPROP_LTE: wC = wA <= wB; break; + case JIM_EXPROP_GTE: wC = wA >= wB; break; + case JIM_EXPROP_LSHIFT: wC = wA << wB; break; + case JIM_EXPROP_RSHIFT: wC = wA >> wB; break; + case JIM_EXPROP_NUMEQ: wC = wA == wB; break; + case JIM_EXPROP_NUMNE: wC = wA != wB; break; + case JIM_EXPROP_BITAND: wC = wA&wB; break; + case JIM_EXPROP_BITXOR: wC = wA^wB; break; + case JIM_EXPROP_BITOR: wC = wA | wB; break; + case JIM_EXPROP_POW: wC = JimPowWide(wA,wB); break; + case JIM_EXPROP_LOGICAND_LEFT: + if (wA == 0) { + i += (int)wB; + wC = 0; + } else { + continue; + } + break; + case JIM_EXPROP_LOGICOR_LEFT: + if (wA != 0) { + i += (int)wB; + wC = 1; + } else { + continue; + } + break; + case JIM_EXPROP_DIV: + if (wB == 0) goto divbyzero; + wC = wA/wB; + break; + case JIM_EXPROP_MOD: + if (wB == 0) goto divbyzero; + wC = wA%wB; + break; + case JIM_EXPROP_ROTL: { + /* uint32_t would be better. But not everyone has inttypes.h?*/ + unsigned long uA = (unsigned long)wA; +#ifdef _MSC_VER + wC = _rotl(uA,(unsigned long)wB); +#else + const unsigned int S = sizeof(unsigned long) * 8; + wC = (unsigned long)((uA << wB) | (uA >> (S-wB))); +#endif + break; + } + case JIM_EXPROP_ROTR: { + unsigned long uA = (unsigned long)wA; +#ifdef _MSC_VER + wC = _rotr(uA,(unsigned long)wB); +#else + const unsigned int S = sizeof(unsigned long) * 8; + wC = (unsigned long)((uA >> wB) | (uA << (S-wB))); +#endif + break; + } + + default: + wC = 0; /* avoid gcc warning */ + break; + } + stack[stacklen] = Jim_NewIntObj(interp, wC); + Jim_IncrRefCount(stack[stacklen]); + stacklen++; + continue; +trydouble: + /* --- Double --- */ + if (Jim_GetDouble(interp, A, &dA) != JIM_OK || + Jim_GetDouble(interp, B, &dB) != JIM_OK) { + + /* Hmmm! For compatibility, maybe convert != and == into ne and eq */ + if (expr->opcode[i] == JIM_EXPROP_NUMNE) { + opcode = JIM_EXPROP_STRNE; + goto retry_as_string; + } + else if (expr->opcode[i] == JIM_EXPROP_NUMEQ) { + opcode = JIM_EXPROP_STREQ; + goto retry_as_string; + } + Jim_DecrRefCount(interp, A); + Jim_DecrRefCount(interp, B); + error = 1; + goto err; + } + Jim_DecrRefCount(interp, A); + Jim_DecrRefCount(interp, B); + switch (expr->opcode[i]) { + case JIM_EXPROP_ROTL: + case JIM_EXPROP_ROTR: + case JIM_EXPROP_LSHIFT: + case JIM_EXPROP_RSHIFT: + case JIM_EXPROP_BITAND: + case JIM_EXPROP_BITXOR: + case JIM_EXPROP_BITOR: + case JIM_EXPROP_MOD: + case JIM_EXPROP_POW: + Jim_SetResultString(interp, + "Got floating-point value where integer was expected", -1); + error = 1; + goto err; + case JIM_EXPROP_ADD: dC = dA + dB; break; + case JIM_EXPROP_SUB: dC = dA-dB; break; + case JIM_EXPROP_MUL: dC = dA*dB; break; + case JIM_EXPROP_LT: dC = dA < dB; break; + case JIM_EXPROP_GT: dC = dA > dB; break; + case JIM_EXPROP_LTE: dC = dA <= dB; break; + case JIM_EXPROP_GTE: dC = dA >= dB; break; + /* FIXME comparing floats for equality/inequality is bad juju */ + case JIM_EXPROP_NUMEQ: dC = dA == dB; break; + case JIM_EXPROP_NUMNE: dC = dA != dB; break; + case JIM_EXPROP_LOGICAND_LEFT: + if (dA == 0) { + i += (int)dB; + dC = 0; + } else { + continue; + } + break; + case JIM_EXPROP_LOGICOR_LEFT: + if (dA != 0) { + i += (int)dB; + dC = 1; + } else { + continue; + } + break; + case JIM_EXPROP_DIV: + if (dB == 0) goto divbyzero; + dC = dA/dB; + break; + default: + dC = 0; /* avoid gcc warning */ + break; + } + stack[stacklen] = Jim_NewDoubleObj(interp, dC); + Jim_IncrRefCount(stack[stacklen]); + stacklen++; + } else if (opcode == JIM_EXPROP_STREQ || opcode == JIM_EXPROP_STRNE) { + B = stack[--stacklen]; + A = stack[--stacklen]; +retry_as_string: + sA = Jim_GetString(A, &Alen); + sB = Jim_GetString(B, &Blen); + switch (opcode) { + case JIM_EXPROP_STREQ: + if (Alen == Blen && memcmp(sA, sB, Alen) ==0) + wC = 1; + else + wC = 0; + break; + case JIM_EXPROP_STRNE: + if (Alen != Blen || memcmp(sA, sB, Alen) != 0) + wC = 1; + else + wC = 0; + break; + default: + wC = 0; /* avoid gcc warning */ + break; + } + Jim_DecrRefCount(interp, A); + Jim_DecrRefCount(interp, B); + stack[stacklen] = Jim_NewIntObj(interp, wC); + Jim_IncrRefCount(stack[stacklen]); + stacklen++; + } else if (opcode == JIM_EXPROP_NOT || + opcode == JIM_EXPROP_BITNOT || + opcode == JIM_EXPROP_LOGICAND_RIGHT || + opcode == JIM_EXPROP_LOGICOR_RIGHT) { + /* Note that there isn't to increment the + * refcount of objects. the references are moved + * from stack to A and B. */ + A = stack[--stacklen]; + + /* --- Integer --- */ + if ((A->typePtr == &doubleObjType && !A->bytes) || + JimGetWideNoErr(interp, A, &wA) != JIM_OK) { + goto trydouble_unary; + } + Jim_DecrRefCount(interp, A); + switch (expr->opcode[i]) { + case JIM_EXPROP_NOT: wC = !wA; break; + case JIM_EXPROP_BITNOT: wC = ~wA; break; + case JIM_EXPROP_LOGICAND_RIGHT: + case JIM_EXPROP_LOGICOR_RIGHT: wC = (wA != 0); break; + default: + wC = 0; /* avoid gcc warning */ + break; + } + stack[stacklen] = Jim_NewIntObj(interp, wC); + Jim_IncrRefCount(stack[stacklen]); + stacklen++; + continue; +trydouble_unary: + /* --- Double --- */ + if (Jim_GetDouble(interp, A, &dA) != JIM_OK) { + Jim_DecrRefCount(interp, A); + error = 1; + goto err; + } + Jim_DecrRefCount(interp, A); + switch (expr->opcode[i]) { + case JIM_EXPROP_NOT: dC = !dA; break; + case JIM_EXPROP_LOGICAND_RIGHT: + case JIM_EXPROP_LOGICOR_RIGHT: dC = (dA != 0); break; + case JIM_EXPROP_BITNOT: + Jim_SetResultString(interp, + "Got floating-point value where integer was expected", -1); + error = 1; + goto err; + break; + default: + dC = 0; /* avoid gcc warning */ + break; + } + stack[stacklen] = Jim_NewDoubleObj(interp, dC); + Jim_IncrRefCount(stack[stacklen]); + stacklen++; + } else { + Jim_Panic(interp,"Unknown opcode in Jim_EvalExpression"); + } + } +err: + /* There is no need to decerement the inUse field because + * this reference is transfered back into the exprObjPtr. */ + Jim_FreeIntRep(interp, exprObjPtr); + exprObjPtr->typePtr = &exprObjType; + Jim_SetIntRepPtr(exprObjPtr, expr); + Jim_DecrRefCount(interp, exprObjPtr); + if (!error) { + *exprResultPtrPtr = stack[0]; + Jim_IncrRefCount(stack[0]); + errRetCode = JIM_OK; + } + for (i = 0; i < stacklen; i++) { + Jim_DecrRefCount(interp, stack[i]); + } + if (stack != staticStack) + Jim_Free(stack); + return errRetCode; +divbyzero: + error = 1; + Jim_SetResultString(interp, "Division by zero", -1); + goto err; +} + +int Jim_GetBoolFromExpr(Jim_Interp *interp, Jim_Obj *exprObjPtr, int *boolPtr) +{ + int retcode; + jim_wide wideValue; + double doubleValue; + Jim_Obj *exprResultPtr; + + retcode = Jim_EvalExpression(interp, exprObjPtr, &exprResultPtr); + if (retcode != JIM_OK) + return retcode; + if (JimGetWideNoErr(interp, exprResultPtr, &wideValue) != JIM_OK) { + if (Jim_GetDouble(interp, exprResultPtr, &doubleValue) != JIM_OK) + { + Jim_DecrRefCount(interp, exprResultPtr); + return JIM_ERR; + } else { + Jim_DecrRefCount(interp, exprResultPtr); + *boolPtr = doubleValue != 0; + return JIM_OK; + } + } + Jim_DecrRefCount(interp, exprResultPtr); + *boolPtr = wideValue != 0; + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * ScanFormat String Object + * ---------------------------------------------------------------------------*/ + +/* This Jim_Obj will held a parsed representation of a format string passed to + * the Jim_ScanString command. For error diagnostics, the scanformat string has + * to be parsed in its entirely first and then, if correct, can be used for + * scanning. To avoid endless re-parsing, the parsed representation will be + * stored in an internal representation and re-used for performance reason. */ + +/* A ScanFmtPartDescr will held the information of /one/ part of the whole + * scanformat string. This part will later be used to extract information + * out from the string to be parsed by Jim_ScanString */ + +typedef struct ScanFmtPartDescr { + char type; /* Type of conversion (e.g. c, d, f) */ + char modifier; /* Modify type (e.g. l - long, h - short */ + size_t width; /* Maximal width of input to be converted */ + int pos; /* -1 - no assign, 0 - natural pos, >0 - XPG3 pos */ + char *arg; /* Specification of a CHARSET conversion */ + char *prefix; /* Prefix to be scanned literally before conversion */ +} ScanFmtPartDescr; + +/* The ScanFmtStringObj will held the internal representation of a scanformat + * string parsed and separated in part descriptions. Furthermore it contains + * the original string representation of the scanformat string to allow for + * fast update of the Jim_Obj's string representation part. + * + * As add-on the internal object representation add some scratch pad area + * for usage by Jim_ScanString to avoid endless allocating and freeing of + * memory for purpose of string scanning. + * + * The error member points to a static allocated string in case of a mal- + * formed scanformat string or it contains '0' (NULL) in case of a valid + * parse representation. + * + * The whole memory of the internal representation is allocated as a single + * area of memory that will be internally separated. So freeing and duplicating + * of such an object is cheap */ + +typedef struct ScanFmtStringObj { + jim_wide size; /* Size of internal repr in bytes */ + char *stringRep; /* Original string representation */ + size_t count; /* Number of ScanFmtPartDescr contained */ + size_t convCount; /* Number of conversions that will assign */ + size_t maxPos; /* Max position index if XPG3 is used */ + const char *error; /* Ptr to error text (NULL if no error */ + char *scratch; /* Some scratch pad used by Jim_ScanString */ + ScanFmtPartDescr descr[1]; /* The vector of partial descriptions */ +} ScanFmtStringObj; + + +static void FreeScanFmtInternalRep(Jim_Interp *interp, Jim_Obj *objPtr); +static void DupScanFmtInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr); +static void UpdateStringOfScanFmt(Jim_Obj *objPtr); + +static Jim_ObjType scanFmtStringObjType = { + "scanformatstring", + FreeScanFmtInternalRep, + DupScanFmtInternalRep, + UpdateStringOfScanFmt, + JIM_TYPE_NONE, +}; + +void FreeScanFmtInternalRep(Jim_Interp *interp, Jim_Obj *objPtr) +{ + JIM_NOTUSED(interp); + Jim_Free((char*)objPtr->internalRep.ptr); + objPtr->internalRep.ptr = 0; +} + +void DupScanFmtInternalRep(Jim_Interp *interp, Jim_Obj *srcPtr, Jim_Obj *dupPtr) +{ + size_t size = (size_t)((ScanFmtStringObj*)srcPtr->internalRep.ptr)->size; + ScanFmtStringObj *newVec = (ScanFmtStringObj*)Jim_Alloc(size); + + JIM_NOTUSED(interp); + memcpy(newVec, srcPtr->internalRep.ptr, size); + dupPtr->internalRep.ptr = newVec; + dupPtr->typePtr = &scanFmtStringObjType; +} + +void UpdateStringOfScanFmt(Jim_Obj *objPtr) +{ + char *bytes = ((ScanFmtStringObj*)objPtr->internalRep.ptr)->stringRep; + + objPtr->bytes = Jim_StrDup(bytes); + objPtr->length = strlen(bytes); +} + +/* SetScanFmtFromAny will parse a given string and create the internal + * representation of the format specification. In case of an error + * the error data member of the internal representation will be set + * to an descriptive error text and the function will be left with + * JIM_ERR to indicate unsucessful parsing (aka. malformed scanformat + * specification */ + +static int SetScanFmtFromAny(Jim_Interp *interp, Jim_Obj *objPtr) +{ + ScanFmtStringObj *fmtObj; + char *buffer; + int maxCount, i, approxSize, lastPos = -1; + const char *fmt = objPtr->bytes; + int maxFmtLen = objPtr->length; + const char *fmtEnd = fmt + maxFmtLen; + int curr; + + Jim_FreeIntRep(interp, objPtr); + /* Count how many conversions could take place maximally */ + for (i = 0, maxCount = 0; i < maxFmtLen; ++i) + if (fmt[i] == '%') + ++maxCount; + /* Calculate an approximation of the memory necessary */ + approxSize = sizeof(ScanFmtStringObj) /* Size of the container */ + + (maxCount + 1) * sizeof(ScanFmtPartDescr) /* Size of all partials */ + + maxFmtLen * sizeof(char) + 3 + 1 /* Scratch + "%n" + '\0' */ + + maxFmtLen * sizeof(char) + 1 /* Original stringrep */ + + maxFmtLen * sizeof(char) /* Arg for CHARSETs */ + + (maxCount +1) * sizeof(char) /* '\0' for every partial */ + + 1; /* safety byte */ + fmtObj = (ScanFmtStringObj*)Jim_Alloc(approxSize); + memset(fmtObj, 0, approxSize); + fmtObj->size = approxSize; + fmtObj->maxPos = 0; + fmtObj->scratch = (char*)&fmtObj->descr[maxCount + 1]; + fmtObj->stringRep = fmtObj->scratch + maxFmtLen + 3 + 1; + memcpy(fmtObj->stringRep, fmt, maxFmtLen); + buffer = fmtObj->stringRep + maxFmtLen + 1; + objPtr->internalRep.ptr = fmtObj; + objPtr->typePtr = &scanFmtStringObjType; + for (i = 0, curr = 0; fmt < fmtEnd; ++fmt) { + int width = 0, skip; + ScanFmtPartDescr *descr = &fmtObj->descr[curr]; + fmtObj->count++; + descr->width = 0; /* Assume width unspecified */ + /* Overread and store any "literal" prefix */ + if (*fmt != '%' || fmt[1] == '%') { + descr->type = 0; + descr->prefix = &buffer[i]; + for (; fmt < fmtEnd; ++fmt) { + if (*fmt == '%') { + if (fmt[1] != '%') break; + ++fmt; + } + buffer[i++] = *fmt; + } + buffer[i++] = 0; + } + /* Skip the conversion introducing '%' sign */ + ++fmt; + /* End reached due to non-conversion literal only? */ + if (fmt >= fmtEnd) + goto done; + descr->pos = 0; /* Assume "natural" positioning */ + if (*fmt == '*') { + descr->pos = -1; /* Okay, conversion will not be assigned */ + ++fmt; + } else + fmtObj->convCount++; /* Otherwise count as assign-conversion */ + /* Check if next token is a number (could be width or pos */ + if (sscanf(fmt, "%d%n", &width, &skip) == 1) { + fmt += skip; + /* Was the number a XPG3 position specifier? */ + if (descr->pos != -1 && *fmt == '$') { + int prev; + ++fmt; + descr->pos = width; + width = 0; + /* Look if "natural" postioning and XPG3 one was mixed */ + if ((lastPos == 0 && descr->pos > 0) + || (lastPos > 0 && descr->pos == 0)) { + fmtObj->error = "cannot mix \"%\" and \"%n$\" conversion specifiers"; + return JIM_ERR; + } + /* Look if this position was already used */ + for (prev = 0; prev < curr; ++prev) { + if (fmtObj->descr[prev].pos == -1) continue; + if (fmtObj->descr[prev].pos == descr->pos) { + fmtObj->error = "same \"%n$\" conversion specifier " + "used more than once"; + return JIM_ERR; + } + } + /* Try to find a width after the XPG3 specifier */ + if (sscanf(fmt, "%d%n", &width, &skip) == 1) { + descr->width = width; + fmt += skip; + } + if (descr->pos > 0 && (size_t)descr->pos > fmtObj->maxPos) + fmtObj->maxPos = descr->pos; + } else { + /* Number was not a XPG3, so it has to be a width */ + descr->width = width; + } + } + /* If positioning mode was undetermined yet, fix this */ + if (lastPos == -1) + lastPos = descr->pos; + /* Handle CHARSET conversion type ... */ + if (*fmt == '[') { + int swapped = 1, beg = i, end, j; + descr->type = '['; + descr->arg = &buffer[i]; + ++fmt; + if (*fmt == '^') buffer[i++] = *fmt++; + if (*fmt == ']') buffer[i++] = *fmt++; + while (*fmt && *fmt != ']') buffer[i++] = *fmt++; + if (*fmt != ']') { + fmtObj->error = "unmatched [ in format string"; + return JIM_ERR; + } + end = i; + buffer[i++] = 0; + /* In case a range fence was given "backwards", swap it */ + while (swapped) { + swapped = 0; + for (j = beg + 1; j < end-1; ++j) { + if (buffer[j] == '-' && buffer[j-1] > buffer[j + 1]) { + char tmp = buffer[j-1]; + buffer[j-1] = buffer[j + 1]; + buffer[j + 1] = tmp; + swapped = 1; + } + } + } + } else { + /* Remember any valid modifier if given */ + if (strchr("hlL", *fmt) != 0) + descr->modifier = tolower((int)*fmt++); + + descr->type = *fmt; + if (strchr("efgcsndoxui", *fmt) == 0) { + fmtObj->error = "bad scan conversion character"; + return JIM_ERR; + } else if (*fmt == 'c' && descr->width != 0) { + fmtObj->error = "field width may not be specified in %c " + "conversion"; + return JIM_ERR; + } else if (*fmt == 'u' && descr->modifier == 'l') { + fmtObj->error = "unsigned wide not supported"; + return JIM_ERR; + } + } + curr++; + } +done: + if (fmtObj->convCount == 0) { + fmtObj->error = "no any conversion specifier given"; + return JIM_ERR; + } + return JIM_OK; +} + +/* Some accessor macros to allow lowlevel access to fields of internal repr */ + +#define FormatGetCnvCount(_fo_) \ + ((ScanFmtStringObj*)((_fo_)->internalRep.ptr))->convCount +#define FormatGetMaxPos(_fo_) \ + ((ScanFmtStringObj*)((_fo_)->internalRep.ptr))->maxPos +#define FormatGetError(_fo_) \ + ((ScanFmtStringObj*)((_fo_)->internalRep.ptr))->error + +/* Some Bit testing/setting/cleaning routines. For now only used in handling + * charsets ([a-z123]) within scanning. Later on perhaps a base for a + * bitvector implementation in Jim? */ + +static int JimTestBit(const char *bitvec, char ch) +{ + div_t pos = div(ch-1, 8); + return bitvec[pos.quot] & (1 << pos.rem); +} + +static void JimSetBit(char *bitvec, char ch) +{ + div_t pos = div(ch-1, 8); + bitvec[pos.quot] |= (1 << pos.rem); +} + +#if 0 /* currently not used */ +static void JimClearBit(char *bitvec, char ch) +{ + div_t pos = div(ch-1, 8); + bitvec[pos.quot] &= ~(1 << pos.rem); +} +#endif + +/* JimScanAString is used to scan an unspecified string that ends with + * next WS, or a string that is specified via a charset. The charset + * is currently implemented in a way to only allow for usage with + * ASCII. Whenever we will switch to UNICODE, another idea has to + * be born :-/ + * + * FIXME: Works only with ASCII */ + +static Jim_Obj * +JimScanAString(Jim_Interp *interp, const char *sdescr, const char *str) +{ + size_t i; + Jim_Obj *result; + char charset[256/8 + 1]; /* A Charset may contain max 256 chars */ + char *buffer = Jim_Alloc(strlen(str) + 1), *anchor = buffer; + + /* First init charset to nothing or all, depending if a specified + * or an unspecified string has to be parsed */ + memset(charset, (sdescr ? 0 : 255), sizeof(charset)); + if (sdescr) { + /* There was a set description given, that means we are parsing + * a specified string. So we have to build a corresponding + * charset reflecting the description */ + int notFlag = 0; + /* Should the set be negated at the end? */ + if (*sdescr == '^') { + notFlag = 1; + ++sdescr; + } + /* Here '-' is meant literally and not to define a range */ + if (*sdescr == '-') { + JimSetBit(charset, '-'); + ++sdescr; + } + while (*sdescr) { + if (sdescr[1] == '-' && sdescr[2] != 0) { + /* Handle range definitions */ + int i_t; + for (i_t = sdescr[0]; i_t <= sdescr[2]; ++i_t) + JimSetBit(charset, (char)i_t); + sdescr += 3; + } else { + /* Handle verbatim character definitions */ + JimSetBit(charset, *sdescr++); + } + } + /* Negate the charset if there was a NOT given */ + for (i = 0; notFlag && i < sizeof(charset); ++i) + charset[i] = ~charset[i]; + } + /* And after all the mess above, the real work begin ... */ + while (str && *str) { + if (!sdescr && isspace((int)*str)) + break; /* EOS via WS if unspecified */ + if (JimTestBit(charset, *str)) *buffer++ = *str++; + else break; /* EOS via mismatch if specified scanning */ + } + *buffer = 0; /* Close the string properly ... */ + result = Jim_NewStringObj(interp, anchor, -1); + Jim_Free(anchor); /* ... and free it afer usage */ + return result; +} + +/* ScanOneEntry will scan one entry out of the string passed as argument. + * It use the sscanf() function for this task. After extracting and + * converting of the value, the count of scanned characters will be + * returned of -1 in case of no conversion tool place and string was + * already scanned thru */ + +static int ScanOneEntry(Jim_Interp *interp, const char *str, long pos, + ScanFmtStringObj *fmtObj, long index_t, Jim_Obj **valObjPtr) +{ +# define MAX_SIZE (sizeof(jim_wide) > sizeof(double) \ + ? sizeof(jim_wide) \ + : sizeof(double)) + char buffer[MAX_SIZE]; + char *value = buffer; + const char *tok; + const ScanFmtPartDescr *descr = &fmtObj->descr[index_t]; + size_t sLen = strlen(&str[pos]), scanned = 0; + size_t anchor = pos; + int i; + + /* First pessimiticly assume, we will not scan anything :-) */ + *valObjPtr = 0; + if (descr->prefix) { + /* There was a prefix given before the conversion, skip it and adjust + * the string-to-be-parsed accordingly */ + for (i = 0; str[pos] && descr->prefix[i]; ++i) { + /* If prefix require, skip WS */ + if (isspace((int)descr->prefix[i])) + while (str[pos] && isspace((int)str[pos])) ++pos; + else if (descr->prefix[i] != str[pos]) + break; /* Prefix do not match here, leave the loop */ + else + ++pos; /* Prefix matched so far, next round */ + } + if (str[pos] == 0) + return -1; /* All of str consumed: EOF condition */ + else if (descr->prefix[i] != 0) + return 0; /* Not whole prefix consumed, no conversion possible */ + } + /* For all but following conversion, skip leading WS */ + if (descr->type != 'c' && descr->type != '[' && descr->type != 'n') + while (isspace((int)str[pos])) ++pos; + /* Determine how much skipped/scanned so far */ + scanned = pos - anchor; + if (descr->type == 'n') { + /* Return pseudo conversion means: how much scanned so far? */ + *valObjPtr = Jim_NewIntObj(interp, anchor + scanned); + } else if (str[pos] == 0) { + /* Cannot scan anything, as str is totally consumed */ + return -1; + } else { + /* Processing of conversions follows ... */ + if (descr->width > 0) { + /* Do not try to scan as fas as possible but only the given width. + * To ensure this, we copy the part that should be scanned. */ + size_t tLen = descr->width > sLen ? sLen : descr->width; + tok = Jim_StrDupLen(&str[pos], tLen); + } else { + /* As no width was given, simply refer to the original string */ + tok = &str[pos]; + } + switch (descr->type) { + case 'c': + *valObjPtr = Jim_NewIntObj(interp, *tok); + scanned += 1; + break; + case 'd': case 'o': case 'x': case 'u': case 'i': { + jim_wide jwvalue = 0; + long lvalue = 0; + char *endp; /* Position where the number finished */ + int base = descr->type == 'o' ? 8 + : descr->type == 'x' ? 16 + : descr->type == 'i' ? 0 + : 10; + + do { + /* Try to scan a number with the given base */ + if (descr->modifier == 'l') + { +#ifdef HAVE_LONG_LONG_INT + jwvalue = JimStrtoll(tok, &endp, base), +#else + jwvalue = strtol(tok, &endp, base), +#endif + memcpy(value, &jwvalue, sizeof(jim_wide)); + } + else + { + if (descr->type == 'u') + lvalue = strtoul(tok, &endp, base); + else + lvalue = strtol(tok, &endp, base); + memcpy(value, &lvalue, sizeof(lvalue)); + } + /* If scanning failed, and base was undetermined, simply + * put it to 10 and try once more. This should catch the + * case where %i begin to parse a number prefix (e.g. + * '0x' but no further digits follows. This will be + * handled as a ZERO followed by a char 'x' by Tcl */ + if (endp == tok && base == 0) base = 10; + else break; + } while (1); + if (endp != tok) { + /* There was some number sucessfully scanned! */ + if (descr->modifier == 'l') + *valObjPtr = Jim_NewIntObj(interp, jwvalue); + else + *valObjPtr = Jim_NewIntObj(interp, lvalue); + /* Adjust the number-of-chars scanned so far */ + scanned += endp - tok; + } else { + /* Nothing was scanned. We have to determine if this + * happened due to e.g. prefix mismatch or input str + * exhausted */ + scanned = *tok ? 0 : -1; + } + break; + } + case 's': case '[': { + *valObjPtr = JimScanAString(interp, descr->arg, tok); + scanned += Jim_Length(*valObjPtr); + break; + } + case 'e': case 'f': case 'g': { + char *endp; + + double dvalue = strtod(tok, &endp); + memcpy(value, &dvalue, sizeof(double)); + if (endp != tok) { + /* There was some number sucessfully scanned! */ + *valObjPtr = Jim_NewDoubleObj(interp, dvalue); + /* Adjust the number-of-chars scanned so far */ + scanned += endp - tok; + } else { + /* Nothing was scanned. We have to determine if this + * happened due to e.g. prefix mismatch or input str + * exhausted */ + scanned = *tok ? 0 : -1; + } + break; + } + } + /* If a substring was allocated (due to pre-defined width) do not + * forget to free it */ + if (tok != &str[pos]) + Jim_Free((char*)tok); + } + return scanned; +} + +/* Jim_ScanString is the workhorse of string scanning. It will scan a given + * string and returns all converted (and not ignored) values in a list back + * to the caller. If an error occured, a NULL pointer will be returned */ + +Jim_Obj *Jim_ScanString(Jim_Interp *interp, Jim_Obj *strObjPtr, + Jim_Obj *fmtObjPtr, int flags) +{ + size_t i, pos; + int scanned = 1; + const char *str = Jim_GetString(strObjPtr, 0); + Jim_Obj *resultList = 0; + Jim_Obj **resultVec =NULL; + int resultc; + Jim_Obj *emptyStr = 0; + ScanFmtStringObj *fmtObj; + + /* If format specification is not an object, convert it! */ + if (fmtObjPtr->typePtr != &scanFmtStringObjType) + SetScanFmtFromAny(interp, fmtObjPtr); + fmtObj = (ScanFmtStringObj*)fmtObjPtr->internalRep.ptr; + /* Check if format specification was valid */ + if (fmtObj->error != 0) { + if (flags & JIM_ERRMSG) + Jim_SetResultString(interp, fmtObj->error, -1); + return 0; + } + /* Allocate a new "shared" empty string for all unassigned conversions */ + emptyStr = Jim_NewEmptyStringObj(interp); + Jim_IncrRefCount(emptyStr); + /* Create a list and fill it with empty strings up to max specified XPG3 */ + resultList = Jim_NewListObj(interp, 0, 0); + if (fmtObj->maxPos > 0) { + for (i = 0; i < fmtObj->maxPos; ++i) + Jim_ListAppendElement(interp, resultList, emptyStr); + JimListGetElements(interp, resultList, &resultc, &resultVec); + } + /* Now handle every partial format description */ + for (i = 0, pos = 0; i < fmtObj->count; ++i) { + ScanFmtPartDescr *descr = &(fmtObj->descr[i]); + Jim_Obj *value = 0; + /* Only last type may be "literal" w/o conversion - skip it! */ + if (descr->type == 0) continue; + /* As long as any conversion could be done, we will proceed */ + if (scanned > 0) + scanned = ScanOneEntry(interp, str, pos, fmtObj, i, &value); + /* In case our first try results in EOF, we will leave */ + if (scanned == -1 && i == 0) + goto eof; + /* Advance next pos-to-be-scanned for the amount scanned already */ + pos += scanned; + /* value == 0 means no conversion took place so take empty string */ + if (value == 0) + value = Jim_NewEmptyStringObj(interp); + /* If value is a non-assignable one, skip it */ + if (descr->pos == -1) { + Jim_FreeNewObj(interp, value); + } else if (descr->pos == 0) + /* Otherwise append it to the result list if no XPG3 was given */ + Jim_ListAppendElement(interp, resultList, value); + else if (resultVec[descr->pos-1] == emptyStr) { + /* But due to given XPG3, put the value into the corr. slot */ + Jim_DecrRefCount(interp, resultVec[descr->pos-1]); + Jim_IncrRefCount(value); + resultVec[descr->pos-1] = value; + } else { + /* Otherwise, the slot was already used - free obj and ERROR */ + Jim_FreeNewObj(interp, value); + goto err; + } + } + Jim_DecrRefCount(interp, emptyStr); + return resultList; +eof: + Jim_DecrRefCount(interp, emptyStr); + Jim_FreeNewObj(interp, resultList); + return (Jim_Obj*)EOF; +err: + Jim_DecrRefCount(interp, emptyStr); + Jim_FreeNewObj(interp, resultList); + return 0; +} + +/* ----------------------------------------------------------------------------- + * Pseudo Random Number Generation + * ---------------------------------------------------------------------------*/ +static void JimPrngSeed(Jim_Interp *interp, const unsigned char *seed, + int seedLen); + +/* Initialize the sbox with the numbers from 0 to 255 */ +static void JimPrngInit(Jim_Interp *interp) +{ + int i; + unsigned int seed[256]; + + interp->prngState = Jim_Alloc(sizeof(Jim_PrngState)); + for (i = 0; i < 256; i++) + seed[i] = (rand() ^ time(NULL) ^ clock()); + JimPrngSeed(interp, (unsigned char*) seed, sizeof(int)*256); +} + +/* Generates N bytes of random data */ +static void JimRandomBytes(Jim_Interp *interp, void *dest, unsigned int len) +{ + Jim_PrngState *prng; + unsigned char *destByte = (unsigned char*) dest; + unsigned int si, sj, x; + + /* initialization, only needed the first time */ + if (interp->prngState == NULL) + JimPrngInit(interp); + prng = interp->prngState; + /* generates 'len' bytes of pseudo-random numbers */ + for (x = 0; x < len; x++) { + prng->i = (prng->i + 1) & 0xff; + si = prng->sbox[prng->i]; + prng->j = (prng->j + si) & 0xff; + sj = prng->sbox[prng->j]; + prng->sbox[prng->i] = sj; + prng->sbox[prng->j] = si; + *destByte++ = prng->sbox[(si + sj)&0xff]; + } +} + +/* Re-seed the generator with user-provided bytes */ +static void JimPrngSeed(Jim_Interp *interp, const unsigned char *seed, + int seedLen) +{ + int i; + unsigned char buf[256]; + Jim_PrngState *prng; + + /* initialization, only needed the first time */ + if (interp->prngState == NULL) + JimPrngInit(interp); + prng = interp->prngState; + + /* Set the sbox[i] with i */ + for (i = 0; i < 256; i++) + prng->sbox[i] = i; + /* Now use the seed to perform a random permutation of the sbox */ + for (i = 0; i < seedLen; i++) { + unsigned char t; + + t = prng->sbox[i&0xFF]; + prng->sbox[i&0xFF] = prng->sbox[seed[i]]; + prng->sbox[seed[i]] = t; + } + prng->i = prng->j = 0; + /* discard the first 256 bytes of stream. */ + JimRandomBytes(interp, buf, 256); +} + +/* ----------------------------------------------------------------------------- + * Dynamic libraries support (WIN32 not supported) + * ---------------------------------------------------------------------------*/ + +#ifdef JIM_DYNLIB +#ifdef WIN32 +#define RTLD_LAZY 0 +void * dlopen(const char *path, int mode) +{ + JIM_NOTUSED(mode); + + return (void *)LoadLibraryA(path); +} +int dlclose(void *handle) +{ + FreeLibrary((HANDLE)handle); + return 0; +} +void *dlsym(void *handle, const char *symbol) +{ + return GetProcAddress((HMODULE)handle, symbol); +} +static char win32_dlerror_string[121]; +const char *dlerror(void) +{ + FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, NULL, GetLastError(), + LANG_NEUTRAL, win32_dlerror_string, 120, NULL); + return win32_dlerror_string; +} +#endif /* WIN32 */ + +int Jim_LoadLibrary(Jim_Interp *interp, const char *pathName) +{ + Jim_Obj *libPathObjPtr; + int prefixc, i; + void *handle; + int (*onload)(Jim_Interp *interp); + + libPathObjPtr = Jim_GetGlobalVariableStr(interp, "jim_libpath", JIM_NONE); + if (libPathObjPtr == NULL) { + prefixc = 0; + libPathObjPtr = NULL; + } else { + Jim_IncrRefCount(libPathObjPtr); + Jim_ListLength(interp, libPathObjPtr, &prefixc); + } + + for (i = -1; i < prefixc; i++) { + if (i < 0) { + handle = dlopen(pathName, RTLD_LAZY); + } else { + FILE *fp; + char buf[JIM_PATH_LEN]; + const char *prefix; + int prefixlen; + Jim_Obj *prefixObjPtr; + + buf[0] = '\0'; + if (Jim_ListIndex(interp, libPathObjPtr, i, + &prefixObjPtr, JIM_NONE) != JIM_OK) + continue; + prefix = Jim_GetString(prefixObjPtr, &prefixlen); + if (prefixlen + strlen(pathName) + 1 >= JIM_PATH_LEN) + continue; + if (*pathName == '/') { + strcpy(buf, pathName); + } + else if (prefixlen && prefix[prefixlen-1] == '/') + sprintf(buf, "%s%s", prefix, pathName); + else + sprintf(buf, "%s/%s", prefix, pathName); + fp = fopen(buf, "r"); + if (fp == NULL) + continue; + fclose(fp); + handle = dlopen(buf, RTLD_LAZY); + } + if (handle == NULL) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "error loading extension \"", pathName, + "\": ", dlerror(), NULL); + if (i < 0) + continue; + goto err; + } + if ((onload = dlsym(handle, "Jim_OnLoad")) == NULL) { + Jim_SetResultString(interp, + "No Jim_OnLoad symbol found on extension", -1); + goto err; + } + if (onload(interp) == JIM_ERR) { + dlclose(handle); + goto err; + } + Jim_SetEmptyResult(interp); + if (libPathObjPtr != NULL) + Jim_DecrRefCount(interp, libPathObjPtr); + return JIM_OK; + } +err: + if (libPathObjPtr != NULL) + Jim_DecrRefCount(interp, libPathObjPtr); + return JIM_ERR; +} +#else /* JIM_DYNLIB */ +int Jim_LoadLibrary(Jim_Interp *interp, const char *pathName) +{ + JIM_NOTUSED(interp); + JIM_NOTUSED(pathName); + + Jim_SetResultString(interp, "the Jim binary has no support for [load]", -1); + return JIM_ERR; +} +#endif/* JIM_DYNLIB */ + +/* ----------------------------------------------------------------------------- + * Packages handling + * ---------------------------------------------------------------------------*/ + +#define JIM_PKG_ANY_VERSION -1 + +/* Convert a string of the type "1.2" into an integer. + * MAJOR.MINOR is converted as MAJOR*100 + MINOR, so "1.2" is converted + * to the integer with value 102 */ +static int JimPackageVersionToInt(Jim_Interp *interp, const char *v, + int *intPtr, int flags) +{ + char *copy; + jim_wide major, minor; + char *majorStr, *minorStr, *p; + + if (v[0] == '\0') { + *intPtr = JIM_PKG_ANY_VERSION; + return JIM_OK; + } + + copy = Jim_StrDup(v); + p = strchr(copy, '.'); + if (p == NULL) goto badfmt; + *p = '\0'; + majorStr = copy; + minorStr = p + 1; + + if (Jim_StringToWide(majorStr, &major, 10) != JIM_OK || + Jim_StringToWide(minorStr, &minor, 10) != JIM_OK) + goto badfmt; + *intPtr = (int)(major*100 + minor); + Jim_Free(copy); + return JIM_OK; + +badfmt: + Jim_Free(copy); + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "invalid package version '", v, "'", NULL); + } + return JIM_ERR; +} + +#define JIM_MATCHVER_EXACT (1 << JIM_PRIV_FLAG_SHIFT) +static int JimPackageMatchVersion(int needed, int actual, int flags) +{ + if (needed == JIM_PKG_ANY_VERSION) return 1; + if (flags & JIM_MATCHVER_EXACT) { + return needed == actual; + } else { + return needed/100 == actual/100 && (needed <= actual); + } +} + +int Jim_PackageProvide(Jim_Interp *interp, const char *name, const char *ver, + int flags) +{ + int intVersion; + /* Check if the version format is ok */ + if (JimPackageVersionToInt(interp, ver, &intVersion, JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + /* If the package was already provided returns an error. */ + if (Jim_FindHashEntry(&interp->packages, name) != NULL) { + if (flags & JIM_ERRMSG) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "package '", name, "' was already provided", NULL); + } + return JIM_ERR; + } + Jim_AddHashEntry(&interp->packages, name, (char*) ver); + return JIM_OK; +} + +#ifndef JIM_ANSIC + +#ifndef WIN32 +# include +# include +#else +# include +/* Posix dirent.h compatiblity layer for WIN32. + * Copyright Kevlin Henney, 1997, 2003. All rights reserved. + * Copyright Salvatore Sanfilippo ,2005. + * + * Permission to use, copy, modify, and distribute this software and its + * documentation for any purpose is hereby granted without fee, provided + * that this copyright and permissions notice appear in all copies and + * derivatives. + * + * This software is supplied "as is" without express or implied warranty. + * This software was modified by Salvatore Sanfilippo for the Jim Interpreter. + */ + +struct dirent { + char *d_name; +}; + +typedef struct DIR { + long handle; /* -1 for failed rewind */ + struct _finddata_t info; + struct dirent result; /* d_name null iff first time */ + char *name; /* null-terminated char string */ +} DIR; + +DIR *opendir(const char *name) +{ + DIR *dir = 0; + + if (name && name[0]) { + size_t base_length = strlen(name); + const char *all = /* search pattern must end with suitable wildcard */ + strchr("/\\", name[base_length - 1]) ? "*" : "/*"; + + if ((dir = (DIR *) Jim_Alloc(sizeof *dir)) != 0 && + (dir->name = (char *) Jim_Alloc(base_length + strlen(all) + 1)) != 0) + { + strcat(strcpy(dir->name, name), all); + + if ((dir->handle = (long) _findfirst(dir->name, &dir->info)) != -1) + dir->result.d_name = 0; + else { /* rollback */ + Jim_Free(dir->name); + Jim_Free(dir); + dir = 0; + } + } else { /* rollback */ + Jim_Free(dir); + dir = 0; + errno = ENOMEM; + } + } else { + errno = EINVAL; + } + return dir; +} + +int closedir(DIR *dir) +{ + int result = -1; + + if (dir) { + if (dir->handle != -1) + result = _findclose(dir->handle); + Jim_Free(dir->name); + Jim_Free(dir); + } + if (result == -1) /* map all errors to EBADF */ + errno = EBADF; + return result; +} + +struct dirent *readdir(DIR *dir) +{ + struct dirent *result = 0; + + if (dir && dir->handle != -1) { + if (!dir->result.d_name || _findnext(dir->handle, &dir->info) != -1) { + result = &dir->result; + result->d_name = dir->info.name; + } + } else { + errno = EBADF; + } + return result; +} + +#endif /* WIN32 */ + +static char *JimFindBestPackage(Jim_Interp *interp, char **prefixes, + int prefixc, const char *pkgName, int pkgVer, int flags) +{ + int bestVer = -1, i; + int pkgNameLen = strlen(pkgName); + char *bestPackage = NULL; + struct dirent *de; + + for (i = 0; i < prefixc; i++) { + DIR *dir; + char buf[JIM_PATH_LEN]; + int prefixLen; + + if (prefixes[i] == NULL) continue; + strncpy(buf, prefixes[i], JIM_PATH_LEN); + buf[JIM_PATH_LEN-1] = '\0'; + prefixLen = strlen(buf); + if (prefixLen && buf[prefixLen-1] == '/') + buf[prefixLen-1] = '\0'; + + if ((dir = opendir(buf)) == NULL) continue; + while ((de = readdir(dir)) != NULL) { + char *fileName = de->d_name; + int fileNameLen = strlen(fileName); + + if (strncmp(fileName, "jim-", 4) == 0 && + strncmp(fileName + 4, pkgName, pkgNameLen) == 0 && + *(fileName + 4+pkgNameLen) == '-' && + fileNameLen > 4 && /* note that this is not really useful */ + (strncmp(fileName + fileNameLen-4, ".tcl", 4) == 0 || + strncmp(fileName + fileNameLen-4, ".dll", 4) == 0 || + strncmp(fileName + fileNameLen-3, ".so", 3) == 0)) + { + char ver[6]; /* xx.yy < nulterm> */ + char *p = strrchr(fileName, '.'); + int verLen, fileVer; + + verLen = p - (fileName + 4+pkgNameLen + 1); + if (verLen < 3 || verLen > 5) continue; + memcpy(ver, fileName + 4+pkgNameLen + 1, verLen); + ver[verLen] = '\0'; + if (JimPackageVersionToInt(interp, ver, &fileVer, JIM_NONE) + != JIM_OK) continue; + if (JimPackageMatchVersion(pkgVer, fileVer, flags) && + (bestVer == -1 || bestVer < fileVer)) + { + bestVer = fileVer; + Jim_Free(bestPackage); + bestPackage = Jim_Alloc(strlen(buf) + strlen(fileName) + 2); + sprintf(bestPackage, "%s/%s", buf, fileName); + } + } + } + closedir(dir); + } + return bestPackage; +} + +#else /* JIM_ANSIC */ + +static char *JimFindBestPackage(Jim_Interp *interp, char **prefixes, + int prefixc, const char *pkgName, int pkgVer, int flags) +{ + JIM_NOTUSED(interp); + JIM_NOTUSED(prefixes); + JIM_NOTUSED(prefixc); + JIM_NOTUSED(pkgName); + JIM_NOTUSED(pkgVer); + JIM_NOTUSED(flags); + return NULL; +} + +#endif /* JIM_ANSIC */ + +/* Search for a suitable package under every dir specified by jim_libpath + * and load it if possible. If a suitable package was loaded with success + * JIM_OK is returned, otherwise JIM_ERR is returned. */ +static int JimLoadPackage(Jim_Interp *interp, const char *name, int ver, + int flags) +{ + Jim_Obj *libPathObjPtr; + char **prefixes, *best; + int prefixc, i, retCode = JIM_OK; + + libPathObjPtr = Jim_GetGlobalVariableStr(interp, "jim_libpath", JIM_NONE); + if (libPathObjPtr == NULL) { + prefixc = 0; + libPathObjPtr = NULL; + } else { + Jim_IncrRefCount(libPathObjPtr); + Jim_ListLength(interp, libPathObjPtr, &prefixc); + } + + prefixes = Jim_Alloc(sizeof(char*)*prefixc); + for (i = 0; i < prefixc; i++) { + Jim_Obj *prefixObjPtr; + if (Jim_ListIndex(interp, libPathObjPtr, i, + &prefixObjPtr, JIM_NONE) != JIM_OK) + { + prefixes[i] = NULL; + continue; + } + prefixes[i] = Jim_StrDup(Jim_GetString(prefixObjPtr, NULL)); + } + /* Scan every directory to find the "best" package. */ + best = JimFindBestPackage(interp, prefixes, prefixc, name, ver, flags); + if (best != NULL) { + char *p = strrchr(best, '.'); + /* Try to load/source it */ + if (p && strcmp(p, ".tcl") == 0) { + retCode = Jim_EvalFile(interp, best); + } else { + retCode = Jim_LoadLibrary(interp, best); + } + } else { + retCode = JIM_ERR; + } + Jim_Free(best); + for (i = 0; i < prefixc; i++) + Jim_Free(prefixes[i]); + Jim_Free(prefixes); + if (libPathObjPtr) + Jim_DecrRefCount(interp, libPathObjPtr); + return retCode; +} + +const char *Jim_PackageRequire(Jim_Interp *interp, const char *name, + const char *ver, int flags) +{ + Jim_HashEntry *he; + int requiredVer; + + /* Start with an empty error string */ + Jim_SetResultString(interp, "", 0); + + if (JimPackageVersionToInt(interp, ver, &requiredVer, JIM_ERRMSG) != JIM_OK) + return NULL; + he = Jim_FindHashEntry(&interp->packages, name); + if (he == NULL) { + /* Try to load the package. */ + if (JimLoadPackage(interp, name, requiredVer, flags) == JIM_OK) { + he = Jim_FindHashEntry(&interp->packages, name); + if (he == NULL) { + return "?"; + } + return he->val; + } + /* No way... return an error. */ + if (flags & JIM_ERRMSG) { + int len; + Jim_GetString(Jim_GetResult(interp), &len); + Jim_AppendStrings(interp, Jim_GetResult(interp), len ? "\n" : "", + "Can't find package '", name, "'", NULL); + } + return NULL; + } else { + int actualVer; + if (JimPackageVersionToInt(interp, he->val, &actualVer, JIM_ERRMSG) + != JIM_OK) + { + return NULL; + } + /* Check if version matches. */ + if (JimPackageMatchVersion(requiredVer, actualVer, flags) == 0) { + Jim_AppendStrings(interp, Jim_GetResult(interp), + "Package '", name, "' already loaded, but with version ", + he->val, NULL); + return NULL; + } + return he->val; + } +} + +/* ----------------------------------------------------------------------------- + * Eval + * ---------------------------------------------------------------------------*/ +#define JIM_EVAL_SARGV_LEN 8 /* static arguments vector length */ +#define JIM_EVAL_SINTV_LEN 8 /* static interpolation vector length */ + +static int JimCallProcedure(Jim_Interp *interp, Jim_Cmd *cmd, int argc, + Jim_Obj *const *argv); + +/* Handle calls to the [unknown] command */ +static int JimUnknown(Jim_Interp *interp, int argc, Jim_Obj *const *argv) +{ + Jim_Obj **v, *sv[JIM_EVAL_SARGV_LEN]; + int retCode; + + /* If JimUnknown() is recursively called (e.g. error in the unknown proc, + * done here + */ + if (interp->unknown_called) { + return JIM_ERR; + } + + /* If the [unknown] command does not exists returns + * just now */ + if (Jim_GetCommand(interp, interp->unknown, JIM_NONE) == NULL) + return JIM_ERR; + + /* The object interp->unknown just contains + * the "unknown" string, it is used in order to + * avoid to lookup the unknown command every time + * but instread to cache the result. */ + if (argc + 1 <= JIM_EVAL_SARGV_LEN) + v = sv; + else + v = Jim_Alloc(sizeof(Jim_Obj*)*(argc + 1)); + /* Make a copy of the arguments vector, but shifted on + * the right of one position. The command name of the + * command will be instead the first argument of the + * [unknonw] call. */ + memcpy(v + 1, argv, sizeof(Jim_Obj*)*argc); + v[0] = interp->unknown; + /* Call it */ + interp->unknown_called++; + retCode = Jim_EvalObjVector(interp, argc + 1, v); + interp->unknown_called--; + + /* Clean up */ + if (v != sv) + Jim_Free(v); + return retCode; +} + +/* Eval the object vector 'objv' composed of 'objc' elements. + * Every element is used as single argument. + * Jim_EvalObj() will call this function every time its object + * argument is of "list" type, with no string representation. + * + * This is possible because the string representation of a + * list object generated by the UpdateStringOfList is made + * in a way that ensures that every list element is a different + * command argument. */ +int Jim_EvalObjVector(Jim_Interp *interp, int objc, Jim_Obj *const *objv) +{ + int i, retcode; + Jim_Cmd *cmdPtr; + + /* Incr refcount of arguments. */ + for (i = 0; i < objc; i++) + Jim_IncrRefCount(objv[i]); + /* Command lookup */ + cmdPtr = Jim_GetCommand(interp, objv[0], JIM_ERRMSG); + if (cmdPtr == NULL) { + retcode = JimUnknown(interp, objc, objv); + } else { + /* Call it -- Make sure result is an empty object. */ + Jim_SetEmptyResult(interp); + if (cmdPtr->cmdProc) { + interp->cmdPrivData = cmdPtr->privData; + retcode = cmdPtr->cmdProc(interp, objc, objv); + if (retcode == JIM_ERR_ADDSTACK) { + //JimAppendStackTrace(interp, "", script->fileName, token[i-argc*2].linenr); + retcode = JIM_ERR; + } + } else { + retcode = JimCallProcedure(interp, cmdPtr, objc, objv); + if (retcode == JIM_ERR) { + JimAppendStackTrace(interp, + Jim_GetString(objv[0], NULL), "", 1); + } + } + } + /* Decr refcount of arguments and return the retcode */ + for (i = 0; i < objc; i++) + Jim_DecrRefCount(interp, objv[i]); + return retcode; +} + +/* Interpolate the given tokens into a unique Jim_Obj returned by reference + * via *objPtrPtr. This function is only called by Jim_EvalObj(). + * The returned object has refcount = 0. */ +int Jim_InterpolateTokens(Jim_Interp *interp, ScriptToken *token, + int tokens, Jim_Obj **objPtrPtr) +{ + int totlen = 0, i, retcode; + Jim_Obj **intv; + Jim_Obj *sintv[JIM_EVAL_SINTV_LEN]; + Jim_Obj *objPtr; + char *s; + + if (tokens <= JIM_EVAL_SINTV_LEN) + intv = sintv; + else + intv = Jim_Alloc(sizeof(Jim_Obj*)* + tokens); + /* Compute every token forming the argument + * in the intv objects vector. */ + for (i = 0; i < tokens; i++) { + switch (token[i].type) { + case JIM_TT_ESC: + case JIM_TT_STR: + intv[i] = token[i].objPtr; + break; + case JIM_TT_VAR: + intv[i] = Jim_GetVariable(interp, token[i].objPtr, JIM_ERRMSG); + if (!intv[i]) { + retcode = JIM_ERR; + goto err; + } + break; + case JIM_TT_DICTSUGAR: + intv[i] = Jim_ExpandDictSugar(interp, token[i].objPtr); + if (!intv[i]) { + retcode = JIM_ERR; + goto err; + } + break; + case JIM_TT_CMD: + retcode = Jim_EvalObj(interp, token[i].objPtr); + if (retcode != JIM_OK) + goto err; + intv[i] = Jim_GetResult(interp); + break; + default: + Jim_Panic(interp, + "default token type reached " + "in Jim_InterpolateTokens()."); + break; + } + Jim_IncrRefCount(intv[i]); + /* Make sure there is a valid + * string rep, and add the string + * length to the total legnth. */ + Jim_GetString(intv[i], NULL); + totlen += intv[i]->length; + } + /* Concatenate every token in an unique + * object. */ + objPtr = Jim_NewStringObjNoAlloc(interp, + NULL, 0); + s = objPtr->bytes = Jim_Alloc(totlen + 1); + objPtr->length = totlen; + for (i = 0; i < tokens; i++) { + memcpy(s, intv[i]->bytes, intv[i]->length); + s += intv[i]->length; + Jim_DecrRefCount(interp, intv[i]); + } + objPtr->bytes[totlen] = '\0'; + /* Free the intv vector if not static. */ + if (tokens > JIM_EVAL_SINTV_LEN) + Jim_Free(intv); + *objPtrPtr = objPtr; + return JIM_OK; +err: + i--; + for (; i >= 0; i--) + Jim_DecrRefCount(interp, intv[i]); + if (tokens > JIM_EVAL_SINTV_LEN) + Jim_Free(intv); + return retcode; +} + +/* Helper of Jim_EvalObj() to perform argument expansion. + * Basically this function append an argument to 'argv' + * (and increments argc by reference accordingly), performing + * expansion of the list object if 'expand' is non-zero, or + * just adding objPtr to argv if 'expand' is zero. */ +void Jim_ExpandArgument(Jim_Interp *interp, Jim_Obj ***argv, + int *argcPtr, int expand, Jim_Obj *objPtr) +{ + if (!expand) { + (*argv) = Jim_Realloc(*argv, sizeof(Jim_Obj*)*((*argcPtr) + 1)); + /* refcount of objPtr not incremented because + * we are actually transfering a reference from + * the old 'argv' to the expanded one. */ + (*argv)[*argcPtr] = objPtr; + (*argcPtr)++; + } else { + int len, i; + + Jim_ListLength(interp, objPtr, &len); + (*argv) = Jim_Realloc(*argv, sizeof(Jim_Obj*)*((*argcPtr) + len)); + for (i = 0; i < len; i++) { + (*argv)[*argcPtr] = objPtr->internalRep.listValue.ele[i]; + Jim_IncrRefCount(objPtr->internalRep.listValue.ele[i]); + (*argcPtr)++; + } + /* The original object reference is no longer needed, + * after the expansion it is no longer present on + * the argument vector, but the single elements are + * in its place. */ + Jim_DecrRefCount(interp, objPtr); + } +} + +int Jim_EvalObj(Jim_Interp *interp, Jim_Obj *scriptObjPtr) +{ + int i, j = 0, len; + ScriptObj *script; + ScriptToken *token; + int *cs; /* command structure array */ + int retcode = JIM_OK; + Jim_Obj *sargv[JIM_EVAL_SARGV_LEN], **argv = NULL, *tmpObjPtr; + + interp->errorFlag = 0; + + /* If the object is of type "list" and there is no + * string representation for this object, we can call + * a specialized version of Jim_EvalObj() */ + if (scriptObjPtr->typePtr == &listObjType && + scriptObjPtr->internalRep.listValue.len && + scriptObjPtr->bytes == NULL) { + Jim_IncrRefCount(scriptObjPtr); + retcode = Jim_EvalObjVector(interp, + scriptObjPtr->internalRep.listValue.len, + scriptObjPtr->internalRep.listValue.ele); + Jim_DecrRefCount(interp, scriptObjPtr); + return retcode; + } + + Jim_IncrRefCount(scriptObjPtr); /* Make sure it's shared. */ + script = Jim_GetScript(interp, scriptObjPtr); + /* Now we have to make sure the internal repr will not be + * freed on shimmering. + * + * Think for example to this: + * + * set x {llength $x; ... some more code ...}; eval $x + * + * In order to preserve the internal rep, we increment the + * inUse field of the script internal rep structure. */ + script->inUse++; + + token = script->token; + len = script->len; + cs = script->cmdStruct; + i = 0; /* 'i' is the current token index. */ + + /* Reset the interpreter result. This is useful to + * return the emtpy result in the case of empty program. */ + Jim_SetEmptyResult(interp); + + /* Execute every command sequentially, returns on + * error (i.e. if a command does not return JIM_OK) */ + while (i < len) { + int expand = 0; + int argc = *cs++; /* Get the number of arguments */ + Jim_Cmd *cmd; + + /* Set the expand flag if needed. */ + if (argc == -1) { + expand++; + argc = *cs++; + } + /* Allocate the arguments vector */ + if (argc <= JIM_EVAL_SARGV_LEN) + argv = sargv; + else + argv = Jim_Alloc(sizeof(Jim_Obj*)*argc); + /* Populate the arguments objects. */ + for (j = 0; j < argc; j++) { + int tokens = *cs++; + + /* tokens is negative if expansion is needed. + * for this argument. */ + if (tokens < 0) { + tokens = (-tokens)-1; + i++; + } + if (tokens == 1) { + /* Fast path if the token does not + * need interpolation */ + switch (token[i].type) { + case JIM_TT_ESC: + case JIM_TT_STR: + argv[j] = token[i].objPtr; + break; + case JIM_TT_VAR: + tmpObjPtr = Jim_GetVariable(interp, token[i].objPtr, + JIM_ERRMSG); + if (!tmpObjPtr) { + retcode = JIM_ERR; + goto err; + } + argv[j] = tmpObjPtr; + break; + case JIM_TT_DICTSUGAR: + tmpObjPtr = Jim_ExpandDictSugar(interp, token[i].objPtr); + if (!tmpObjPtr) { + retcode = JIM_ERR; + goto err; + } + argv[j] = tmpObjPtr; + break; + case JIM_TT_CMD: + retcode = Jim_EvalObj(interp, token[i].objPtr); + if (retcode != JIM_OK) + goto err; + argv[j] = Jim_GetResult(interp); + break; + default: + Jim_Panic(interp, + "default token type reached " + "in Jim_EvalObj()."); + break; + } + Jim_IncrRefCount(argv[j]); + i += 2; + } else { + /* For interpolation we call an helper + * function doing the work for us. */ + if ((retcode = Jim_InterpolateTokens(interp, + token + i, tokens, &tmpObjPtr)) != JIM_OK) + { + goto err; + } + argv[j] = tmpObjPtr; + Jim_IncrRefCount(argv[j]); + i += tokens + 1; + } + } + /* Handle {expand} expansion */ + if (expand) { + int *ecs = cs - argc; + int eargc = 0; + Jim_Obj **eargv = NULL; + + for (j = 0; j < argc; j++) { + Jim_ExpandArgument(interp, &eargv, &eargc, + ecs[j] < 0, argv[j]); + } + if (argv != sargv) + Jim_Free(argv); + argc = eargc; + argv = eargv; + j = argc; + if (argc == 0) { + /* Nothing to do with zero args. */ + Jim_Free(eargv); + continue; + } + } + /* Lookup the command to call */ + cmd = Jim_GetCommand(interp, argv[0], JIM_ERRMSG); + if (cmd != NULL) { + /* Call it -- Make sure result is an empty object. */ + Jim_SetEmptyResult(interp); + if (cmd->cmdProc) { + interp->cmdPrivData = cmd->privData; + retcode = cmd->cmdProc(interp, argc, argv); + if ((retcode == JIM_ERR)||(retcode == JIM_ERR_ADDSTACK)) { + JimAppendStackTrace(interp, "", script->fileName, token[i-argc*2].linenr); + retcode = JIM_ERR; + } + } else { + retcode = JimCallProcedure(interp, cmd, argc, argv); + if (retcode == JIM_ERR) { + JimAppendStackTrace(interp, + Jim_GetString(argv[0], NULL), script->fileName, + token[i-argc*2].linenr); + } + } + } else { + /* Call [unknown] */ + retcode = JimUnknown(interp, argc, argv); + if (retcode == JIM_ERR) { + JimAppendStackTrace(interp, + "", script->fileName, + token[i-argc*2].linenr); + } + } + if (retcode != JIM_OK) { + i -= argc*2; /* point to the command name. */ + goto err; + } + /* Decrement the arguments count */ + for (j = 0; j < argc; j++) { + Jim_DecrRefCount(interp, argv[j]); + } + + if (argv != sargv) { + Jim_Free(argv); + argv = NULL; + } + } + /* Note that we don't have to decrement inUse, because the + * following code transfers our use of the reference again to + * the script object. */ + j = 0; /* on normal termination, the argv array is already + Jim_DecrRefCount-ed. */ +err: + /* Handle errors. */ + if (retcode == JIM_ERR && !interp->errorFlag) { + interp->errorFlag = 1; + JimSetErrorFileName(interp, script->fileName); + JimSetErrorLineNumber(interp, token[i].linenr); + JimResetStackTrace(interp); + } + Jim_FreeIntRep(interp, scriptObjPtr); + scriptObjPtr->typePtr = &scriptObjType; + Jim_SetIntRepPtr(scriptObjPtr, script); + Jim_DecrRefCount(interp, scriptObjPtr); + for (i = 0; i < j; i++) { + Jim_DecrRefCount(interp, argv[i]); + } + if (argv != sargv) + Jim_Free(argv); + return retcode; +} + +/* Call a procedure implemented in Tcl. + * It's possible to speed-up a lot this function, currently + * the callframes are not cached, but allocated and + * destroied every time. What is expecially costly is + * to create/destroy the local vars hash table every time. + * + * This can be fixed just implementing callframes caching + * in JimCreateCallFrame() and JimFreeCallFrame(). */ +int JimCallProcedure(Jim_Interp *interp, Jim_Cmd *cmd, int argc, + Jim_Obj *const *argv) +{ + int i, retcode; + Jim_CallFrame *callFramePtr; + int num_args; + + /* Check arity */ + if (argc < cmd->arityMin || (cmd->arityMax != -1 && + argc > cmd->arityMax)) { + Jim_Obj *objPtr = Jim_NewEmptyStringObj(interp); + Jim_AppendStrings(interp, objPtr, + "wrong # args: should be \"", Jim_GetString(argv[0], NULL), + (cmd->arityMin > 1) ? " " : "", + Jim_GetString(cmd->argListObjPtr, NULL), "\"", NULL); + Jim_SetResult(interp, objPtr); + return JIM_ERR; + } + /* Check if there are too nested calls */ + if (interp->numLevels == interp->maxNestingDepth) { + Jim_SetResultString(interp, + "Too many nested calls. Infinite recursion?", -1); + return JIM_ERR; + } + /* Create a new callframe */ + callFramePtr = JimCreateCallFrame(interp); + callFramePtr->parentCallFrame = interp->framePtr; + callFramePtr->argv = argv; + callFramePtr->argc = argc; + callFramePtr->procArgsObjPtr = cmd->argListObjPtr; + callFramePtr->procBodyObjPtr = cmd->bodyObjPtr; + callFramePtr->staticVars = cmd->staticVars; + Jim_IncrRefCount(cmd->argListObjPtr); + Jim_IncrRefCount(cmd->bodyObjPtr); + interp->framePtr = callFramePtr; + interp->numLevels ++; + + /* Set arguments */ + Jim_ListLength(interp, cmd->argListObjPtr, &num_args); + + /* If last argument is 'args', don't set it here */ + if (cmd->arityMax == -1) { + num_args--; + } + + for (i = 0; i < num_args; i++) { + Jim_Obj *argObjPtr=NULL; + Jim_Obj *nameObjPtr=NULL; + Jim_Obj *valueObjPtr=NULL; + + Jim_ListIndex(interp, cmd->argListObjPtr, i, &argObjPtr, JIM_NONE); + if (i + 1 >= cmd->arityMin) { + /* The name is the first element of the list */ + Jim_ListIndex(interp, argObjPtr, 0, &nameObjPtr, JIM_NONE); + } + else { + /* The element arg is the name */ + nameObjPtr = argObjPtr; + } + + if (i + 1 >= argc) { + /* No more values, so use default */ + /* The value is the second element of the list */ + Jim_ListIndex(interp, argObjPtr, 1, &valueObjPtr, JIM_NONE); + } + else { + valueObjPtr = argv[i + 1]; + } + Jim_SetVariable(interp, nameObjPtr, valueObjPtr); + } + /* Set optional arguments */ + if (cmd->arityMax == -1) { + Jim_Obj *listObjPtr=NULL, *objPtr=NULL; + + i++; + listObjPtr = Jim_NewListObj(interp, argv + i, argc-i); + Jim_ListIndex(interp, cmd->argListObjPtr, num_args, &objPtr, JIM_NONE); + Jim_SetVariable(interp, objPtr, listObjPtr); + } + /* Eval the body */ + retcode = Jim_EvalObj(interp, cmd->bodyObjPtr); + + /* Destroy the callframe */ + interp->numLevels --; + interp->framePtr = interp->framePtr->parentCallFrame; + if (callFramePtr->vars.size != JIM_HT_INITIAL_SIZE) { + JimFreeCallFrame(interp, callFramePtr, JIM_FCF_NONE); + } else { + JimFreeCallFrame(interp, callFramePtr, JIM_FCF_NOHT); + } + /* Handle the JIM_EVAL return code */ + if (retcode == JIM_EVAL && interp->evalRetcodeLevel != interp->numLevels) { + int savedLevel = interp->evalRetcodeLevel; + + interp->evalRetcodeLevel = interp->numLevels; + while (retcode == JIM_EVAL) { + Jim_Obj *resultScriptObjPtr = Jim_GetResult(interp); + Jim_IncrRefCount(resultScriptObjPtr); + retcode = Jim_EvalObj(interp, resultScriptObjPtr); + Jim_DecrRefCount(interp, resultScriptObjPtr); + } + interp->evalRetcodeLevel = savedLevel; + } + /* Handle the JIM_RETURN return code */ + if (retcode == JIM_RETURN) { + retcode = interp->returnCode; + interp->returnCode = JIM_OK; + } + return retcode; +} + +int Jim_Eval_Named(Jim_Interp *interp, const char *script, const char *filename, int lineno) +{ + int retval; + Jim_Obj *scriptObjPtr; + + scriptObjPtr = Jim_NewStringObj(interp, script, -1); + Jim_IncrRefCount(scriptObjPtr); + + + if (filename) { + JimSetSourceInfo(interp, scriptObjPtr, filename, lineno); + } + + retval = Jim_EvalObj(interp, scriptObjPtr); + Jim_DecrRefCount(interp, scriptObjPtr); + return retval; +} + +int Jim_Eval(Jim_Interp *interp, const char *script) +{ + return Jim_Eval_Named(interp, script, NULL, 0); +} + + + +/* Execute script in the scope of the global level */ +int Jim_EvalGlobal(Jim_Interp *interp, const char *script) +{ + Jim_CallFrame *savedFramePtr; + int retval; + + savedFramePtr = interp->framePtr; + interp->framePtr = interp->topFramePtr; + retval = Jim_Eval(interp, script); + interp->framePtr = savedFramePtr; + return retval; +} + +int Jim_EvalObjBackground(Jim_Interp *interp, Jim_Obj *scriptObjPtr) +{ + Jim_CallFrame *savedFramePtr; + int retval; + + savedFramePtr = interp->framePtr; + interp->framePtr = interp->topFramePtr; + retval = Jim_EvalObj(interp, scriptObjPtr); + interp->framePtr = savedFramePtr; + /* Try to report the error (if any) via the bgerror proc */ + if (retval != JIM_OK) { + Jim_Obj *objv[2]; + + objv[0] = Jim_NewStringObj(interp, "bgerror", -1); + objv[1] = Jim_GetResult(interp); + Jim_IncrRefCount(objv[0]); + Jim_IncrRefCount(objv[1]); + if (Jim_EvalObjVector(interp, 2, objv) != JIM_OK) { + /* Report the error to stderr. */ + Jim_fprintf(interp, interp->cookie_stderr, "Background error:" JIM_NL); + Jim_PrintErrorMessage(interp); + } + Jim_DecrRefCount(interp, objv[0]); + Jim_DecrRefCount(interp, objv[1]); + } + return retval; +} + +int Jim_EvalFile(Jim_Interp *interp, const char *filename) +{ + char *prg = NULL; + FILE *fp; + int nread, totread, maxlen, buflen; + int retval; + Jim_Obj *scriptObjPtr; + + if ((fp = fopen(filename, "r")) == NULL) { + const int cwd_len = 2048; + char *cwd = malloc(cwd_len); + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + if (!getcwd(cwd, cwd_len)) strcpy(cwd, "unknown"); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "Error loading script \"", filename, "\"", + " cwd: ", cwd, + " err: ", strerror(errno), NULL); + free(cwd); + return JIM_ERR; + } + buflen = 1024; + maxlen = totread = 0; + while (1) { + if (maxlen < totread + buflen + 1) { + maxlen = totread + buflen + 1; + prg = Jim_Realloc(prg, maxlen); + } + /* do not use Jim_fread() - this is really a file */ + if ((nread = fread(prg + totread, 1, buflen, fp)) == 0) break; + totread += nread; + } + prg[totread] = '\0'; + /* do not use Jim_fclose() - this is really a file */ + fclose(fp); + + scriptObjPtr = Jim_NewStringObjNoAlloc(interp, prg, totread); + JimSetSourceInfo(interp, scriptObjPtr, filename, 1); + Jim_IncrRefCount(scriptObjPtr); + retval = Jim_EvalObj(interp, scriptObjPtr); + Jim_DecrRefCount(interp, scriptObjPtr); + return retval; +} + +/* ----------------------------------------------------------------------------- + * Subst + * ---------------------------------------------------------------------------*/ +static int JimParseSubstStr(struct JimParserCtx *pc) +{ + pc->tstart = pc->p; + pc->tline = pc->linenr; + while (*pc->p && *pc->p != '$' && *pc->p != '[') { + pc->p++; pc->len--; + } + pc->tend = pc->p-1; + pc->tt = JIM_TT_ESC; + return JIM_OK; +} + +static int JimParseSubst(struct JimParserCtx *pc, int flags) +{ + int retval; + + if (pc->len == 0) { + pc->tstart = pc->tend = pc->p; + pc->tline = pc->linenr; + pc->tt = JIM_TT_EOL; + pc->eof = 1; + return JIM_OK; + } + switch (*pc->p) { + case '[': + retval = JimParseCmd(pc); + if (flags & JIM_SUBST_NOCMD) { + pc->tstart--; + pc->tend++; + pc->tt = (flags & JIM_SUBST_NOESC) ? + JIM_TT_STR : JIM_TT_ESC; + } + return retval; + break; + case '$': + if (JimParseVar(pc) == JIM_ERR) { + pc->tstart = pc->tend = pc->p++; pc->len--; + pc->tline = pc->linenr; + pc->tt = JIM_TT_STR; + } else { + if (flags & JIM_SUBST_NOVAR) { + pc->tstart--; + if (flags & JIM_SUBST_NOESC) + pc->tt = JIM_TT_STR; + else + pc->tt = JIM_TT_ESC; + if (*pc->tstart == '{') { + pc->tstart--; + if (*(pc->tend + 1)) + pc->tend++; + } + } + } + break; + default: + retval = JimParseSubstStr(pc); + if (flags & JIM_SUBST_NOESC) + pc->tt = JIM_TT_STR; + return retval; + break; + } + return JIM_OK; +} + +/* The subst object type reuses most of the data structures and functions + * of the script object. Script's data structures are a bit more complex + * for what is needed for [subst]itution tasks, but the reuse helps to + * deal with a single data structure at the cost of some more memory + * usage for substitutions. */ +static Jim_ObjType substObjType = { + "subst", + FreeScriptInternalRep, + DupScriptInternalRep, + NULL, + JIM_TYPE_REFERENCES, +}; + +/* This method takes the string representation of an object + * as a Tcl string where to perform [subst]itution, and generates + * the pre-parsed internal representation. */ +int SetSubstFromAny(Jim_Interp *interp, struct Jim_Obj *objPtr, int flags) +{ + int scriptTextLen; + const char *scriptText = Jim_GetString(objPtr, &scriptTextLen); + struct JimParserCtx parser; + struct ScriptObj *script = Jim_Alloc(sizeof(*script)); + + script->len = 0; + script->csLen = 0; + script->commands = 0; + script->token = NULL; + script->cmdStruct = NULL; + script->inUse = 1; + script->substFlags = flags; + script->fileName = NULL; + + JimParserInit(&parser, scriptText, scriptTextLen, 1); + while (1) { + char *token; + int len, type, linenr; + + JimParseSubst(&parser, flags); + if (JimParserEof(&parser)) break; + token = JimParserGetToken(&parser, &len, &type, &linenr); + ScriptObjAddToken(interp, script, token, len, type, + NULL, linenr); + } + /* Free the old internal rep and set the new one. */ + Jim_FreeIntRep(interp, objPtr); + Jim_SetIntRepPtr(objPtr, script); + objPtr->typePtr = &scriptObjType; + return JIM_OK; +} + +ScriptObj *Jim_GetSubst(Jim_Interp *interp, Jim_Obj *objPtr, int flags) +{ + struct ScriptObj *script = Jim_GetIntRepPtr(objPtr); + + if (objPtr->typePtr != &substObjType || script->substFlags != flags) + SetSubstFromAny(interp, objPtr, flags); + return (ScriptObj*) Jim_GetIntRepPtr(objPtr); +} + +/* Performs commands,variables,blackslashes substitution, + * storing the result object (with refcount 0) into + * resObjPtrPtr. */ +int Jim_SubstObj(Jim_Interp *interp, Jim_Obj *substObjPtr, + Jim_Obj **resObjPtrPtr, int flags) +{ + ScriptObj *script; + ScriptToken *token; + int i, len, retcode = JIM_OK; + Jim_Obj *resObjPtr, *savedResultObjPtr; + + script = Jim_GetSubst(interp, substObjPtr, flags); +#ifdef JIM_OPTIMIZATION + /* Fast path for a very common case with array-alike syntax, + * that's: $foo($bar) */ + if (script->len == 1 && script->token[0].type == JIM_TT_VAR) { + Jim_Obj *varObjPtr = script->token[0].objPtr; + + Jim_IncrRefCount(varObjPtr); + resObjPtr = Jim_GetVariable(interp, varObjPtr, JIM_ERRMSG); + if (resObjPtr == NULL) { + Jim_DecrRefCount(interp, varObjPtr); + return JIM_ERR; + } + Jim_DecrRefCount(interp, varObjPtr); + *resObjPtrPtr = resObjPtr; + return JIM_OK; + } +#endif + + Jim_IncrRefCount(substObjPtr); /* Make sure it's shared. */ + /* In order to preserve the internal rep, we increment the + * inUse field of the script internal rep structure. */ + script->inUse++; + + token = script->token; + len = script->len; + + /* Save the interp old result, to set it again before + * to return. */ + savedResultObjPtr = interp->result; + Jim_IncrRefCount(savedResultObjPtr); + + /* Perform the substitution. Starts with an empty object + * and adds every token (performing the appropriate + * var/command/escape substitution). */ + resObjPtr = Jim_NewStringObj(interp, "", 0); + for (i = 0; i < len; i++) { + Jim_Obj *objPtr; + + switch (token[i].type) { + case JIM_TT_STR: + case JIM_TT_ESC: + Jim_AppendObj(interp, resObjPtr, token[i].objPtr); + break; + case JIM_TT_VAR: + objPtr = Jim_GetVariable(interp, token[i].objPtr, JIM_ERRMSG); + if (objPtr == NULL) goto err; + Jim_IncrRefCount(objPtr); + Jim_AppendObj(interp, resObjPtr, objPtr); + Jim_DecrRefCount(interp, objPtr); + break; + case JIM_TT_DICTSUGAR: + objPtr = Jim_ExpandDictSugar(interp, token[i].objPtr); + if (!objPtr) { + retcode = JIM_ERR; + goto err; + } + break; + case JIM_TT_CMD: + if (Jim_EvalObj(interp, token[i].objPtr) != JIM_OK) + goto err; + Jim_AppendObj(interp, resObjPtr, interp->result); + break; + default: + Jim_Panic(interp, + "default token type (%d) reached " + "in Jim_SubstObj().", token[i].type); + break; + } + } +ok: + if (retcode == JIM_OK) + Jim_SetResult(interp, savedResultObjPtr); + Jim_DecrRefCount(interp, savedResultObjPtr); + /* Note that we don't have to decrement inUse, because the + * following code transfers our use of the reference again to + * the script object. */ + Jim_FreeIntRep(interp, substObjPtr); + substObjPtr->typePtr = &scriptObjType; + Jim_SetIntRepPtr(substObjPtr, script); + Jim_DecrRefCount(interp, substObjPtr); + *resObjPtrPtr = resObjPtr; + return retcode; +err: + Jim_FreeNewObj(interp, resObjPtr); + retcode = JIM_ERR; + goto ok; +} + +/* ----------------------------------------------------------------------------- + * API Input/Export functions + * ---------------------------------------------------------------------------*/ + +int Jim_GetApi(Jim_Interp *interp, const char *funcname, void *targetPtrPtr) +{ + Jim_HashEntry *he; + + he = Jim_FindHashEntry(&interp->stub, funcname); + if (!he) + return JIM_ERR; + memcpy(targetPtrPtr, &he->val, sizeof(void*)); + return JIM_OK; +} + +int Jim_RegisterApi(Jim_Interp *interp, const char *funcname, void *funcptr) +{ + return Jim_AddHashEntry(&interp->stub, funcname, funcptr); +} + +#define JIM_REGISTER_API(name) \ + Jim_RegisterApi(interp, "Jim_" #name, (void *)Jim_ ## name) + +void JimRegisterCoreApi(Jim_Interp *interp) +{ + interp->getApiFuncPtr = Jim_GetApi; + JIM_REGISTER_API(Alloc); + JIM_REGISTER_API(Free); + JIM_REGISTER_API(Eval); + JIM_REGISTER_API(Eval_Named); + JIM_REGISTER_API(EvalGlobal); + JIM_REGISTER_API(EvalFile); + JIM_REGISTER_API(EvalObj); + JIM_REGISTER_API(EvalObjBackground); + JIM_REGISTER_API(EvalObjVector); + JIM_REGISTER_API(InitHashTable); + JIM_REGISTER_API(ExpandHashTable); + JIM_REGISTER_API(AddHashEntry); + JIM_REGISTER_API(ReplaceHashEntry); + JIM_REGISTER_API(DeleteHashEntry); + JIM_REGISTER_API(FreeHashTable); + JIM_REGISTER_API(FindHashEntry); + JIM_REGISTER_API(ResizeHashTable); + JIM_REGISTER_API(GetHashTableIterator); + JIM_REGISTER_API(NextHashEntry); + JIM_REGISTER_API(NewObj); + JIM_REGISTER_API(FreeObj); + JIM_REGISTER_API(InvalidateStringRep); + JIM_REGISTER_API(InitStringRep); + JIM_REGISTER_API(DuplicateObj); + JIM_REGISTER_API(GetString); + JIM_REGISTER_API(Length); + JIM_REGISTER_API(InvalidateStringRep); + JIM_REGISTER_API(NewStringObj); + JIM_REGISTER_API(NewStringObjNoAlloc); + JIM_REGISTER_API(AppendString); + JIM_REGISTER_API(AppendString_sprintf); + JIM_REGISTER_API(AppendObj); + JIM_REGISTER_API(AppendStrings); + JIM_REGISTER_API(StringEqObj); + JIM_REGISTER_API(StringMatchObj); + JIM_REGISTER_API(StringRangeObj); + JIM_REGISTER_API(FormatString); + JIM_REGISTER_API(CompareStringImmediate); + JIM_REGISTER_API(NewReference); + JIM_REGISTER_API(GetReference); + JIM_REGISTER_API(SetFinalizer); + JIM_REGISTER_API(GetFinalizer); + JIM_REGISTER_API(CreateInterp); + JIM_REGISTER_API(FreeInterp); + JIM_REGISTER_API(GetExitCode); + JIM_REGISTER_API(SetStdin); + JIM_REGISTER_API(SetStdout); + JIM_REGISTER_API(SetStderr); + JIM_REGISTER_API(CreateCommand); + JIM_REGISTER_API(CreateProcedure); + JIM_REGISTER_API(DeleteCommand); + JIM_REGISTER_API(RenameCommand); + JIM_REGISTER_API(GetCommand); + JIM_REGISTER_API(SetVariable); + JIM_REGISTER_API(SetVariableStr); + JIM_REGISTER_API(SetGlobalVariableStr); + JIM_REGISTER_API(SetVariableStrWithStr); + JIM_REGISTER_API(SetVariableLink); + JIM_REGISTER_API(GetVariable); + JIM_REGISTER_API(GetCallFrameByLevel); + JIM_REGISTER_API(Collect); + JIM_REGISTER_API(CollectIfNeeded); + JIM_REGISTER_API(GetIndex); + JIM_REGISTER_API(NewListObj); + JIM_REGISTER_API(ListAppendElement); + JIM_REGISTER_API(ListAppendList); + JIM_REGISTER_API(ListLength); + JIM_REGISTER_API(ListIndex); + JIM_REGISTER_API(SetListIndex); + JIM_REGISTER_API(ConcatObj); + JIM_REGISTER_API(NewDictObj); + JIM_REGISTER_API(DictKey); + JIM_REGISTER_API(DictKeysVector); + JIM_REGISTER_API(GetIndex); + JIM_REGISTER_API(GetReturnCode); + JIM_REGISTER_API(EvalExpression); + JIM_REGISTER_API(GetBoolFromExpr); + JIM_REGISTER_API(GetWide); + JIM_REGISTER_API(GetLong); + JIM_REGISTER_API(SetWide); + JIM_REGISTER_API(NewIntObj); + JIM_REGISTER_API(GetDouble); + JIM_REGISTER_API(SetDouble); + JIM_REGISTER_API(NewDoubleObj); + JIM_REGISTER_API(WrongNumArgs); + JIM_REGISTER_API(SetDictKeysVector); + JIM_REGISTER_API(SubstObj); + JIM_REGISTER_API(RegisterApi); + JIM_REGISTER_API(PrintErrorMessage); + JIM_REGISTER_API(InteractivePrompt); + JIM_REGISTER_API(RegisterCoreCommands); + JIM_REGISTER_API(GetSharedString); + JIM_REGISTER_API(ReleaseSharedString); + JIM_REGISTER_API(Panic); + JIM_REGISTER_API(StrDup); + JIM_REGISTER_API(UnsetVariable); + JIM_REGISTER_API(GetVariableStr); + JIM_REGISTER_API(GetGlobalVariable); + JIM_REGISTER_API(GetGlobalVariableStr); + JIM_REGISTER_API(GetAssocData); + JIM_REGISTER_API(SetAssocData); + JIM_REGISTER_API(DeleteAssocData); + JIM_REGISTER_API(GetEnum); + JIM_REGISTER_API(ScriptIsComplete); + JIM_REGISTER_API(PackageRequire); + JIM_REGISTER_API(PackageProvide); + JIM_REGISTER_API(InitStack); + JIM_REGISTER_API(FreeStack); + JIM_REGISTER_API(StackLen); + JIM_REGISTER_API(StackPush); + JIM_REGISTER_API(StackPop); + JIM_REGISTER_API(StackPeek); + JIM_REGISTER_API(FreeStackElements); + JIM_REGISTER_API(fprintf); + JIM_REGISTER_API(vfprintf); + JIM_REGISTER_API(fwrite); + JIM_REGISTER_API(fread); + JIM_REGISTER_API(fflush); + JIM_REGISTER_API(fgets); + JIM_REGISTER_API(GetNvp); + JIM_REGISTER_API(Nvp_name2value); + JIM_REGISTER_API(Nvp_name2value_simple); + JIM_REGISTER_API(Nvp_name2value_obj); + JIM_REGISTER_API(Nvp_name2value_nocase); + JIM_REGISTER_API(Nvp_name2value_obj_nocase); + + JIM_REGISTER_API(Nvp_value2name); + JIM_REGISTER_API(Nvp_value2name_simple); + JIM_REGISTER_API(Nvp_value2name_obj); + + JIM_REGISTER_API(GetOpt_Setup); + JIM_REGISTER_API(GetOpt_Debug); + JIM_REGISTER_API(GetOpt_Obj); + JIM_REGISTER_API(GetOpt_String); + JIM_REGISTER_API(GetOpt_Double); + JIM_REGISTER_API(GetOpt_Wide); + JIM_REGISTER_API(GetOpt_Nvp); + JIM_REGISTER_API(GetOpt_NvpUnknown); + JIM_REGISTER_API(GetOpt_Enum); + + JIM_REGISTER_API(Debug_ArgvString); + JIM_REGISTER_API(SetResult_sprintf); + JIM_REGISTER_API(SetResult_NvpUnknown); + +} + +/* ----------------------------------------------------------------------------- + * Core commands utility functions + * ---------------------------------------------------------------------------*/ +void Jim_WrongNumArgs(Jim_Interp *interp, int argc, Jim_Obj *const *argv, + const char *msg) +{ + int i; + Jim_Obj *objPtr = Jim_NewEmptyStringObj(interp); + + Jim_AppendString(interp, objPtr, "wrong # args: should be \"", -1); + for (i = 0; i < argc; i++) { + Jim_AppendObj(interp, objPtr, argv[i]); + if (!(i + 1 == argc && msg[0] == '\0')) + Jim_AppendString(interp, objPtr, " ", 1); + } + Jim_AppendString(interp, objPtr, msg, -1); + Jim_AppendString(interp, objPtr, "\"", 1); + Jim_SetResult(interp, objPtr); +} + +static Jim_Obj *JimCommandsList(Jim_Interp *interp, Jim_Obj *patternObjPtr) +{ + Jim_HashTableIterator *htiter; + Jim_HashEntry *he; + Jim_Obj *listObjPtr = Jim_NewListObj(interp, NULL, 0); + const char *pattern; + int patternLen=0; + + pattern = patternObjPtr ? Jim_GetString(patternObjPtr, &patternLen) : NULL; + htiter = Jim_GetHashTableIterator(&interp->commands); + while ((he = Jim_NextHashEntry(htiter)) != NULL) { + if (pattern && !JimStringMatch(pattern, patternLen, he->key, + strlen((const char*)he->key), 0)) + continue; + Jim_ListAppendElement(interp, listObjPtr, + Jim_NewStringObj(interp, he->key, -1)); + } + Jim_FreeHashTableIterator(htiter); + return listObjPtr; +} + +#define JIM_VARLIST_GLOBALS 0 +#define JIM_VARLIST_LOCALS 1 +#define JIM_VARLIST_VARS 2 + +static Jim_Obj *JimVariablesList(Jim_Interp *interp, Jim_Obj *patternObjPtr, + int mode) +{ + Jim_HashTableIterator *htiter; + Jim_HashEntry *he; + Jim_Obj *listObjPtr = Jim_NewListObj(interp, NULL, 0); + const char *pattern; + int patternLen=0; + + pattern = patternObjPtr ? Jim_GetString(patternObjPtr, &patternLen) : NULL; + if (mode == JIM_VARLIST_GLOBALS) { + htiter = Jim_GetHashTableIterator(&interp->topFramePtr->vars); + } else { + /* For [info locals], if we are at top level an emtpy list + * is returned. I don't agree, but we aim at compatibility (SS) */ + if (mode == JIM_VARLIST_LOCALS && + interp->framePtr == interp->topFramePtr) + return listObjPtr; + htiter = Jim_GetHashTableIterator(&interp->framePtr->vars); + } + while ((he = Jim_NextHashEntry(htiter)) != NULL) { + Jim_Var *varPtr = (Jim_Var*) he->val; + if (mode == JIM_VARLIST_LOCALS) { + if (varPtr->linkFramePtr != NULL) + continue; + } + if (pattern && !JimStringMatch(pattern, patternLen, he->key, + strlen((const char*)he->key), 0)) + continue; + Jim_ListAppendElement(interp, listObjPtr, + Jim_NewStringObj(interp, he->key, -1)); + } + Jim_FreeHashTableIterator(htiter); + return listObjPtr; +} + +static int JimInfoLevel(Jim_Interp *interp, Jim_Obj *levelObjPtr, + Jim_Obj **objPtrPtr) +{ + Jim_CallFrame *targetCallFrame; + + if (JimGetCallFrameByInteger(interp, levelObjPtr, &targetCallFrame) + != JIM_OK) + return JIM_ERR; + /* No proc call at toplevel callframe */ + if (targetCallFrame == interp->topFramePtr) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad level \"", + Jim_GetString(levelObjPtr, NULL), "\"", NULL); + return JIM_ERR; + } + *objPtrPtr = Jim_NewListObj(interp, + targetCallFrame->argv, + targetCallFrame->argc); + return JIM_OK; +} + +/* ----------------------------------------------------------------------------- + * Core commands + * ---------------------------------------------------------------------------*/ + +/* fake [puts] -- not the real puts, just for debugging. */ +static int Jim_PutsCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *str; + int len, nonewline = 0; + + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "-nonewline string"); + return JIM_ERR; + } + if (argc == 3) { + if (!Jim_CompareStringImmediate(interp, argv[1], "-nonewline")) + { + Jim_SetResultString(interp, "The second argument must " + "be -nonewline", -1); + return JIM_OK; + } else { + nonewline = 1; + argv++; + } + } + str = Jim_GetString(argv[1], &len); + Jim_fwrite(interp, str, 1, len, interp->cookie_stdout); + if (!nonewline) Jim_fprintf(interp, interp->cookie_stdout, JIM_NL); + return JIM_OK; +} + +/* Helper for [+] and [*] */ +static int Jim_AddMulHelper(Jim_Interp *interp, int argc, + Jim_Obj *const *argv, int op) +{ + jim_wide wideValue, res; + double doubleValue, doubleRes; + int i; + + res = (op == JIM_EXPROP_ADD) ? 0 : 1; + + for (i = 1; i < argc; i++) { + if (Jim_GetWide(interp, argv[i], &wideValue) != JIM_OK) + goto trydouble; + if (op == JIM_EXPROP_ADD) + res += wideValue; + else + res *= wideValue; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, res)); + return JIM_OK; +trydouble: + doubleRes = (double) res; + for (;i < argc; i++) { + if (Jim_GetDouble(interp, argv[i], &doubleValue) != JIM_OK) + return JIM_ERR; + if (op == JIM_EXPROP_ADD) + doubleRes += doubleValue; + else + doubleRes *= doubleValue; + } + Jim_SetResult(interp, Jim_NewDoubleObj(interp, doubleRes)); + return JIM_OK; +} + +/* Helper for [-] and [/] */ +static int Jim_SubDivHelper(Jim_Interp *interp, int argc, + Jim_Obj *const *argv, int op) +{ + jim_wide wideValue, res = 0; + double doubleValue, doubleRes = 0; + int i = 2; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "number ?number ... number?"); + return JIM_ERR; + } else if (argc == 2) { + /* The arity = 2 case is different. For [- x] returns -x, + * while [/ x] returns 1/x. */ + if (Jim_GetWide(interp, argv[1], &wideValue) != JIM_OK) { + if (Jim_GetDouble(interp, argv[1], &doubleValue) != + JIM_OK) + { + return JIM_ERR; + } else { + if (op == JIM_EXPROP_SUB) + doubleRes = -doubleValue; + else + doubleRes = 1.0/doubleValue; + Jim_SetResult(interp, Jim_NewDoubleObj(interp, + doubleRes)); + return JIM_OK; + } + } + if (op == JIM_EXPROP_SUB) { + res = -wideValue; + Jim_SetResult(interp, Jim_NewIntObj(interp, res)); + } else { + doubleRes = 1.0/wideValue; + Jim_SetResult(interp, Jim_NewDoubleObj(interp, + doubleRes)); + } + return JIM_OK; + } else { + if (Jim_GetWide(interp, argv[1], &res) != JIM_OK) { + if (Jim_GetDouble(interp, argv[1], &doubleRes) + != JIM_OK) { + return JIM_ERR; + } else { + goto trydouble; + } + } + } + for (i = 2; i < argc; i++) { + if (Jim_GetWide(interp, argv[i], &wideValue) != JIM_OK) { + doubleRes = (double) res; + goto trydouble; + } + if (op == JIM_EXPROP_SUB) + res -= wideValue; + else + res /= wideValue; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, res)); + return JIM_OK; +trydouble: + for (;i < argc; i++) { + if (Jim_GetDouble(interp, argv[i], &doubleValue) != JIM_OK) + return JIM_ERR; + if (op == JIM_EXPROP_SUB) + doubleRes -= doubleValue; + else + doubleRes /= doubleValue; + } + Jim_SetResult(interp, Jim_NewDoubleObj(interp, doubleRes)); + return JIM_OK; +} + + +/* [+] */ +static int Jim_AddCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return Jim_AddMulHelper(interp, argc, argv, JIM_EXPROP_ADD); +} + +/* [*] */ +static int Jim_MulCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return Jim_AddMulHelper(interp, argc, argv, JIM_EXPROP_MUL); +} + +/* [-] */ +static int Jim_SubCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return Jim_SubDivHelper(interp, argc, argv, JIM_EXPROP_SUB); +} + +/* [/] */ +static int Jim_DivCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return Jim_SubDivHelper(interp, argc, argv, JIM_EXPROP_DIV); +} + +/* [set] */ +static int Jim_SetCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?newValue?"); + return JIM_ERR; + } + if (argc == 2) { + Jim_Obj *objPtr; + objPtr = Jim_GetVariable(interp, argv[1], JIM_ERRMSG); + if (!objPtr) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; + } + /* argc == 3 case. */ + if (Jim_SetVariable(interp, argv[1], argv[2]) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, argv[2]); + return JIM_OK; +} + +/* [unset] */ +static int Jim_UnsetCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int i; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?varName ...?"); + return JIM_ERR; + } + for (i = 1; i < argc; i++) { + if (Jim_UnsetVariable(interp, argv[i], JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + } + return JIM_OK; +} + +/* [incr] */ +static int Jim_IncrCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + jim_wide wideValue, increment = 1; + Jim_Obj *intObjPtr; + + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?increment?"); + return JIM_ERR; + } + if (argc == 3) { + if (Jim_GetWide(interp, argv[2], &increment) != JIM_OK) + return JIM_ERR; + } + intObjPtr = Jim_GetVariable(interp, argv[1], JIM_ERRMSG); + if (!intObjPtr) return JIM_ERR; + if (Jim_GetWide(interp, intObjPtr, &wideValue) != JIM_OK) + return JIM_ERR; + if (Jim_IsShared(intObjPtr)) { + intObjPtr = Jim_NewIntObj(interp, wideValue + increment); + if (Jim_SetVariable(interp, argv[1], intObjPtr) != JIM_OK) { + Jim_FreeNewObj(interp, intObjPtr); + return JIM_ERR; + } + } else { + Jim_SetWide(interp, intObjPtr, wideValue + increment); + /* The following step is required in order to invalidate the + * string repr of "FOO" if the var name is on the form of "FOO(IDX)" */ + if (Jim_SetVariable(interp, argv[1], intObjPtr) != JIM_OK) { + return JIM_ERR; + } + } + Jim_SetResult(interp, intObjPtr); + return JIM_OK; +} + +/* [while] */ +static int Jim_WhileCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "condition body"); + return JIM_ERR; + } + /* Try to run a specialized version of while if the expression + * is in one of the following forms: + * + * $a < CONST, $a < $b + * $a <= CONST, $a <= $b + * $a > CONST, $a > $b + * $a >= CONST, $a >= $b + * $a != CONST, $a != $b + * $a == CONST, $a == $b + * $a + * !$a + * CONST + */ + +#ifdef JIM_OPTIMIZATION + { + ExprByteCode *expr; + Jim_Obj *varAObjPtr = NULL, *varBObjPtr = NULL, *objPtr; + int exprLen, retval; + + /* STEP 1 -- Check if there are the conditions to run the specialized + * version of while */ + + if ((expr = Jim_GetExpression(interp, argv[1])) == NULL) goto noopt; + if (expr->len <= 0 || expr->len > 3) goto noopt; + switch (expr->len) { + case 1: + if (expr->opcode[0] != JIM_EXPROP_VARIABLE && + expr->opcode[0] != JIM_EXPROP_NUMBER) + goto noopt; + break; + case 2: + if (expr->opcode[1] != JIM_EXPROP_NOT || + expr->opcode[0] != JIM_EXPROP_VARIABLE) + goto noopt; + break; + case 3: + if (expr->opcode[0] != JIM_EXPROP_VARIABLE || + (expr->opcode[1] != JIM_EXPROP_NUMBER && + expr->opcode[1] != JIM_EXPROP_VARIABLE)) + goto noopt; + switch (expr->opcode[2]) { + case JIM_EXPROP_LT: + case JIM_EXPROP_LTE: + case JIM_EXPROP_GT: + case JIM_EXPROP_GTE: + case JIM_EXPROP_NUMEQ: + case JIM_EXPROP_NUMNE: + /* nothing to do */ + break; + default: + goto noopt; + } + break; + default: + Jim_Panic(interp, + "Unexpected default reached in Jim_WhileCoreCommand()"); + break; + } + + /* STEP 2 -- conditions meet. Initialization. Take different + * branches for different expression lengths. */ + exprLen = expr->len; + + if (exprLen == 1) { + jim_wide wideValue=0; + + if (expr->opcode[0] == JIM_EXPROP_VARIABLE) { + varAObjPtr = expr->obj[0]; + Jim_IncrRefCount(varAObjPtr); + } else { + if (Jim_GetWide(interp, expr->obj[0], &wideValue) != JIM_OK) + goto noopt; + } + while (1) { + if (varAObjPtr) { + if (!(objPtr = + Jim_GetVariable(interp, varAObjPtr, JIM_NONE)) || + Jim_GetWide(interp, objPtr, &wideValue) != JIM_OK) + { + Jim_DecrRefCount(interp, varAObjPtr); + goto noopt; + } + } + if (!wideValue) break; + if ((retval = Jim_EvalObj(interp, argv[2])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + if (varAObjPtr) + Jim_DecrRefCount(interp, varAObjPtr); + goto out; + break; + case JIM_CONTINUE: + continue; + break; + default: + if (varAObjPtr) + Jim_DecrRefCount(interp, varAObjPtr); + return retval; + } + } + } + if (varAObjPtr) + Jim_DecrRefCount(interp, varAObjPtr); + } else if (exprLen == 3) { + jim_wide wideValueA, wideValueB=0, cmpRes = 0; + int cmpType = expr->opcode[2]; + + varAObjPtr = expr->obj[0]; + Jim_IncrRefCount(varAObjPtr); + if (expr->opcode[1] == JIM_EXPROP_VARIABLE) { + varBObjPtr = expr->obj[1]; + Jim_IncrRefCount(varBObjPtr); + } else { + if (Jim_GetWide(interp, expr->obj[1], &wideValueB) != JIM_OK) + goto noopt; + } + while (1) { + if (!(objPtr = Jim_GetVariable(interp, varAObjPtr, JIM_NONE)) || + Jim_GetWide(interp, objPtr, &wideValueA) != JIM_OK) + { + Jim_DecrRefCount(interp, varAObjPtr); + if (varBObjPtr) + Jim_DecrRefCount(interp, varBObjPtr); + goto noopt; + } + if (varBObjPtr) { + if (!(objPtr = + Jim_GetVariable(interp, varBObjPtr, JIM_NONE)) || + Jim_GetWide(interp, objPtr, &wideValueB) != JIM_OK) + { + Jim_DecrRefCount(interp, varAObjPtr); + Jim_DecrRefCount(interp, varBObjPtr); + goto noopt; + } + } + switch (cmpType) { + case JIM_EXPROP_LT: + cmpRes = wideValueA < wideValueB; break; + case JIM_EXPROP_LTE: + cmpRes = wideValueA <= wideValueB; break; + case JIM_EXPROP_GT: + cmpRes = wideValueA > wideValueB; break; + case JIM_EXPROP_GTE: + cmpRes = wideValueA >= wideValueB; break; + case JIM_EXPROP_NUMEQ: + cmpRes = wideValueA == wideValueB; break; + case JIM_EXPROP_NUMNE: + cmpRes = wideValueA != wideValueB; break; + } + if (!cmpRes) break; + if ((retval = Jim_EvalObj(interp, argv[2])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + Jim_DecrRefCount(interp, varAObjPtr); + if (varBObjPtr) + Jim_DecrRefCount(interp, varBObjPtr); + goto out; + break; + case JIM_CONTINUE: + continue; + break; + default: + Jim_DecrRefCount(interp, varAObjPtr); + if (varBObjPtr) + Jim_DecrRefCount(interp, varBObjPtr); + return retval; + } + } + } + Jim_DecrRefCount(interp, varAObjPtr); + if (varBObjPtr) + Jim_DecrRefCount(interp, varBObjPtr); + } else { + /* TODO: case for len == 2 */ + goto noopt; + } + Jim_SetEmptyResult(interp); + return JIM_OK; + } +noopt: +#endif + + /* The general purpose implementation of while starts here */ + while (1) { + int boolean, retval; + + if ((retval = Jim_GetBoolFromExpr(interp, argv[1], + &boolean)) != JIM_OK) + return retval; + if (!boolean) break; + if ((retval = Jim_EvalObj(interp, argv[2])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + goto out; + break; + case JIM_CONTINUE: + continue; + break; + default: + return retval; + } + } + } +out: + Jim_SetEmptyResult(interp); + return JIM_OK; +} + +/* [for] */ +static int Jim_ForCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int retval; + + if (argc != 5) { + Jim_WrongNumArgs(interp, 1, argv, "start test next body"); + return JIM_ERR; + } + /* Check if the for is on the form: + * for {set i CONST} {$i < CONST} {incr i} + * for {set i CONST} {$i < $j} {incr i} + * for {set i CONST} {$i <= CONST} {incr i} + * for {set i CONST} {$i <= $j} {incr i} + * XXX: NOTE: if variable traces are implemented, this optimization + * need to be modified to check for the proc epoch at every variable + * update. */ +#ifdef JIM_OPTIMIZATION + { + ScriptObj *initScript, *incrScript; + ExprByteCode *expr; + jim_wide start, stop=0, currentVal; + unsigned jim_wide procEpoch = interp->procEpoch; + Jim_Obj *varNamePtr, *stopVarNamePtr = NULL, *objPtr; + int cmpType; + struct Jim_Cmd *cmdPtr; + + /* Do it only if there aren't shared arguments */ + if (argv[1] == argv[2] || argv[2] == argv[3] || argv[1] == argv[3]) + goto evalstart; + initScript = Jim_GetScript(interp, argv[1]); + expr = Jim_GetExpression(interp, argv[2]); + incrScript = Jim_GetScript(interp, argv[3]); + + /* Ensure proper lengths to start */ + if (initScript->len != 6) goto evalstart; + if (incrScript->len != 4) goto evalstart; + if (expr->len != 3) goto evalstart; + /* Ensure proper token types. */ + if (initScript->token[2].type != JIM_TT_ESC || + initScript->token[4].type != JIM_TT_ESC || + incrScript->token[2].type != JIM_TT_ESC || + expr->opcode[0] != JIM_EXPROP_VARIABLE || + (expr->opcode[1] != JIM_EXPROP_NUMBER && + expr->opcode[1] != JIM_EXPROP_VARIABLE) || + (expr->opcode[2] != JIM_EXPROP_LT && + expr->opcode[2] != JIM_EXPROP_LTE)) + goto evalstart; + cmpType = expr->opcode[2]; + /* Initialization command must be [set] */ + cmdPtr = Jim_GetCommand(interp, initScript->token[0].objPtr, JIM_NONE); + if (cmdPtr == NULL || cmdPtr->cmdProc != Jim_SetCoreCommand) + goto evalstart; + /* Update command must be incr */ + cmdPtr = Jim_GetCommand(interp, incrScript->token[0].objPtr, JIM_NONE); + if (cmdPtr == NULL || cmdPtr->cmdProc != Jim_IncrCoreCommand) + goto evalstart; + /* set, incr, expression must be about the same variable */ + if (!Jim_StringEqObj(initScript->token[2].objPtr, + incrScript->token[2].objPtr, 0)) + goto evalstart; + if (!Jim_StringEqObj(initScript->token[2].objPtr, + expr->obj[0], 0)) + goto evalstart; + /* Check that the initialization and comparison are valid integers */ + if (Jim_GetWide(interp, initScript->token[4].objPtr, &start) == JIM_ERR) + goto evalstart; + if (expr->opcode[1] == JIM_EXPROP_NUMBER && + Jim_GetWide(interp, expr->obj[1], &stop) == JIM_ERR) + { + goto evalstart; + } + + /* Initialization */ + varNamePtr = expr->obj[0]; + if (expr->opcode[1] == JIM_EXPROP_VARIABLE) { + stopVarNamePtr = expr->obj[1]; + Jim_IncrRefCount(stopVarNamePtr); + } + Jim_IncrRefCount(varNamePtr); + + /* --- OPTIMIZED FOR --- */ + /* Start to loop */ + objPtr = Jim_NewIntObj(interp, start); + if (Jim_SetVariable(interp, varNamePtr, objPtr) != JIM_OK) { + Jim_DecrRefCount(interp, varNamePtr); + if (stopVarNamePtr) Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_FreeNewObj(interp, objPtr); + goto evalstart; + } + while (1) { + /* === Check condition === */ + /* Common code: */ + objPtr = Jim_GetVariable(interp, varNamePtr, JIM_NONE); + if (objPtr == NULL || + Jim_GetWide(interp, objPtr, ¤tVal) != JIM_OK) + { + Jim_DecrRefCount(interp, varNamePtr); + if (stopVarNamePtr) Jim_DecrRefCount(interp, stopVarNamePtr); + goto testcond; + } + /* Immediate or Variable? get the 'stop' value if the latter. */ + if (stopVarNamePtr) { + objPtr = Jim_GetVariable(interp, stopVarNamePtr, JIM_NONE); + if (objPtr == NULL || + Jim_GetWide(interp, objPtr, &stop) != JIM_OK) + { + Jim_DecrRefCount(interp, varNamePtr); + Jim_DecrRefCount(interp, stopVarNamePtr); + goto testcond; + } + } + if (cmpType == JIM_EXPROP_LT) { + if (currentVal >= stop) break; + } else { + if (currentVal > stop) break; + } + /* Eval body */ + if ((retval = Jim_EvalObj(interp, argv[4])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + goto out; + case JIM_CONTINUE: + /* nothing to do */ + break; + default: + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + return retval; + } + } + /* If there was a change in procedures/command continue + * with the usual [for] command implementation */ + if (procEpoch != interp->procEpoch) { + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + goto evalnext; + } + /* Increment */ + objPtr = Jim_GetVariable(interp, varNamePtr, JIM_ERRMSG); + if (objPtr->refCount == 1 && objPtr->typePtr == &intObjType) { + objPtr->internalRep.wideValue ++; + Jim_InvalidateStringRep(objPtr); + } else { + Jim_Obj *auxObjPtr; + + if (Jim_GetWide(interp, objPtr, ¤tVal) == JIM_ERR) { + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + goto evalnext; + } + auxObjPtr = Jim_NewIntObj(interp, currentVal + 1); + if (Jim_SetVariable(interp, varNamePtr, auxObjPtr) == JIM_ERR) { + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + Jim_FreeNewObj(interp, auxObjPtr); + goto evalnext; + } + } + } + if (stopVarNamePtr) + Jim_DecrRefCount(interp, stopVarNamePtr); + Jim_DecrRefCount(interp, varNamePtr); + Jim_SetEmptyResult(interp); + return JIM_OK; + } +#endif +evalstart: + /* Eval start */ + if ((retval = Jim_EvalObj(interp, argv[1])) != JIM_OK) + return retval; + while (1) { + int boolean; +testcond: + /* Test the condition */ + if ((retval = Jim_GetBoolFromExpr(interp, argv[2], &boolean)) + != JIM_OK) + return retval; + if (!boolean) break; + /* Eval body */ + if ((retval = Jim_EvalObj(interp, argv[4])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + goto out; + break; + case JIM_CONTINUE: + /* Nothing to do */ + break; + default: + return retval; + } + } +evalnext: + /* Eval next */ + if ((retval = Jim_EvalObj(interp, argv[3])) != JIM_OK) { + switch (retval) { + case JIM_BREAK: + goto out; + break; + case JIM_CONTINUE: + continue; + break; + default: + return retval; + } + } + } +out: + Jim_SetEmptyResult(interp); + return JIM_OK; +} + +/* foreach + lmap implementation. */ +static int JimForeachMapHelper(Jim_Interp *interp, int argc, + Jim_Obj *const *argv, int doMap) +{ + int result = JIM_ERR, i, nbrOfLists, *listsIdx, *listsEnd; + int nbrOfLoops = 0; + Jim_Obj *emptyStr, *script, *mapRes = NULL; + + if (argc < 4 || argc % 2 != 0) { + Jim_WrongNumArgs(interp, 1, argv, "varList list ?varList list ...? script"); + return JIM_ERR; + } + if (doMap) { + mapRes = Jim_NewListObj(interp, NULL, 0); + Jim_IncrRefCount(mapRes); + } + emptyStr = Jim_NewEmptyStringObj(interp); + Jim_IncrRefCount(emptyStr); + script = argv[argc-1]; /* Last argument is a script */ + nbrOfLists = (argc - 1 - 1) / 2; /* argc - 'foreach' - script */ + listsIdx = (int*)Jim_Alloc(nbrOfLists * sizeof(int)); + listsEnd = (int*)Jim_Alloc(nbrOfLists*2 * sizeof(int)); + /* Initialize iterators and remember max nbr elements each list */ + memset(listsIdx, 0, nbrOfLists * sizeof(int)); + /* Remember lengths of all lists and calculate how much rounds to loop */ + for (i = 0; i < nbrOfLists*2; i += 2) { + div_t cnt; + int count; + Jim_ListLength(interp, argv[i + 1], &listsEnd[i]); + Jim_ListLength(interp, argv[i + 2], &listsEnd[i + 1]); + if (listsEnd[i] == 0) { + Jim_SetResultString(interp, "foreach varlist is empty", -1); + goto err; + } + cnt = div(listsEnd[i + 1], listsEnd[i]); + count = cnt.quot + (cnt.rem ? 1 : 0); + if (count > nbrOfLoops) + nbrOfLoops = count; + } + for (; nbrOfLoops-- > 0;) { + for (i = 0; i < nbrOfLists; ++i) { + int varIdx = 0, var = i * 2; + while (varIdx < listsEnd[var]) { + Jim_Obj *varName, *ele; + int lst = i * 2 + 1; + if (Jim_ListIndex(interp, argv[var + 1], varIdx, &varName, JIM_ERRMSG) + != JIM_OK) + goto err; + if (listsIdx[i] < listsEnd[lst]) { + if (Jim_ListIndex(interp, argv[lst + 1], listsIdx[i], &ele, JIM_ERRMSG) + != JIM_OK) + goto err; + if (Jim_SetVariable(interp, varName, ele) != JIM_OK) { + Jim_SetResultString(interp, "couldn't set loop variable: ", -1); + goto err; + } + ++listsIdx[i]; /* Remember next iterator of current list */ + } else if (Jim_SetVariable(interp, varName, emptyStr) != JIM_OK) { + Jim_SetResultString(interp, "couldn't set loop variable: ", -1); + goto err; + } + ++varIdx; /* Next variable */ + } + } + switch (result = Jim_EvalObj(interp, script)) { + case JIM_OK: + if (doMap) + Jim_ListAppendElement(interp, mapRes, interp->result); + break; + case JIM_CONTINUE: + break; + case JIM_BREAK: + goto out; + break; + default: + goto err; + } + } +out: + result = JIM_OK; + if (doMap) + Jim_SetResult(interp, mapRes); + else + Jim_SetEmptyResult(interp); +err: + if (doMap) + Jim_DecrRefCount(interp, mapRes); + Jim_DecrRefCount(interp, emptyStr); + Jim_Free(listsIdx); + Jim_Free(listsEnd); + return result; +} + +/* [foreach] */ +static int Jim_ForeachCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return JimForeachMapHelper(interp, argc, argv, 0); +} + +/* [lmap] */ +static int Jim_LmapCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + return JimForeachMapHelper(interp, argc, argv, 1); +} + +/* [if] */ +static int Jim_IfCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int boolean, retval, current = 1, falsebody = 0; + if (argc >= 3) { + while (1) { + /* Far not enough arguments given! */ + if (current >= argc) goto err; + if ((retval = Jim_GetBoolFromExpr(interp, + argv[current++], &boolean)) + != JIM_OK) + return retval; + /* There lacks something, isn't it? */ + if (current >= argc) goto err; + if (Jim_CompareStringImmediate(interp, argv[current], + "then")) current++; + /* Tsk tsk, no then-clause? */ + if (current >= argc) goto err; + if (boolean) + return Jim_EvalObj(interp, argv[current]); + /* Ok: no else-clause follows */ + if (++current >= argc) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + return JIM_OK; + } + falsebody = current++; + if (Jim_CompareStringImmediate(interp, argv[falsebody], + "else")) { + /* IIICKS - else-clause isn't last cmd? */ + if (current != argc-1) goto err; + return Jim_EvalObj(interp, argv[current]); + } else if (Jim_CompareStringImmediate(interp, + argv[falsebody], "elseif")) + /* Ok: elseif follows meaning all the stuff + * again (how boring...) */ + continue; + /* OOPS - else-clause is not last cmd?*/ + else if (falsebody != argc-1) + goto err; + return Jim_EvalObj(interp, argv[falsebody]); + } + return JIM_OK; + } +err: + Jim_WrongNumArgs(interp, 1, argv, "condition ?then? trueBody ?elseif ...? ?else? falseBody"); + return JIM_ERR; +} + +enum {SWITCH_EXACT, SWITCH_GLOB, SWITCH_RE, SWITCH_CMD, SWITCH_UNKNOWN}; + +/* [switch] */ +static int Jim_SwitchCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int retcode = JIM_ERR, matchOpt = SWITCH_EXACT, opt = 1, patCount, i; + Jim_Obj *command = 0, *const *caseList = 0, *strObj; + Jim_Obj *script = 0; + if (argc < 3) goto wrongnumargs; + for (opt = 1; opt < argc; ++opt) { + const char *option = Jim_GetString(argv[opt], 0); + if (*option != '-') break; + else if (strncmp(option, "--", 2) == 0) { ++opt; break; } + else if (strncmp(option, "-exact", 2) == 0) matchOpt = SWITCH_EXACT; + else if (strncmp(option, "-glob", 2) == 0) matchOpt = SWITCH_GLOB; + else if (strncmp(option, "-regexp", 2) == 0) matchOpt = SWITCH_RE; + else if (strncmp(option, "-command", 2) == 0) { matchOpt = SWITCH_CMD; + if ((argc - opt) < 2) goto wrongnumargs; + command = argv[++opt]; + } else { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad option \"", option, "\": must be -exact, -glob, " + "-regexp, -command procname or --", 0); + goto err; + } + if ((argc - opt) < 2) goto wrongnumargs; + } + strObj = argv[opt++]; + patCount = argc - opt; + if (patCount == 1) { + Jim_Obj **vector; + JimListGetElements(interp, argv[opt], &patCount, &vector); + caseList = vector; + } else + caseList = &argv[opt]; + if (patCount == 0 || patCount % 2 != 0) goto wrongnumargs; + for (i = 0; script == 0 && i < patCount; i += 2) { + Jim_Obj *patObj = caseList[i]; + if (!Jim_CompareStringImmediate(interp, patObj, "default") + || i < (patCount-2)) { + switch (matchOpt) { + case SWITCH_EXACT: + if (Jim_StringEqObj(strObj, patObj, 0)) + script = caseList[i + 1]; + break; + case SWITCH_GLOB: + if (Jim_StringMatchObj(patObj, strObj, 0)) + script = caseList[i + 1]; + break; + case SWITCH_RE: + command = Jim_NewStringObj(interp, "regexp", -1); + /* Fall thru intentionally */ + case SWITCH_CMD: { + Jim_Obj *parms[] = {command, patObj, strObj}; + int rc = Jim_EvalObjVector(interp, 3, parms); + long matching; + /* After the execution of a command we need to + * make sure to reconvert the object into a list + * again. Only for the single-list style [switch]. */ + if (argc-opt == 1) { + Jim_Obj **vector; + JimListGetElements(interp, argv[opt], &patCount, + &vector); + caseList = vector; + } + /* command is here already decref'd */ + if (rc != JIM_OK) { + retcode = rc; + goto err; + } + rc = Jim_GetLong(interp, Jim_GetResult(interp), &matching); + if (rc != JIM_OK) { + retcode = rc; + goto err; + } + if (matching) + script = caseList[i + 1]; + break; + } + default: + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "internal error: no such option implemented", 0); + goto err; + } + } else { + script = caseList[i + 1]; + } + } + for (; i < patCount && Jim_CompareStringImmediate(interp, script, "-"); + i += 2) + script = caseList[i + 1]; + if (script && Jim_CompareStringImmediate(interp, script, "-")) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "no body specified for pattern \"", + Jim_GetString(caseList[i-2], 0), "\"", 0); + goto err; + } + retcode = JIM_OK; + Jim_SetEmptyResult(interp); + if (script != 0) + retcode = Jim_EvalObj(interp, script); + return retcode; +wrongnumargs: + Jim_WrongNumArgs(interp, 1, argv, "?options? string " + "pattern body ... ?default body? or " + "{pattern body ?pattern body ...?}"); +err: + return retcode; +} + +/* [list] */ +static int Jim_ListCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *listObjPtr; + + listObjPtr = Jim_NewListObj(interp, argv + 1, argc-1); + Jim_SetResult(interp, listObjPtr); + return JIM_OK; +} + +/* [lindex] */ +static int Jim_LindexCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *objPtr, *listObjPtr; + int i; + int index_t; + + if (argc < 3) { + Jim_WrongNumArgs(interp, 1, argv, "list index ?...?"); + return JIM_ERR; + } + objPtr = argv[1]; + Jim_IncrRefCount(objPtr); + for (i = 2; i < argc; i++) { + listObjPtr = objPtr; + if (Jim_GetIndex(interp, argv[i], &index_t) != JIM_OK) { + Jim_DecrRefCount(interp, listObjPtr); + return JIM_ERR; + } + if (Jim_ListIndex(interp, listObjPtr, index_t, &objPtr, + JIM_NONE) != JIM_OK) { + /* Returns an empty object if the index + * is out of range. */ + Jim_DecrRefCount(interp, listObjPtr); + Jim_SetEmptyResult(interp); + return JIM_OK; + } + Jim_IncrRefCount(objPtr); + Jim_DecrRefCount(interp, listObjPtr); + } + Jim_SetResult(interp, objPtr); + Jim_DecrRefCount(interp, objPtr); + return JIM_OK; +} + +/* [llength] */ +static int Jim_LlengthCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int len; + + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "list"); + return JIM_ERR; + } + Jim_ListLength(interp, argv[1], &len); + Jim_SetResult(interp, Jim_NewIntObj(interp, len)); + return JIM_OK; +} + +/* [lappend] */ +static int Jim_LappendCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *listObjPtr; + int shared, i; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?value value ...?"); + return JIM_ERR; + } + listObjPtr = Jim_GetVariable(interp, argv[1], JIM_NONE); + if (!listObjPtr) { + /* Create the list if it does not exists */ + listObjPtr = Jim_NewListObj(interp, NULL, 0); + if (Jim_SetVariable(interp, argv[1], listObjPtr) != JIM_OK) { + Jim_FreeNewObj(interp, listObjPtr); + return JIM_ERR; + } + } + shared = Jim_IsShared(listObjPtr); + if (shared) + listObjPtr = Jim_DuplicateObj(interp, listObjPtr); + for (i = 2; i < argc; i++) + Jim_ListAppendElement(interp, listObjPtr, argv[i]); + if (Jim_SetVariable(interp, argv[1], listObjPtr) != JIM_OK) { + if (shared) + Jim_FreeNewObj(interp, listObjPtr); + return JIM_ERR; + } + Jim_SetResult(interp, listObjPtr); + return JIM_OK; +} + +/* [linsert] */ +static int Jim_LinsertCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int index_t, len; + Jim_Obj *listPtr; + + if (argc < 4) { + Jim_WrongNumArgs(interp, 1, argv, "list index element " + "?element ...?"); + return JIM_ERR; + } + listPtr = argv[1]; + if (Jim_IsShared(listPtr)) + listPtr = Jim_DuplicateObj(interp, listPtr); + if (Jim_GetIndex(interp, argv[2], &index_t) != JIM_OK) + goto err; + Jim_ListLength(interp, listPtr, &len); + if (index_t >= len) + index_t = len; + else if (index_t < 0) + index_t = len + index_t + 1; + Jim_ListInsertElements(interp, listPtr, index_t, argc-3, &argv[3]); + Jim_SetResult(interp, listPtr); + return JIM_OK; +err: + if (listPtr != argv[1]) { + Jim_FreeNewObj(interp, listPtr); + } + return JIM_ERR; +} + +/* [lset] */ +static int Jim_LsetCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc < 3) { + Jim_WrongNumArgs(interp, 1, argv, "listVar ?index...? newVal"); + return JIM_ERR; + } else if (argc == 3) { + if (Jim_SetVariable(interp, argv[1], argv[2]) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, argv[2]); + return JIM_OK; + } + if (Jim_SetListIndex(interp, argv[1], argv + 2, argc-3, argv[argc-1]) + == JIM_ERR) return JIM_ERR; + return JIM_OK; +} + +/* [lsort] */ +static int Jim_LsortCoreCommand(Jim_Interp *interp, int argc, Jim_Obj *const argv[]) +{ + const char *options[] = { + "-ascii", "-nocase", "-increasing", "-decreasing", NULL + }; + enum {OPT_ASCII, OPT_NOCASE, OPT_INCREASING, OPT_DECREASING}; + Jim_Obj *resObj; + int i, lsortType = JIM_LSORT_ASCII; /* default sort type */ + int decreasing = 0; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "?options? list"); + return JIM_ERR; + } + for (i = 1; i < (argc-1); i++) { + int option; + + if (Jim_GetEnum(interp, argv[i], options, &option, "option", JIM_ERRMSG) + != JIM_OK) + return JIM_ERR; + switch (option) { + case OPT_ASCII: lsortType = JIM_LSORT_ASCII; break; + case OPT_NOCASE: lsortType = JIM_LSORT_NOCASE; break; + case OPT_INCREASING: decreasing = 0; break; + case OPT_DECREASING: decreasing = 1; break; + } + } + if (decreasing) { + switch (lsortType) { + case JIM_LSORT_ASCII: lsortType = JIM_LSORT_ASCII_DECR; break; + case JIM_LSORT_NOCASE: lsortType = JIM_LSORT_NOCASE_DECR; break; + } + } + resObj = Jim_DuplicateObj(interp, argv[argc-1]); + ListSortElements(interp, resObj, lsortType); + Jim_SetResult(interp, resObj); + return JIM_OK; +} + +/* [append] */ +static int Jim_AppendCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *stringObjPtr; + int shared, i; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?value value ...?"); + return JIM_ERR; + } + if (argc == 2) { + stringObjPtr = Jim_GetVariable(interp, argv[1], JIM_ERRMSG); + if (!stringObjPtr) return JIM_ERR; + } else { + stringObjPtr = Jim_GetVariable(interp, argv[1], JIM_NONE); + if (!stringObjPtr) { + /* Create the string if it does not exists */ + stringObjPtr = Jim_NewEmptyStringObj(interp); + if (Jim_SetVariable(interp, argv[1], stringObjPtr) + != JIM_OK) { + Jim_FreeNewObj(interp, stringObjPtr); + return JIM_ERR; + } + } + } + shared = Jim_IsShared(stringObjPtr); + if (shared) + stringObjPtr = Jim_DuplicateObj(interp, stringObjPtr); + for (i = 2; i < argc; i++) + Jim_AppendObj(interp, stringObjPtr, argv[i]); + if (Jim_SetVariable(interp, argv[1], stringObjPtr) != JIM_OK) { + if (shared) + Jim_FreeNewObj(interp, stringObjPtr); + return JIM_ERR; + } + Jim_SetResult(interp, stringObjPtr); + return JIM_OK; +} + +/* [debug] */ +static int Jim_DebugCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *options[] = { + "refcount", "objcount", "objects", "invstr", "scriptlen", "exprlen", + "exprbc", + NULL + }; + enum { + OPT_REFCOUNT, OPT_OBJCOUNT, OPT_OBJECTS, OPT_INVSTR, OPT_SCRIPTLEN, + OPT_EXPRLEN, OPT_EXPRBC + }; + int option; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "option ?...?"); + return JIM_ERR; + } + if (Jim_GetEnum(interp, argv[1], options, &option, "option", + JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + if (option == OPT_REFCOUNT) { + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "object"); + return JIM_ERR; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, argv[2]->refCount)); + return JIM_OK; + } else if (option == OPT_OBJCOUNT) { + int freeobj = 0, liveobj = 0; + char buf[256]; + Jim_Obj *objPtr; + + if (argc != 2) { + Jim_WrongNumArgs(interp, 2, argv, ""); + return JIM_ERR; + } + /* Count the number of free objects. */ + objPtr = interp->freeList; + while (objPtr) { + freeobj++; + objPtr = objPtr->nextObjPtr; + } + /* Count the number of live objects. */ + objPtr = interp->liveList; + while (objPtr) { + liveobj++; + objPtr = objPtr->nextObjPtr; + } + /* Set the result string and return. */ + sprintf(buf, "free %d used %d", freeobj, liveobj); + Jim_SetResultString(interp, buf, -1); + return JIM_OK; + } else if (option == OPT_OBJECTS) { + Jim_Obj *objPtr, *listObjPtr, *subListObjPtr; + /* Count the number of live objects. */ + objPtr = interp->liveList; + listObjPtr = Jim_NewListObj(interp, NULL, 0); + while (objPtr) { + char buf[128]; + const char *type = objPtr->typePtr ? + objPtr->typePtr->name : ""; + subListObjPtr = Jim_NewListObj(interp, NULL, 0); + sprintf(buf, "%p", objPtr); + Jim_ListAppendElement(interp, subListObjPtr, + Jim_NewStringObj(interp, buf, -1)); + Jim_ListAppendElement(interp, subListObjPtr, + Jim_NewStringObj(interp, type, -1)); + Jim_ListAppendElement(interp, subListObjPtr, + Jim_NewIntObj(interp, objPtr->refCount)); + Jim_ListAppendElement(interp, subListObjPtr, objPtr); + Jim_ListAppendElement(interp, listObjPtr, subListObjPtr); + objPtr = objPtr->nextObjPtr; + } + Jim_SetResult(interp, listObjPtr); + return JIM_OK; + } else if (option == OPT_INVSTR) { + Jim_Obj *objPtr; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "object"); + return JIM_ERR; + } + objPtr = argv[2]; + if (objPtr->typePtr != NULL) + Jim_InvalidateStringRep(objPtr); + Jim_SetEmptyResult(interp); + return JIM_OK; + } else if (option == OPT_SCRIPTLEN) { + ScriptObj *script; + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "script"); + return JIM_ERR; + } + script = Jim_GetScript(interp, argv[2]); + Jim_SetResult(interp, Jim_NewIntObj(interp, script->len)); + return JIM_OK; + } else if (option == OPT_EXPRLEN) { + ExprByteCode *expr; + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "expression"); + return JIM_ERR; + } + expr = Jim_GetExpression(interp, argv[2]); + if (expr == NULL) + return JIM_ERR; + Jim_SetResult(interp, Jim_NewIntObj(interp, expr->len)); + return JIM_OK; + } else if (option == OPT_EXPRBC) { + Jim_Obj *objPtr; + ExprByteCode *expr; + int i; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "expression"); + return JIM_ERR; + } + expr = Jim_GetExpression(interp, argv[2]); + if (expr == NULL) + return JIM_ERR; + objPtr = Jim_NewListObj(interp, NULL, 0); + for (i = 0; i < expr->len; i++) { + const char *type; + Jim_ExprOperator *op; + + switch (expr->opcode[i]) { + case JIM_EXPROP_NUMBER: type = "number"; break; + case JIM_EXPROP_COMMAND: type = "command"; break; + case JIM_EXPROP_VARIABLE: type = "variable"; break; + case JIM_EXPROP_DICTSUGAR: type = "dictsugar"; break; + case JIM_EXPROP_SUBST: type = "subst"; break; + case JIM_EXPROP_STRING: type = "string"; break; + default: + op = JimExprOperatorInfo(Jim_GetString(expr->obj[i], NULL)); + if (op == NULL) { + type = "private"; + } else { + type = "operator"; + } + break; + } + Jim_ListAppendElement(interp, objPtr, + Jim_NewStringObj(interp, type, -1)); + Jim_ListAppendElement(interp, objPtr, expr->obj[i]); + } + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else { + Jim_SetResultString(interp, + "bad option. Valid options are refcount, " + "objcount, objects, invstr", -1); + return JIM_ERR; + } + return JIM_OK; /* unreached */ +} + +/* [eval] */ +static int Jim_EvalCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc == 2) { + return Jim_EvalObj(interp, argv[1]); + } else if (argc > 2) { + Jim_Obj *objPtr; + int retcode; + + objPtr = Jim_ConcatObj(interp, argc-1, argv + 1); + Jim_IncrRefCount(objPtr); + retcode = Jim_EvalObj(interp, objPtr); + Jim_DecrRefCount(interp, objPtr); + return retcode; + } else { + Jim_WrongNumArgs(interp, 1, argv, "script ?...?"); + return JIM_ERR; + } +} + +/* [uplevel] */ +static int Jim_UplevelCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc >= 2) { + int retcode, newLevel, oldLevel; + Jim_CallFrame *savedCallFrame, *targetCallFrame; + Jim_Obj *objPtr; + const char *str; + + /* Save the old callframe pointer */ + savedCallFrame = interp->framePtr; + + /* Lookup the target frame pointer */ + str = Jim_GetString(argv[1], NULL); + if ((str[0] >= '0' && str[0] <= '9') || str[0] == '#') + { + if (Jim_GetCallFrameByLevel(interp, argv[1], + &targetCallFrame, + &newLevel) != JIM_OK) + return JIM_ERR; + argc--; + argv++; + } else { + if (Jim_GetCallFrameByLevel(interp, NULL, + &targetCallFrame, + &newLevel) != JIM_OK) + return JIM_ERR; + } + if (argc < 2) { + argc++; + argv--; + Jim_WrongNumArgs(interp, 1, argv, + "?level? command ?arg ...?"); + return JIM_ERR; + } + /* Eval the code in the target callframe. */ + interp->framePtr = targetCallFrame; + oldLevel = interp->numLevels; + interp->numLevels = newLevel; + if (argc == 2) { + retcode = Jim_EvalObj(interp, argv[1]); + } else { + objPtr = Jim_ConcatObj(interp, argc-1, argv + 1); + Jim_IncrRefCount(objPtr); + retcode = Jim_EvalObj(interp, objPtr); + Jim_DecrRefCount(interp, objPtr); + } + interp->numLevels = oldLevel; + interp->framePtr = savedCallFrame; + return retcode; + } else { + Jim_WrongNumArgs(interp, 1, argv, "?level? command ?arg ...?"); + return JIM_ERR; + } +} + +/* [expr] */ +static int Jim_ExprCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *exprResultPtr; + int retcode; + + if (argc == 2) { + retcode = Jim_EvalExpression(interp, argv[1], &exprResultPtr); + } else if (argc > 2) { + Jim_Obj *objPtr; + + objPtr = Jim_ConcatObj(interp, argc-1, argv + 1); + Jim_IncrRefCount(objPtr); + retcode = Jim_EvalExpression(interp, objPtr, &exprResultPtr); + Jim_DecrRefCount(interp, objPtr); + } else { + Jim_WrongNumArgs(interp, 1, argv, "expression ?...?"); + return JIM_ERR; + } + if (retcode != JIM_OK) return retcode; + Jim_SetResult(interp, exprResultPtr); + Jim_DecrRefCount(interp, exprResultPtr); + return JIM_OK; +} + +/* [break] */ +static int Jim_BreakCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 1) { + Jim_WrongNumArgs(interp, 1, argv, ""); + return JIM_ERR; + } + return JIM_BREAK; +} + +/* [continue] */ +static int Jim_ContinueCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 1) { + Jim_WrongNumArgs(interp, 1, argv, ""); + return JIM_ERR; + } + return JIM_CONTINUE; +} + +/* [return] */ +static int Jim_ReturnCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc == 1) { + return JIM_RETURN; + } else if (argc == 2) { + Jim_SetResult(interp, argv[1]); + interp->returnCode = JIM_OK; + return JIM_RETURN; + } else if (argc == 3 || argc == 4) { + int returnCode; + if (Jim_GetReturnCode(interp, argv[2], &returnCode) == JIM_ERR) + return JIM_ERR; + interp->returnCode = returnCode; + if (argc == 4) + Jim_SetResult(interp, argv[3]); + return JIM_RETURN; + } else { + Jim_WrongNumArgs(interp, 1, argv, "?-code code? ?result?"); + return JIM_ERR; + } + return JIM_RETURN; /* unreached */ +} + +/* [tailcall] */ +static int Jim_TailcallCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *objPtr; + + objPtr = Jim_NewListObj(interp, argv + 1, argc-1); + Jim_SetResult(interp, objPtr); + return JIM_EVAL; +} + +/* [proc] */ +static int Jim_ProcCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int argListLen; + int arityMin, arityMax; + + if (argc != 4 && argc != 5) { + Jim_WrongNumArgs(interp, 1, argv, "name arglist ?statics? body"); + return JIM_ERR; + } + Jim_ListLength(interp, argv[2], &argListLen); + arityMin = arityMax = argListLen + 1; + + if (argListLen) { + const char *str; + int len; + Jim_Obj *argPtr=NULL; + + /* Check for 'args' and adjust arityMin and arityMax if necessary */ + Jim_ListIndex(interp, argv[2], argListLen-1, &argPtr, JIM_NONE); + str = Jim_GetString(argPtr, &len); + if (len == 4 && memcmp(str, "args", 4) == 0) { + arityMin--; + arityMax = -1; + } + + /* Check for default arguments and reduce arityMin if necessary */ + while (arityMin > 1) { + Jim_ListIndex(interp, argv[2], arityMin - 2, &argPtr, JIM_NONE); + Jim_ListLength(interp, argPtr, &len); + if (len != 2) { + /* No default argument */ + break; + } + arityMin--; + } + } + if (argc == 4) { + return Jim_CreateProcedure(interp, Jim_GetString(argv[1], NULL), + argv[2], NULL, argv[3], arityMin, arityMax); + } else { + return Jim_CreateProcedure(interp, Jim_GetString(argv[1], NULL), + argv[2], argv[3], argv[4], arityMin, arityMax); + } +} + +/* [concat] */ +static int Jim_ConcatCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_SetResult(interp, Jim_ConcatObj(interp, argc-1, argv + 1)); + return JIM_OK; +} + +/* [upvar] */ +static int Jim_UpvarCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *str; + int i; + Jim_CallFrame *targetCallFrame; + + /* Lookup the target frame pointer */ + str = Jim_GetString(argv[1], NULL); + if (argc > 3 && + ((str[0] >= '0' && str[0] <= '9') || str[0] == '#')) + { + if (Jim_GetCallFrameByLevel(interp, argv[1], + &targetCallFrame, NULL) != JIM_OK) + return JIM_ERR; + argc--; + argv++; + } else { + if (Jim_GetCallFrameByLevel(interp, NULL, + &targetCallFrame, NULL) != JIM_OK) + return JIM_ERR; + } + /* Check for arity */ + if (argc < 3 || ((argc-1)%2) != 0) { + Jim_WrongNumArgs(interp, 1, argv, "?level? otherVar localVar ?otherVar localVar ...?"); + return JIM_ERR; + } + /* Now... for every other/local couple: */ + for (i = 1; i < argc; i += 2) { + if (Jim_SetVariableLink(interp, argv[i + 1], argv[i], + targetCallFrame) != JIM_OK) return JIM_ERR; + } + return JIM_OK; +} + +/* [global] */ +static int Jim_GlobalCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int i; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "varName ?varName ...?"); + return JIM_ERR; + } + /* Link every var to the toplevel having the same name */ + if (interp->numLevels == 0) return JIM_OK; /* global at toplevel... */ + for (i = 1; i < argc; i++) { + if (Jim_SetVariableLink(interp, argv[i], argv[i], + interp->topFramePtr) != JIM_OK) return JIM_ERR; + } + return JIM_OK; +} + +/* does the [string map] operation. On error NULL is returned, + * otherwise a new string object with the result, having refcount = 0, + * is returned. */ +static Jim_Obj *JimStringMap(Jim_Interp *interp, Jim_Obj *mapListObjPtr, + Jim_Obj *objPtr, int nocase) +{ + int numMaps; + const char **key, *str, *noMatchStart = NULL; + Jim_Obj **value; + int *keyLen, strLen, i; + Jim_Obj *resultObjPtr; + + Jim_ListLength(interp, mapListObjPtr, &numMaps); + if (numMaps % 2) { + Jim_SetResultString(interp, + "list must contain an even number of elements", -1); + return NULL; + } + /* Initialization */ + numMaps /= 2; + key = Jim_Alloc(sizeof(char*)*numMaps); + keyLen = Jim_Alloc(sizeof(int)*numMaps); + value = Jim_Alloc(sizeof(Jim_Obj*)*numMaps); + resultObjPtr = Jim_NewStringObj(interp, "", 0); + for (i = 0; i < numMaps; i++) { + Jim_Obj *eleObjPtr=NULL; + + Jim_ListIndex(interp, mapListObjPtr, i*2, &eleObjPtr, JIM_NONE); + key[i] = Jim_GetString(eleObjPtr, &keyLen[i]); + Jim_ListIndex(interp, mapListObjPtr, i*2 + 1, &eleObjPtr, JIM_NONE); + value[i] = eleObjPtr; + } + str = Jim_GetString(objPtr, &strLen); + /* Map it */ + while (strLen) { + for (i = 0; i < numMaps; i++) { + if (strLen >= keyLen[i] && keyLen[i]) { + if (!JimStringCompare(str, keyLen[i], key[i], keyLen[i], + nocase)) + { + if (noMatchStart) { + Jim_AppendString(interp, resultObjPtr, + noMatchStart, str-noMatchStart); + noMatchStart = NULL; + } + Jim_AppendObj(interp, resultObjPtr, value[i]); + str += keyLen[i]; + strLen -= keyLen[i]; + break; + } + } + } + if (i == numMaps) { /* no match */ + if (noMatchStart == NULL) + noMatchStart = str; + str ++; + strLen --; + } + } + if (noMatchStart) { + Jim_AppendString(interp, resultObjPtr, + noMatchStart, str-noMatchStart); + } + Jim_Free((void*)key); + Jim_Free(keyLen); + Jim_Free(value); + return resultObjPtr; +} + +/* [string] */ +static int Jim_StringCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int option; + const char *options[] = { + "length", "compare", "match", "equal", "range", "map", "repeat", + "index", "first", "tolower", "toupper", NULL + }; + enum { + OPT_LENGTH, OPT_COMPARE, OPT_MATCH, OPT_EQUAL, OPT_RANGE, + OPT_MAP, OPT_REPEAT, OPT_INDEX, OPT_FIRST, OPT_TOLOWER, OPT_TOUPPER + }; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "option ?arguments ...?"); + return JIM_ERR; + } + if (Jim_GetEnum(interp, argv[1], options, &option, "option", + JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + + if (option == OPT_LENGTH) { + int len; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "string"); + return JIM_ERR; + } + Jim_GetString(argv[2], &len); + Jim_SetResult(interp, Jim_NewIntObj(interp, len)); + return JIM_OK; + } else if (option == OPT_COMPARE) { + int nocase = 0; + if ((argc != 4 && argc != 5) || + (argc == 5 && Jim_CompareStringImmediate(interp, + argv[2], "-nocase") == 0)) { + Jim_WrongNumArgs(interp, 2, argv, "string1 string2"); + return JIM_ERR; + } + if (argc == 5) { + nocase = 1; + argv++; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, + Jim_StringCompareObj(argv[2], + argv[3], nocase))); + return JIM_OK; + } else if (option == OPT_MATCH) { + int nocase = 0; + if ((argc != 4 && argc != 5) || + (argc == 5 && Jim_CompareStringImmediate(interp, + argv[2], "-nocase") == 0)) { + Jim_WrongNumArgs(interp, 2, argv, "?-nocase? pattern " + "string"); + return JIM_ERR; + } + if (argc == 5) { + nocase = 1; + argv++; + } + Jim_SetResult(interp, + Jim_NewIntObj(interp, Jim_StringMatchObj(argv[2], + argv[3], nocase))); + return JIM_OK; + } else if (option == OPT_EQUAL) { + if (argc != 4) { + Jim_WrongNumArgs(interp, 2, argv, "string1 string2"); + return JIM_ERR; + } + Jim_SetResult(interp, + Jim_NewIntObj(interp, Jim_StringEqObj(argv[2], + argv[3], 0))); + return JIM_OK; + } else if (option == OPT_RANGE) { + Jim_Obj *objPtr; + + if (argc != 5) { + Jim_WrongNumArgs(interp, 2, argv, "string first last"); + return JIM_ERR; + } + objPtr = Jim_StringRangeObj(interp, argv[2], argv[3], argv[4]); + if (objPtr == NULL) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else if (option == OPT_MAP) { + int nocase = 0; + Jim_Obj *objPtr; + + if ((argc != 4 && argc != 5) || + (argc == 5 && Jim_CompareStringImmediate(interp, + argv[2], "-nocase") == 0)) { + Jim_WrongNumArgs(interp, 2, argv, "?-nocase? mapList " + "string"); + return JIM_ERR; + } + if (argc == 5) { + nocase = 1; + argv++; + } + objPtr = JimStringMap(interp, argv[2], argv[3], nocase); + if (objPtr == NULL) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else if (option == OPT_REPEAT) { + Jim_Obj *objPtr; + jim_wide count; + + if (argc != 4) { + Jim_WrongNumArgs(interp, 2, argv, "string count"); + return JIM_ERR; + } + if (Jim_GetWide(interp, argv[3], &count) != JIM_OK) + return JIM_ERR; + objPtr = Jim_NewStringObj(interp, "", 0); + while (count--) { + Jim_AppendObj(interp, objPtr, argv[2]); + } + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else if (option == OPT_INDEX) { + int index_t, len; + const char *str; + + if (argc != 4) { + Jim_WrongNumArgs(interp, 2, argv, "string index"); + return JIM_ERR; + } + if (Jim_GetIndex(interp, argv[3], &index_t) != JIM_OK) + return JIM_ERR; + str = Jim_GetString(argv[2], &len); + if (index_t != INT_MIN && index_t != INT_MAX) + index_t = JimRelToAbsIndex(len, index_t); + if (index_t < 0 || index_t >= len) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + return JIM_OK; + } else { + Jim_SetResult(interp, Jim_NewStringObj(interp, str + index_t, 1)); + return JIM_OK; + } + } else if (option == OPT_FIRST) { + int index_t = 0, l1, l2; + const char *s1, *s2; + + if (argc != 4 && argc != 5) { + Jim_WrongNumArgs(interp, 2, argv, "subString string ?startIndex?"); + return JIM_ERR; + } + s1 = Jim_GetString(argv[2], &l1); + s2 = Jim_GetString(argv[3], &l2); + if (argc == 5) { + if (Jim_GetIndex(interp, argv[4], &index_t) != JIM_OK) + return JIM_ERR; + index_t = JimRelToAbsIndex(l2, index_t); + } + Jim_SetResult(interp, Jim_NewIntObj(interp, + JimStringFirst(s1, l1, s2, l2, index_t))); + return JIM_OK; + } else if (option == OPT_TOLOWER) { + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "string"); + return JIM_ERR; + } + Jim_SetResult(interp, JimStringToLower(interp, argv[2])); + } else if (option == OPT_TOUPPER) { + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "string"); + return JIM_ERR; + } + Jim_SetResult(interp, JimStringToUpper(interp, argv[2])); + } + return JIM_OK; +} + +/* [time] */ +static int Jim_TimeCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + long i, count = 1; + jim_wide start, elapsed; + char buf [256]; + const char *fmt = "%" JIM_WIDE_MODIFIER " microseconds per iteration"; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "script ?count?"); + return JIM_ERR; + } + if (argc == 3) { + if (Jim_GetLong(interp, argv[2], &count) != JIM_OK) + return JIM_ERR; + } + if (count < 0) + return JIM_OK; + i = count; + start = JimClock(); + while (i-- > 0) { + int retval; + + if ((retval = Jim_EvalObj(interp, argv[1])) != JIM_OK) + return retval; + } + elapsed = JimClock() - start; + sprintf(buf, fmt, elapsed/count); + Jim_SetResultString(interp, buf, -1); + return JIM_OK; +} + +/* [exit] */ +static int Jim_ExitCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + long exitCode = 0; + + if (argc > 2) { + Jim_WrongNumArgs(interp, 1, argv, "?exitCode?"); + return JIM_ERR; + } + if (argc == 2) { + if (Jim_GetLong(interp, argv[1], &exitCode) != JIM_OK) + return JIM_ERR; + } + interp->exitCode = exitCode; + return JIM_EXIT; +} + +/* [catch] */ +static int Jim_CatchCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int exitCode = 0; + + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "script ?varName?"); + return JIM_ERR; + } + exitCode = Jim_EvalObj(interp, argv[1]); + if (argc == 3) { + if (Jim_SetVariable(interp, argv[2], Jim_GetResult(interp)) + != JIM_OK) + return JIM_ERR; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, exitCode)); + return JIM_OK; +} + +/* [ref] */ +static int Jim_RefCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 3 && argc != 4) { + Jim_WrongNumArgs(interp, 1, argv, "string tag ?finalizer?"); + return JIM_ERR; + } + if (argc == 3) { + Jim_SetResult(interp, Jim_NewReference(interp, argv[1], argv[2], NULL)); + } else { + Jim_SetResult(interp, Jim_NewReference(interp, argv[1], argv[2], + argv[3])); + } + return JIM_OK; +} + +/* [getref] */ +static int Jim_GetrefCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Reference *refPtr; + + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "reference"); + return JIM_ERR; + } + if ((refPtr = Jim_GetReference(interp, argv[1])) == NULL) + return JIM_ERR; + Jim_SetResult(interp, refPtr->objPtr); + return JIM_OK; +} + +/* [setref] */ +static int Jim_SetrefCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Reference *refPtr; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "reference newValue"); + return JIM_ERR; + } + if ((refPtr = Jim_GetReference(interp, argv[1])) == NULL) + return JIM_ERR; + Jim_IncrRefCount(argv[2]); + Jim_DecrRefCount(interp, refPtr->objPtr); + refPtr->objPtr = argv[2]; + Jim_SetResult(interp, argv[2]); + return JIM_OK; +} + +/* [collect] */ +static int Jim_CollectCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 1) { + Jim_WrongNumArgs(interp, 1, argv, ""); + return JIM_ERR; + } + Jim_SetResult(interp, Jim_NewIntObj(interp, Jim_Collect(interp))); + return JIM_OK; +} + +/* [finalize] reference ?newValue? */ +static int Jim_FinalizeCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "reference ?finalizerProc?"); + return JIM_ERR; + } + if (argc == 2) { + Jim_Obj *cmdNamePtr; + + if (Jim_GetFinalizer(interp, argv[1], &cmdNamePtr) != JIM_OK) + return JIM_ERR; + if (cmdNamePtr != NULL) /* otherwise the null string is returned. */ + Jim_SetResult(interp, cmdNamePtr); + } else { + if (Jim_SetFinalizer(interp, argv[1], argv[2]) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, argv[2]); + } + return JIM_OK; +} + +/* TODO */ +/* [info references] (list of all the references/finalizers) */ + +/* [rename] */ +static int Jim_RenameCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *oldName, *newName; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "oldName newName"); + return JIM_ERR; + } + oldName = Jim_GetString(argv[1], NULL); + newName = Jim_GetString(argv[2], NULL); + if (Jim_RenameCommand(interp, oldName, newName) != JIM_OK) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "can't rename \"", oldName, "\": ", + "command doesn't exist", NULL); + return JIM_ERR; + } + return JIM_OK; +} + +/* [dict] */ +static int Jim_DictCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int option; + const char *options[] = { + "create", "get", "set", "unset", "exists", NULL + }; + enum { + OPT_CREATE, OPT_GET, OPT_SET, OPT_UNSET, OPT_EXIST + }; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "option ?arguments ...?"); + return JIM_ERR; + } + + if (Jim_GetEnum(interp, argv[1], options, &option, "option", + JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + + if (option == OPT_CREATE) { + Jim_Obj *objPtr; + + if (argc % 2) { + Jim_WrongNumArgs(interp, 2, argv, "?key value ...?"); + return JIM_ERR; + } + objPtr = Jim_NewDictObj(interp, argv + 2, argc-2); + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else if (option == OPT_GET) { + Jim_Obj *objPtr; + + if (Jim_DictKeysVector(interp, argv[2], argv + 3, argc-3, &objPtr, + JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; + } else if (option == OPT_SET) { + if (argc < 5) { + Jim_WrongNumArgs(interp, 2, argv, "varName key ?key ...? value"); + return JIM_ERR; + } + return Jim_SetDictKeysVector(interp, argv[2], argv + 3, argc-4, + argv[argc-1]); + } else if (option == OPT_UNSET) { + if (argc < 4) { + Jim_WrongNumArgs(interp, 2, argv, "varName key ?key ...?"); + return JIM_ERR; + } + return Jim_SetDictKeysVector(interp, argv[2], argv + 3, argc-3, + NULL); + } else if (option == OPT_EXIST) { + Jim_Obj *objPtr; + int exists; + + if (Jim_DictKeysVector(interp, argv[2], argv + 3, argc-3, &objPtr, + JIM_ERRMSG) == JIM_OK) + exists = 1; + else + exists = 0; + Jim_SetResult(interp, Jim_NewIntObj(interp, exists)); + return JIM_OK; + } else { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad option \"", Jim_GetString(argv[1], NULL), "\":", + " must be create, get, set", NULL); + return JIM_ERR; + } + return JIM_OK; +} + +/* [load] */ +static int Jim_LoadCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "libaryFile"); + return JIM_ERR; + } + return Jim_LoadLibrary(interp, Jim_GetString(argv[1], NULL)); +} + +/* [subst] */ +static int Jim_SubstCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int i, flags = 0; + Jim_Obj *objPtr; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, + "?-nobackslashes? ?-nocommands? ?-novariables? string"); + return JIM_ERR; + } + i = argc-2; + while (i--) { + if (Jim_CompareStringImmediate(interp, argv[i + 1], + "-nobackslashes")) + flags |= JIM_SUBST_NOESC; + else if (Jim_CompareStringImmediate(interp, argv[i + 1], + "-novariables")) + flags |= JIM_SUBST_NOVAR; + else if (Jim_CompareStringImmediate(interp, argv[i + 1], + "-nocommands")) + flags |= JIM_SUBST_NOCMD; + else { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "bad option \"", Jim_GetString(argv[i + 1], NULL), + "\": must be -nobackslashes, -nocommands, or " + "-novariables", NULL); + return JIM_ERR; + } + } + if (Jim_SubstObj(interp, argv[argc-1], &objPtr, flags) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; +} + +/* [info] */ +static int Jim_InfoCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int cmd, result = JIM_OK; + static const char *commands[] = { + "body", "commands", "exists", "globals", "level", "locals", + "vars", "version", "complete", "args", "hostname", NULL + }; + enum {INFO_BODY, INFO_COMMANDS, INFO_EXISTS, INFO_GLOBALS, INFO_LEVEL, + INFO_LOCALS, INFO_VARS, INFO_VERSION, INFO_COMPLETE, INFO_ARGS, INFO_HOSTNAME}; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "command ?args ...?"); + return JIM_ERR; + } + if (Jim_GetEnum(interp, argv[1], commands, &cmd, "command", JIM_ERRMSG) + != JIM_OK) { + return JIM_ERR; + } + + if (cmd == INFO_COMMANDS) { + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "?pattern?"); + return JIM_ERR; + } + if (argc == 3) + Jim_SetResult(interp,JimCommandsList(interp, argv[2])); + else + Jim_SetResult(interp, JimCommandsList(interp, NULL)); + } else if (cmd == INFO_EXISTS) { + Jim_Obj *exists; + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "varName"); + return JIM_ERR; + } + exists = Jim_GetVariable(interp, argv[2], 0); + Jim_SetResult(interp, Jim_NewIntObj(interp, exists != 0)); + } else if (cmd == INFO_GLOBALS || cmd == INFO_LOCALS || cmd == INFO_VARS) { + int mode; + switch (cmd) { + case INFO_GLOBALS: mode = JIM_VARLIST_GLOBALS; break; + case INFO_LOCALS: mode = JIM_VARLIST_LOCALS; break; + case INFO_VARS: mode = JIM_VARLIST_VARS; break; + default: mode = 0; /* avoid warning */; break; + } + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "?pattern?"); + return JIM_ERR; + } + if (argc == 3) + Jim_SetResult(interp,JimVariablesList(interp, argv[2], mode)); + else + Jim_SetResult(interp, JimVariablesList(interp, NULL, mode)); + } else if (cmd == INFO_LEVEL) { + Jim_Obj *objPtr; + switch (argc) { + case 2: + Jim_SetResult(interp, + Jim_NewIntObj(interp, interp->numLevels)); + break; + case 3: + if (JimInfoLevel(interp, argv[2], &objPtr) != JIM_OK) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + break; + default: + Jim_WrongNumArgs(interp, 2, argv, "?levelNum?"); + return JIM_ERR; + } + } else if (cmd == INFO_BODY || cmd == INFO_ARGS) { + Jim_Cmd *cmdPtr; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "procname"); + return JIM_ERR; + } + if ((cmdPtr = Jim_GetCommand(interp, argv[2], JIM_ERRMSG)) == NULL) + return JIM_ERR; + if (cmdPtr->cmdProc != NULL) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "command \"", Jim_GetString(argv[2], NULL), + "\" is not a procedure", NULL); + return JIM_ERR; + } + if (cmd == INFO_BODY) + Jim_SetResult(interp, cmdPtr->bodyObjPtr); + else + Jim_SetResult(interp, cmdPtr->argListObjPtr); + } else if (cmd == INFO_VERSION) { + char buf[(JIM_INTEGER_SPACE * 2) + 1]; + sprintf(buf, "%d.%d", + JIM_VERSION / 100, JIM_VERSION % 100); + Jim_SetResultString(interp, buf, -1); + } else if (cmd == INFO_COMPLETE) { + const char *s; + int len; + + if (argc != 3) { + Jim_WrongNumArgs(interp, 2, argv, "script"); + return JIM_ERR; + } + s = Jim_GetString(argv[2], &len); + Jim_SetResult(interp, + Jim_NewIntObj(interp, Jim_ScriptIsComplete(s, len, NULL))); + } else if (cmd == INFO_HOSTNAME) { + /* Redirect to os.hostname if it exists */ + Jim_Obj *command = Jim_NewStringObj(interp, "os.gethostname", -1); + result = Jim_EvalObjVector(interp, 1, &command); + } + return result; +} + +/* [split] */ +static int Jim_SplitCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *str, *splitChars, *noMatchStart; + int splitLen, strLen, i; + Jim_Obj *resObjPtr; + + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "string ?splitChars?"); + return JIM_ERR; + } + /* Init */ + if (argc == 2) { + splitChars = " \n\t\r"; + splitLen = 4; + } else { + splitChars = Jim_GetString(argv[2], &splitLen); + } + str = Jim_GetString(argv[1], &strLen); + if (!strLen) return JIM_OK; + noMatchStart = str; + resObjPtr = Jim_NewListObj(interp, NULL, 0); + /* Split */ + if (splitLen) { + while (strLen) { + for (i = 0; i < splitLen; i++) { + if (*str == splitChars[i]) { + Jim_Obj *objPtr; + + objPtr = Jim_NewStringObj(interp, noMatchStart, + (str-noMatchStart)); + Jim_ListAppendElement(interp, resObjPtr, objPtr); + noMatchStart = str + 1; + break; + } + } + str ++; + strLen --; + } + Jim_ListAppendElement(interp, resObjPtr, + Jim_NewStringObj(interp, noMatchStart, (str-noMatchStart))); + } else { + /* This handles the special case of splitchars eq {}. This + * is trivial but we want to perform object sharing as Tcl does. */ + Jim_Obj *objCache[256]; + const unsigned char *u = (unsigned char*) str; + memset(objCache, 0, sizeof(objCache)); + for (i = 0; i < strLen; i++) { + int c = u[i]; + + if (objCache[c] == NULL) + objCache[c] = Jim_NewStringObj(interp, (char*)u + i, 1); + Jim_ListAppendElement(interp, resObjPtr, objCache[c]); + } + } + Jim_SetResult(interp, resObjPtr); + return JIM_OK; +} + +/* [join] */ +static int Jim_JoinCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *joinStr; + int joinStrLen, i, listLen; + Jim_Obj *resObjPtr; + + if (argc != 2 && argc != 3) { + Jim_WrongNumArgs(interp, 1, argv, "list ?joinString?"); + return JIM_ERR; + } + /* Init */ + if (argc == 2) { + joinStr = " "; + joinStrLen = 1; + } else { + joinStr = Jim_GetString(argv[2], &joinStrLen); + } + Jim_ListLength(interp, argv[1], &listLen); + resObjPtr = Jim_NewStringObj(interp, NULL, 0); + /* Split */ + for (i = 0; i < listLen; i++) { + Jim_Obj *objPtr=NULL; + + Jim_ListIndex(interp, argv[1], i, &objPtr, JIM_NONE); + Jim_AppendObj(interp, resObjPtr, objPtr); + if (i + 1 != listLen) { + Jim_AppendString(interp, resObjPtr, joinStr, joinStrLen); + } + } + Jim_SetResult(interp, resObjPtr); + return JIM_OK; +} + +/* [format] */ +static int Jim_FormatCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *objPtr; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "formatString ?arg arg ...?"); + return JIM_ERR; + } + objPtr = Jim_FormatString(interp, argv[1], argc-2, argv + 2); + if (objPtr == NULL) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; +} + +/* [scan] */ +static int Jim_ScanCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *listPtr, **outVec; + int outc, i, count = 0; + + if (argc < 3) { + Jim_WrongNumArgs(interp, 1, argv, "string formatString ?varName ...?"); + return JIM_ERR; + } + if (argv[2]->typePtr != &scanFmtStringObjType) + SetScanFmtFromAny(interp, argv[2]); + if (FormatGetError(argv[2]) != 0) { + Jim_SetResultString(interp, FormatGetError(argv[2]), -1); + return JIM_ERR; + } + if (argc > 3) { + int maxPos = FormatGetMaxPos(argv[2]); + int arg_count = FormatGetCnvCount(argv[2]); + if (maxPos > argc-3) { + Jim_SetResultString(interp, "\"%n$\" argument index out of range", -1); + return JIM_ERR; + } else if (arg_count != 0 && arg_count < argc-3) { + Jim_SetResultString(interp, "variable is not assigned by any " + "conversion specifiers", -1); + return JIM_ERR; + } else if (arg_count > argc-3) { + Jim_SetResultString(interp, "different numbers of variable names and " + "field specifiers", -1); + return JIM_ERR; + } + } + listPtr = Jim_ScanString(interp, argv[1], argv[2], JIM_ERRMSG); + if (listPtr == 0) + return JIM_ERR; + if (argc > 3) { + int len = 0; + if (listPtr != 0 && listPtr != (Jim_Obj*)EOF) + Jim_ListLength(interp, listPtr, &len); + if (listPtr == (Jim_Obj*)EOF || len == 0) { // XXX + Jim_SetResult(interp, Jim_NewIntObj(interp, -1)); + return JIM_OK; + } + JimListGetElements(interp, listPtr, &outc, &outVec); + for (i = 0; i < outc; ++i) { + if (Jim_Len + +/* [error] */ +static int Jim_ErrorCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "message"); + return JIM_ERR; + } + Jim_SetResult(interp, argv[1]); + return JIM_ERR; +} + +/* [lrange] */ +static int Jim_LrangeCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *objPtr; + + if (argc != 4) { + Jim_WrongNumArgs(interp, 1, argv, "list first last"); + return JIM_ERR; + } + if ((objPtr = Jim_ListRange(interp, argv[1], argv[2], argv[3])) == NULL) + return JIM_ERR; + Jim_SetResult(interp, objPtr); + return JIM_OK; +} + +/* [env] */ +static int Jim_EnvCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + const char *key; + char *val; + + if (argc == 1) { + +#ifdef NEED_ENVIRON_EXTERN + extern char **environ; +#endif + + int i; + Jim_Obj *listObjPtr = Jim_NewListObj(interp, NULL, 0); + + for (i = 0; environ[i]; i++) { + const char *equals = strchr(environ[i], '='); + if (equals) { + Jim_ListAppendElement(interp, listObjPtr, Jim_NewStringObj(interp, environ[i], equals - environ[i])); + Jim_ListAppendElement(interp, listObjPtr, Jim_NewStringObj(interp, equals + 1, -1)); + } + } + + Jim_SetResult(interp, listObjPtr); + return JIM_OK; + } + + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "varName"); + return JIM_ERR; + } + key = Jim_GetString(argv[1], NULL); + val = getenv(key); + if (val == NULL) { + Jim_SetResult(interp, Jim_NewEmptyStringObj(interp)); + Jim_AppendStrings(interp, Jim_GetResult(interp), + "environment variable \"", + key, "\" does not exist", NULL); + return JIM_ERR; + } + Jim_SetResult(interp, Jim_NewStringObj(interp, val, -1)); + return JIM_OK; +} + +/* [source] */ +static int Jim_SourceCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int retval; + + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "fileName"); + return JIM_ERR; + } + retval = Jim_EvalFile(interp, Jim_GetString(argv[1], NULL)); + if (retval == JIM_ERR) { + return JIM_ERR_ADDSTACK; + } + if (retval == JIM_RETURN) + return JIM_OK; + return retval; +} + +/* [lreverse] */ +static int Jim_LreverseCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + Jim_Obj *revObjPtr, **ele; + int len; + + if (argc != 2) { + Jim_WrongNumArgs(interp, 1, argv, "list"); + return JIM_ERR; + } + JimListGetElements(interp, argv[1], &len, &ele); + len--; + revObjPtr = Jim_NewListObj(interp, NULL, 0); + while (len >= 0) + ListAppendElement(revObjPtr, ele[len--]); + Jim_SetResult(interp, revObjPtr); + return JIM_OK; +} + +static int JimRangeLen(jim_wide start, jim_wide end, jim_wide step) +{ + jim_wide len; + + if (step == 0) return -1; + if (start == end) return 0; + else if (step > 0 && start > end) return -1; + else if (step < 0 && end > start) return -1; + len = end-start; + if (len < 0) len = -len; /* abs(len) */ + if (step < 0) step = -step; /* abs(step) */ + len = 1 + ((len-1)/step); + /* We can truncate safely to INT_MAX, the range command + * will always return an error for a such long range + * because Tcl lists can't be so long. */ + if (len > INT_MAX) len = INT_MAX; + return (int)((len < 0) ? -1 : len); +} + +/* [range] */ +static int Jim_RangeCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + jim_wide start = 0, end, step = 1; + int len, i; + Jim_Obj *objPtr; + + if (argc < 2 || argc > 4) { + Jim_WrongNumArgs(interp, 1, argv, "?start? end ?step?"); + return JIM_ERR; + } + if (argc == 2) { + if (Jim_GetWide(interp, argv[1], &end) != JIM_OK) + return JIM_ERR; + } else { + if (Jim_GetWide(interp, argv[1], &start) != JIM_OK || + Jim_GetWide(interp, argv[2], &end) != JIM_OK) + return JIM_ERR; + if (argc == 4 && Jim_GetWide(interp, argv[3], &step) != JIM_OK) + return JIM_ERR; + } + if ((len = JimRangeLen(start, end, step)) == -1) { + Jim_SetResultString(interp, "Invalid (infinite?) range specified", -1); + return JIM_ERR; + } + objPtr = Jim_NewListObj(interp, NULL, 0); + for (i = 0; i < len; i++) + ListAppendElement(objPtr, Jim_NewIntObj(interp, start + i*step)); + Jim_SetResult(interp, objPtr); + return JIM_OK; +} + +/* [rand] */ +static int Jim_RandCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + jim_wide min = 0, max =0, len, maxMul; + + if (argc < 1 || argc > 3) { + Jim_WrongNumArgs(interp, 1, argv, "?min? max"); + return JIM_ERR; + } + if (argc == 1) { + max = JIM_WIDE_MAX; + } else if (argc == 2) { + if (Jim_GetWide(interp, argv[1], &max) != JIM_OK) + return JIM_ERR; + } else if (argc == 3) { + if (Jim_GetWide(interp, argv[1], &min) != JIM_OK || + Jim_GetWide(interp, argv[2], &max) != JIM_OK) + return JIM_ERR; + } + len = max-min; + if (len < 0) { + Jim_SetResultString(interp, "Invalid arguments (max < min)", -1); + return JIM_ERR; + } + maxMul = JIM_WIDE_MAX - (len ? (JIM_WIDE_MAX%len) : 0); + while (1) { + jim_wide r; + + JimRandomBytes(interp, &r, sizeof(jim_wide)); + if (r < 0 || r >= maxMul) continue; + r = (len == 0) ? 0 : r%len; + Jim_SetResult(interp, Jim_NewIntObj(interp, min + r)); + return JIM_OK; + } +} + +/* [package] */ +static int Jim_PackageCoreCommand(Jim_Interp *interp, int argc, + Jim_Obj *const *argv) +{ + int option; + const char *options[] = { + "require", "provide", NULL + }; + enum {OPT_REQUIRE, OPT_PROVIDE}; + + if (argc < 2) { + Jim_WrongNumArgs(interp, 1, argv, "option ?arguments ...?"); + return JIM_ERR; + } + if (Jim_GetEnum(interp, argv[1], options, &option, "option", + JIM_ERRMSG) != JIM_OK) + return JIM_ERR; + + if (option == OPT_REQUIRE) { + int exact = 0; + const char *ver; + + if (Jim_CompareStringImmediate(interp, argv[2], "-exact")) { + exact = 1; + argv++; + argc--; + } + if (argc != 3 && argc != 4) { + Jim_WrongNumArgs(interp, 2, argv, "?-exact? package ?version?"); + return JIM_ERR; + } + ver = Jim_PackageRequire(interp, Jim_GetString(argv[2], NULL), + argc == 4 ? Jim_GetString(argv[3], NULL) : "", + JIM_ERRMSG); + if (ver == NULL) + return JIM_ERR_ADDSTACK; + Jim_SetResultString(interp, ver, -1); + } else if (option == OPT_PROVIDE) { + if (argc != 4) { + Jim_WrongNumArgs(interp, 2, argv, "package version"); + return JIM_ERR; + } + return Jim_PackageProvide(interp, Jim_GetString(argv[2], NULL), + Jim_GetString(argv[3], NULL), JIM_ERRMSG); + } + return JIM_OK; +} + +static struct { + const char *name; + Jim_CmdProc cmdProc; +} Jim_CoreCommandsTable[] = { + {"set", Jim_SetCoreCommand}, + {"unset", Jim_UnsetCoreCommand}, + {"puts", Jim_PutsCoreCommand}, + {"+", Jim_AddCoreCommand}, + {"*", Jim_MulCoreCommand}, + {"-", Jim_SubCoreCommand}, + {"/", Jim_DivCoreCommand}, + {"incr", Jim_IncrCoreCommand}, + {"while", Jim_WhileCoreCommand}, + {"for", Jim_ForCoreCommand}, + {"foreach", Jim_ForeachCoreCommand}, + {"lmap", Jim_LmapCoreCommand}, + {"if", Jim_IfCoreCommand}, + {"switch", Jim_SwitchCoreCommand}, + {"list", Jim_ListCoreCommand}, + {"lindex", Jim_LindexCoreCommand}, + {"lset", Jim_LsetCoreCommand}, + {"llength", Jim_LlengthCoreCommand}, + {"lappend", Jim_LappendCoreCommand}, + {"linsert", Jim_LinsertCoreCommand}, + {"lsort", Jim_LsortCoreCommand}, + {"append", Jim_AppendCoreCommand}, + {"debug", Jim_DebugCoreCommand}, + {"eval", Jim_EvalCoreCommand}, + {"uplevel", Jim_UplevelCoreCommand}, + {"expr", Jim_ExprCoreCommand}, + {"break", Jim_BreakCoreCommand}, + {"continue", Jim_ContinueCoreCommand}, + {"proc", Jim_ProcCoreCommand}, + {"concat", Jim_ConcatCoreCommand}, + {"return", Jim_ReturnCoreCommand}, + {"upvar", Jim_UpvarCoreCommand}, + {"global", Jim_GlobalCoreCommand}, + {"string", Jim_StringCoreCommand}, + {"time", Jim_TimeCoreCommand}, + {"exit", Jim_ExitCoreCommand}, + {"catch", Jim_CatchCoreCommand}, + {"ref", Jim_RefCoreCommand}, + {"getref", Jim_GetrefCoreCommand}, + {"setref", Jim_SetrefCoreCommand}, + {"finalize", Jim_FinalizeCoreCommand}, + {"collect", Jim_CollectCoreCommand}, + {"rename", Jim_RenameCoreCommand}, + {"dict", Jim_DictCoreCommand}, + {"load", Jim_LoadCoreCommand}, + {"subst", Jim_SubstCoreCommand}, + {"info", Jim_InfoCoreCommand}, + {"split", Jim_SplitCoreCommand}, + {"join", Jim_JoinCoreCommand}, + {"format", Jim_FormatCoreCommand}, + {"scan", Jim_ScanCoreCommand}, + {"error", Jim_ErrorCoreCommand}, + {"lrange", Jim_LrangeCoreCommand}, + {"env", Jim_EnvCoreCommand}, + {"source", Jim_SourceCoreCommand}, + {"lreverse", Jim_LreverseCoreCommand}, + {"range", Jim_RangeCoreCommand}, + {"rand", Jim_RandCoreCommand}, + {"package", Jim_PackageCoreCommand}, + {"tailcall", Jim_TailcallCoreCommand}, + {NULL, NULL}, +}; + +/* Some Jim core command is actually a procedure written in Jim itself. */ +static void Jim_RegisterCoreProcedures(Jim_Interp *interp) +{ + Jim_Eval(interp, (char*) +"proc lambda {arglist args} {\n" +" set name [ref {} function lambdaFinalizer]\n" +" uplevel 1 [list proc $name $arglist {expand}$args]\n" +" return $name\n" +"}\n" +"proc lambdaFinalizer {name val} {\n" +" rename $name {}\n" +"}\n" +); +} + +void Jim_RegisterCoreCommands(Jim_Interp *interp) +{ + int i = 0; + + while (Jim_CoreCommandsTable[i].name != NULL) { + Jim_CreateCommand(interp, + Jim_CoreCommandsTable[i].name, + Jim_CoreCommandsTable[i].cmdProc, + NULL, NULL); + i++; + } + Jim_RegisterCoreProcedures(interp); +} + +/* ----------------------------------------------------------------------------- + * Interactive prompt + * ---------------------------------------------------------------------------*/ +void Jim_PrintErrorMessage(Jim_Interp *interp) +{ + int len, i; + + if (*interp->errorFileName) { + Jim_fprintf(interp, interp->cookie_stderr, "Runtime error, file \"%s\", line %d:" JIM_NL " ", + interp->errorFileName, interp->errorLine); + } + Jim_fprintf(interp,interp->cookie_stderr, "%s" JIM_NL, + Jim_GetString(interp->result, NULL)); + Jim_ListLength(interp, interp->stackTrace, &len); + for (i = len-3; i >= 0; i-= 3) { + Jim_Obj *objPtr=NULL; + const char *proc, *file, *line; + + Jim_ListIndex(interp, interp->stackTrace, i, &objPtr, JIM_NONE); + proc = Jim_GetString(objPtr, NULL); + Jim_ListIndex(interp, interp->stackTrace, i + 1, &objPtr, + JIM_NONE); + file = Jim_GetString(objPtr, NULL); + Jim_ListIndex(interp, interp->stackTrace, i + 2, &objPtr, + JIM_NONE); + line = Jim_GetString(objPtr, NULL); + if (*proc) { + Jim_fprintf(interp, interp->cookie_stderr, + "in procedure '%s' ", proc); + } + if (*file) { + Jim_fprintf(interp, interp->cookie_stderr, + "called at file \"%s\", line %s", + file, line); + } + if (*file || *proc) { + Jim_fprintf(interp, interp->cookie_stderr, JIM_NL); + } + } +} + +int Jim_InteractivePrompt(Jim_Interp *interp) +{ + int retcode = JIM_OK; + Jim_Obj *scriptObjPtr; + + Jim_fprintf(interp,interp->cookie_stdout, "Welcome to Jim version %d.%d, " + "Copyright (c) 2005-8 Salvatore Sanfilippo" JIM_NL, + JIM_VERSION / 100, JIM_VERSION % 100); + Jim_SetVariableStrWithStr(interp, "jim_interactive", "1"); + while (1) { + char buf[1024]; + const char *result; + const char *retcodestr[] = { + "ok", "error", "return", "break", "continue", "eval", "exit" + }; + int reslen; + + if (retcode != 0) { + if (retcode >= 2 && retcode <= 6) + Jim_fprintf(interp,interp->cookie_stdout, "[%s] . ", retcodestr[retcode]); + else + Jim_fprintf(interp,interp->cookie_stdout, "[%d] . ", retcode); + } else + Jim_fprintf(interp, interp->cookie_stdout, ". "); + Jim_fflush(interp, interp->cookie_stdout); + scriptObjPtr = Jim_NewStringObj(interp, "", 0); + Jim_IncrRefCount(scriptObjPtr); + while (1) { + const char *str; + char state; + int len; + + if (Jim_fgets(interp, buf, 1024, interp->cookie_stdin) == NULL) { + Jim_DecrRefCount(interp, scriptObjPtr); + goto out; + } + Jim_AppendString(interp, scriptObjPtr, buf, -1); + str = Jim_GetString(scriptObjPtr, &len); + if (Jim_ScriptIsComplete(str, len, &state)) + break; + Jim_fprintf(interp, interp->cookie_stdout, "%c> ", state); + Jim_fflush(interp, interp->cookie_stdout); + } + retcode = Jim_EvalObj(interp, scriptObjPtr); + Jim_DecrRefCount(interp, scriptObjPtr); + result = Jim_GetString(Jim_GetResult(interp), &reslen); + if (retcode == JIM_ERR) { + Jim_PrintErrorMessage(interp); + } else if (retcode == JIM_EXIT) { + exit(Jim_GetExitCode(interp)); + } else { + if (reslen) { + Jim_fwrite(interp, result, 1, reslen, interp->cookie_stdout); + Jim_fprintf(interp,interp->cookie_stdout, JIM_NL); + } + } + } +out: + return 0; +} + +/* ----------------------------------------------------------------------------- + * Jim's idea of STDIO.. + * ---------------------------------------------------------------------------*/ + +int Jim_fprintf(Jim_Interp *interp, void *cookie, const char *fmt, ...) +{ + int r; + + va_list ap; + va_start(ap,fmt); + r = Jim_vfprintf(interp, cookie, fmt,ap); + va_end(ap); + return r; +} + +int Jim_vfprintf(Jim_Interp *interp, void *cookie, const char *fmt, va_list ap) +{ + if ((interp == NULL) || (interp->cb_vfprintf == NULL)) { + errno = ENOTSUP; + return -1; + } + return (*(interp->cb_vfprintf))(cookie, fmt, ap); +} + +size_t Jim_fwrite(Jim_Interp *interp, const void *ptr, size_t size, size_t n, void *cookie) +{ + if ((interp == NULL) || (interp->cb_fwrite == NULL)) { + errno = ENOTSUP; + return 0; + } + return (*(interp->cb_fwrite))(ptr, size, n, cookie); +} + +size_t Jim_fread(Jim_Interp *interp, void *ptr, size_t size, size_t n, void *cookie) +{ + if ((interp == NULL) || (interp->cb_fread == NULL)) { + errno = ENOTSUP; + return 0; + } + return (*(interp->cb_fread))(ptr, size, n, cookie); +} + +int Jim_fflush(Jim_Interp *interp, void *cookie) +{ + if ((interp == NULL) || (interp->cb_fflush == NULL)) { + /* pretend all is well */ + return 0; + } + return (*(interp->cb_fflush))(cookie); +} + +char* Jim_fgets(Jim_Interp *interp, char *s, int size, void *cookie) +{ + if ((interp == NULL) || (interp->cb_fgets == NULL)) { + errno = ENOTSUP; + return NULL; + } + return (*(interp->cb_fgets))(s, size, cookie); +} +Jim_Nvp * +Jim_Nvp_name2value_simple(const Jim_Nvp *p, const char *name) +{ + while (p->name) { + if (0 == strcmp(name, p->name)) { + break; + } + p++; + } + return ((Jim_Nvp *)(p)); +} + +Jim_Nvp * +Jim_Nvp_name2value_nocase_simple(const Jim_Nvp *p, const char *name) +{ + while (p->name) { + if (0 == strcasecmp(name, p->name)) { + break; + } + p++; + } + return ((Jim_Nvp *)(p)); +} + +int +Jim_Nvp_name2value_obj(Jim_Interp *interp, + const Jim_Nvp *p, + Jim_Obj *o, + Jim_Nvp **result) +{ + return Jim_Nvp_name2value(interp, p, Jim_GetString(o, NULL), result); +} + + +int +Jim_Nvp_name2value(Jim_Interp *interp, + const Jim_Nvp *_p, + const char *name, + Jim_Nvp **result) +{ + const Jim_Nvp *p; + + p = Jim_Nvp_name2value_simple(_p, name); + + /* result */ + if (result) { + *result = (Jim_Nvp *)(p); + } + + /* found? */ + if (p->name) { + return JIM_OK; + } else { + return JIM_ERR; + } +} + +int +Jim_Nvp_name2value_obj_nocase(Jim_Interp *interp, const Jim_Nvp *p, Jim_Obj *o, Jim_Nvp **puthere) +{ + return Jim_Nvp_name2value_nocase(interp, p, Jim_GetString(o, NULL), puthere); +} + +int +Jim_Nvp_name2value_nocase(Jim_Interp *interp, const Jim_Nvp *_p, const char *name, Jim_Nvp **puthere) +{ + const Jim_Nvp *p; + + p = Jim_Nvp_name2value_nocase_simple(_p, name); + + if (puthere) { + *puthere = (Jim_Nvp *)(p); + } + /* found */ + if (p->name) { + return JIM_OK; + } else { + return JIM_ERR; + } +} + + +int +Jim_Nvp_value2name_obj(Jim_Interp *interp, const Jim_Nvp *p, Jim_Obj *o, Jim_Nvp **result) +{ + int e;; + jim_wide w; + + e = Jim_GetWide(interp, o, &w); + if (e != JIM_OK) { + return e; + } + + return Jim_Nvp_value2name(interp, p, w, result); +} + +Jim_Nvp * +Jim_Nvp_value2name_simple(const Jim_Nvp *p, int value) +{ + while (p->name) { + if (value == p->value) { + break; + } + p++; + } + return ((Jim_Nvp *)(p)); +} + + +int +Jim_Nvp_value2name(Jim_Interp *interp, const Jim_Nvp *_p, int value, Jim_Nvp **result) +{ + const Jim_Nvp *p; + + p = Jim_Nvp_value2name_simple(_p, value); + + if (result) { + *result = (Jim_Nvp *)(p); + } + + if (p->name) { + return JIM_OK; + } else { + return JIM_ERR; + } +} + + +int +Jim_GetOpt_Setup(Jim_GetOptInfo *p, Jim_Interp *interp, int argc, Jim_Obj * const * argv) +{ + memset(p, 0, sizeof(*p)); + p->interp = interp; + p->argc = argc; + p->argv = argv; + + return JIM_OK; +} + +void +Jim_GetOpt_Debug(Jim_GetOptInfo *p) +{ + int x; + + Jim_fprintf(p->interp, p->interp->cookie_stderr, "---args---\n"); + for (x = 0 ; x < p->argc ; x++) { + Jim_fprintf(p->interp, p->interp->cookie_stderr, + "%2d) %s\n", + x, + Jim_GetString(p->argv[x], NULL)); + } + Jim_fprintf(p->interp, p->interp->cookie_stderr, "-------\n"); +} + + +int +Jim_GetOpt_Obj(Jim_GetOptInfo *goi, Jim_Obj **puthere) +{ + Jim_Obj *o; + + o = NULL; // failure + if (goi->argc) { + // success + o = goi->argv[0]; + goi->argc -= 1; + goi->argv += 1; + } + if (puthere) { + *puthere = o; + } + if (o != NULL) { + return JIM_OK; + } else { + return JIM_ERR; + } +} + +int +Jim_GetOpt_String(Jim_GetOptInfo *goi, char **puthere, int *len) +{ + int r; + Jim_Obj *o; + const char *cp; + + + r = Jim_GetOpt_Obj(goi, &o); + if (r == JIM_OK) { + cp = Jim_GetString(o, len); + if (puthere) { + /* remove const */ + *puthere = (char *)(cp); + } + } + return r; +} + +int +Jim_GetOpt_Double(Jim_GetOptInfo *goi, double *puthere) +{ + int r; + Jim_Obj *o; + double _safe; + + if (puthere == NULL) { + puthere = &_safe; + } + + r = Jim_GetOpt_Obj(goi, &o); + if (r == JIM_OK) { + r = Jim_GetDouble(goi->interp, o, puthere); + if (r != JIM_OK) { + Jim_SetResult_sprintf(goi->interp, + "not a number: %s", + Jim_GetString(o, NULL)); + } + } + return r; +} + +int +Jim_GetOpt_Wide(Jim_GetOptInfo *goi, jim_wide *puthere) +{ + int r; + Jim_Obj *o; + jim_wide _safe; + + if (puthere == NULL) { + puthere = &_safe; + } + + r = Jim_GetOpt_Obj(goi, &o); + if (r == JIM_OK) { + r = Jim_GetWide(goi->interp, o, puthere); + } + return r; +} + +int Jim_GetOpt_Nvp(Jim_GetOptInfo *goi, + const Jim_Nvp *nvp, + Jim_Nvp **puthere) +{ + Jim_Nvp *_safe; + Jim_Obj *o; + int e; + + if (puthere == NULL) { + puthere = &_safe; + } + + e = Jim_GetOpt_Obj(goi, &o); + if (e == JIM_OK) { + e = Jim_Nvp_name2value_obj(goi->interp, + nvp, + o, + puthere); + } + + return e; +} + +void +Jim_GetOpt_NvpUnknown(Jim_GetOptInfo *goi, + const Jim_Nvp *nvptable, + int hadprefix) +{ + if (hadprefix) { + Jim_SetResult_NvpUnknown(goi->interp, + goi->argv[-2], + goi->argv[-1], + nvptable); + } else { + Jim_SetResult_NvpUnknown(goi->interp, + NULL, + goi->argv[-1], + nvptable); + } +} + + +int +Jim_GetOpt_Enum(Jim_GetOptInfo *goi, + const char * const * lookup, + int *puthere) +{ + int _safe; + Jim_Obj *o; + int e; + + if (puthere == NULL) { + puthere = &_safe; + } + e = Jim_GetOpt_Obj(goi, &o); + if (e == JIM_OK) { + e = Jim_GetEnum(goi->interp, + o, + lookup, + puthere, + "option", + JIM_ERRMSG); + } + return e; +} + + + +int +Jim_SetResult_sprintf(Jim_Interp *interp, const char *fmt,...) +{ + va_list ap; + char *buf; + + va_start(ap,fmt); + buf = jim_vasprintf(fmt, ap); + va_end(ap); + if (buf) { + Jim_SetResultString(interp, buf, -1); + jim_vasprintf_done(buf); + } + return JIM_OK; +} + + +void +Jim_SetResult_NvpUnknown(Jim_Interp *interp, + Jim_Obj *param_name, + Jim_Obj *param_value, + const Jim_Nvp *nvp) +{ + if (param_name) { + Jim_SetResult_sprintf(interp, + "%s: Unknown: %s, try one of: ", + Jim_GetString(param_name, NULL), + Jim_GetString(param_value, NULL)); + } else { + Jim_SetResult_sprintf(interp, + "Unknown param: %s, try one of: ", + Jim_GetString(param_value, NULL)); + } + while (nvp->name) { + const char *a; + const char *b; + + if ((nvp + 1)->name) { + a = nvp->name; + b = ", "; + } else { + a = "or "; + b = nvp->name; + } + Jim_AppendStrings(interp, + Jim_GetResult(interp), + a, b, NULL); + nvp++; + } +} + + +static Jim_Obj *debug_string_obj; + +const char * +Jim_Debug_ArgvString(Jim_Interp *interp, int argc, Jim_Obj *const *argv) +{ + int x; + + if (debug_string_obj) { + Jim_FreeObj(interp, debug_string_obj); + } + + debug_string_obj = Jim_NewEmptyStringObj(interp); + for (x = 0 ; x < argc ; x++) { + Jim_AppendStrings(interp, + debug_string_obj, + Jim_GetString(argv[x], NULL), + " ", + NULL); + } + + return Jim_GetString(debug_string_obj, NULL); +} -- cgit v1.2.3