parser/lexer:

o Support special '/DISCARD/' as NAME.
o Add 'SORT' as a token. It is equivalent with SORT_BY_NAME. The new rule sort_by_name coverts both.
Ld:
o Don't die on unknown sections, they can be referenced before declared. Doesn't matter for this app anyway.

5 files changed, 74 insertions, 14 deletions

diff --git a/GnuLdLexer.g4 b/GnuLdLexer.g4
index 16c4a9e..eb6090f 100644
--- a/GnuLdLexer.g4
+++ b/GnuLdLexer.g4
@@ -131,6 +131,7 @@ SIZEOF_HEADERS : 'SIZEOF_HEADERS';
 SIZEOF : 'SIZEOF';
 SORT_BY_ALIGNMENT : 'SORT_BY_ALIGNMENT';
 SORT_BY_INIT_PRIORITY : 'SORT_BY_INIT_PRIORITY';
+SORT : 'SORT';
 SORT_BY_NAME : 'SORT_BY_NAME';
 SORT_NONE : 'SORT_NONE';
 SPECIAL : 'SPECIAL';
@@ -159,7 +160,7 @@ QUOTE : '"' -> skip, pushMode(STRING);
 //  | NAME                     # namePlain;
 
 //NAME : [\._a-zA-Z][\._a-zA-Z0-9]*;
-NAME : [*\._a-zA-Z][*\.\/_a-zA-Z0-9]*;
+NAME : [*\._a-zA-Z][*\.\/_a-zA-Z0-9]* | '/DISCARD/';
 
 // TODO: ld supports some really fancy expressions here, like "0101010b", "ffH", "ffx", "$Aa" etc
 //INT : '0x' [0-9a-fA-F]+
diff --git a/GnuLdParser.g4 b/GnuLdParser.g4
index 36d4515..0d2bf4b 100644
--- a/GnuLdParser.g4
+++ b/GnuLdParser.g4
@@ -213,17 +213,23 @@ wildcard_name:
     |   QUESTION
     ;
 
+// SORT(*) and SORT_BY_NAME(*) both mean the same
+sort_by_name:
+        SORT
+    |   SORT_BY_NAME
+    ;
+
 wildcard_spec:
         wildcard_name
     |   EXCLUDE_FILE LPAREN exclude_name_list RPAREN wildcard_name
-    |   SORT_BY_NAME LPAREN wildcard_name RPAREN
+    |   sort_by_name LPAREN wildcard_name RPAREN
     |   SORT_BY_ALIGNMENT LPAREN wildcard_name RPAREN
     |   SORT_NONE LPAREN wildcard_name RPAREN
-    |   SORT_BY_NAME LPAREN SORT_BY_ALIGNMENT LPAREN wildcard_name RPAREN RPAREN
-    |   SORT_BY_NAME LPAREN SORT_BY_NAME LPAREN wildcard_name RPAREN RPAREN
-    |   SORT_BY_ALIGNMENT LPAREN SORT_BY_NAME LPAREN wildcard_name RPAREN RPAREN
+    |   sort_by_name LPAREN SORT_BY_ALIGNMENT LPAREN wildcard_name RPAREN RPAREN
+    |   sort_by_name LPAREN sort_by_name LPAREN wildcard_name RPAREN RPAREN
+    |   SORT_BY_ALIGNMENT LPAREN sort_by_name LPAREN wildcard_name RPAREN RPAREN
     |   SORT_BY_ALIGNMENT LPAREN SORT_BY_ALIGNMENT LPAREN wildcard_name RPAREN RPAREN
-    |   SORT_BY_NAME LPAREN EXCLUDE_FILE LPAREN exclude_name_list RPAREN wildcard_name RPAREN
+    |   sort_by_name LPAREN EXCLUDE_FILE LPAREN exclude_name_list RPAREN wildcard_name RPAREN
     |   SORT_BY_INIT_PRIORITY LPAREN wildcard_name RPAREN
     ;
 
@@ -265,7 +271,7 @@ statement:
     |   CREATE_OBJECT_SYMBOLS
     |   SEMICOLON
     |   CONSTRUCTORS
-    |   SORT_BY_NAME LPAREN CONSTRUCTORS RPAREN
+    |   sort_by_name LPAREN CONSTRUCTORS RPAREN
     |   input_section_spec
     |   length LPAREN mustbe_exp RPAREN
     |   FILL LPAREN fill_exp RPAREN
diff --git a/Ld.cpp b/Ld.cpp
index a60656c..8bacaea 100644
--- a/Ld.cpp
+++ b/Ld.cpp
@@ -39,14 +39,18 @@ public:
     }
 
     virtual V get(ParseTree *const node) {
+        return _annotations.at(node);
+    }
+
+//    virtual V get(ParseTree *const node) {
 //        try {
 //             cout << "node = " << node->getText() << endl;
-            return _annotations.at(node);
+//        return _annotations.at(node);
 //        } catch (std::out_of_range &e) {
 //             cout << "out of range: " << node->getText() << endl;
 //            throw e;
 //        }
-    }
+//    }
 
     virtual void put(ParseTree *const node, V value) {
         // cout << "put(" << node << ", " << value << ")" << endl;
@@ -88,7 +92,8 @@ public:
                 return s;
             }
         }
-        throw out_of_range("No such section: " + name);
+        sections.emplace_back(Section{.name = name});
+        return sections[sections.size() - 1];
     }
 
     static uint64_t parseInt(const string &s) {
@@ -218,8 +223,34 @@ public:
 LdScriptLoader::LdScriptLoader() : debug_(false) {
 }
 
+class LdErrorListener : public BaseErrorListener {
+public:
+    std::vector<std::string> messages;
+
+    void syntaxError(IRecognizer *recognizer, Token *offendingSymbol, size_t line, int charPositionInLine,
+                     const std::string &msg, std::exception_ptr e) override {
+        messages.push_back(msg);
+    }
+};
+
 LdScript LdScriptLoader::load(std::string path) {
-    ANTLRFileStream input(path);
+    std::ifstream stream(path, std::ios::binary);
+
+    if (!stream.good() || stream.eof()) {
+        return {};
+    }
+
+    string data;
+    while (!stream.eof()) {
+        char buf[1000];
+        stream.read(buf, sizeof(buf));
+        auto count = stream.gcount();
+        for (int i = 0; i < count; i++) {
+            data.push_back((char) (buf[i] & 0x7f));
+        }
+    }
+
+    ANTLRInputStream input(data);
     GnuLdLexer lexer(&input);
     CommonTokenStream tokens(&lexer);
     tokens.fill();
@@ -233,9 +264,17 @@ LdScript LdScriptLoader::load(std::string path) {
     GnuLdParser parser(&tokens);
     ElfinfoGnuLdBaseListener listener;
     parser.addParseListener(&listener);
+
+    LdErrorListener ldErrorListener;
+    parser.addErrorListener(&ldErrorListener);
+
     auto file = parser.file();
 
-    if (debug_) {
+    if (!ldErrorListener.messages.empty()) {
+        throw LdParseException(ldErrorListener.messages);
+    }
+
+    if (debug_ && parser.getNumberOfSyntaxErrors() == 0) {
         std::cout << file->toStringTree(&parser) << std::endl << std::endl;
     }
 
diff --git a/elfinfo.cpp b/elfinfo.cpp
index 6ddbceb..985a2de 100644
--- a/elfinfo.cpp
+++ b/elfinfo.cpp
@@ -192,10 +192,17 @@ int main(int argc, char **argv) {
             LdScript file = loader.load(ld_filename);
 
             memoryAreas.insert(memoryAreas.end(), file.memoryAreas.begin(), file.memoryAreas.end());
+        } catch (LdParseException &e) {
+            for (auto &m: e.messages) {
+                fprintf(stderr, m.c_str());
+            }
+            errx(EX_DATAERR, "Could not parse ld script");
         } catch (std::exception &e) {
-            errx(EX_DATAERR, "Unhandled exception when parsing LD script. The parser is new and fragile, so please file a bug and send me your LD script: %s", e.what());
+            errx(EX_DATAERR, "Unhandled exception when parsing LD script. "
+                "The parser is new and fragile, so please file a bug and send me your LD script: %s", e.what());
         } catch (...) {
-            errx(EX_DATAERR, "Unhandled exception when parsing LD script. The parser is new and fragile, so please file a bug and send me your LD script.");
+            errx(EX_DATAERR, "Unhandled exception when parsing LD script. "
+                "The parser is new and fragile, so please file a bug and send me your LD script.");
         }
     }
 
diff --git a/includes/trygvis/elfinfo/Ld.h b/includes/trygvis/elfinfo/Ld.h
index cab466d..09dcaec 100644
--- a/includes/trygvis/elfinfo/Ld.h
+++ b/includes/trygvis/elfinfo/Ld.h
@@ -8,6 +8,13 @@
 namespace trygvis {
 namespace elfinfo {
 
+class LdParseException : public std::runtime_error {
+public:
+    explicit LdParseException(const std::vector<std::string> messages) : runtime_error("Parse error"), messages(messages) {}
+
+    const std::vector<std::string> messages;
+};
+
 enum class MemoryAttribute {
     R, W, X
 };