From 3492bff64a809b3a2a2376b83f41e099e16d22f6 Mon Sep 17 00:00:00 2001 From: Richard Purdie Date: Mon, 2 Aug 2010 10:20:20 +0100 Subject: bitbake: Add codeparser for parsing shell and python functions This commit is derived from Chris Larson's checksum work, turned into a standalone piece of code for parsing python and shell functions. The deindent code has been replaced with code to work around indentation for speed. The original NodeVisitor in the ast was replaced with a faster class walk call. Signed-off-by: Richard Purdie --- bitbake/lib/bb/codeparser.py | 273 +++++++++++++++++++++++++++++++++++++++++++ bitbake/lib/bb/data_smart.py | 6 +- 2 files changed, 276 insertions(+), 3 deletions(-) create mode 100644 bitbake/lib/bb/codeparser.py (limited to 'bitbake/lib/bb') diff --git a/bitbake/lib/bb/codeparser.py b/bitbake/lib/bb/codeparser.py new file mode 100644 index 000000000..88a26c82a --- /dev/null +++ b/bitbake/lib/bb/codeparser.py @@ -0,0 +1,273 @@ +from pysh import pyshyacc, pyshlex +from itertools import chain +from bb import msg, utils +import ast +import codegen + +def check_indent(codestr): + """If the code is indented, add a top level piece of code to 'remove' the indentation""" + + if codestr[0] is " " or codestr[0] is " ": + return "if 1:\n" + codestr + + return codestr + +pythonparsecache = {} + +class PythonParser(): + class ValueVisitor(): + """Visitor to traverse a python abstract syntax tree and obtain + the variables referenced via bitbake metadata APIs, and the external + functions called. + """ + + getvars = ("d.getVar", "bb.data.getVar", "data.getVar") + expands = ("d.expand", "bb.data.expand", "data.expand") + execs = ("bb.build.exec_func", "bb.build.exec_task") + + @classmethod + def _compare_name(cls, strparts, node): + """Given a sequence of strings representing a python name, + where the last component is the actual Name and the prior + elements are Attribute nodes, determine if the supplied node + matches. + """ + + if not strparts: + return True + + current, rest = strparts[0], strparts[1:] + if isinstance(node, ast.Attribute): + if current == node.attr: + return cls._compare_name(rest, node.value) + elif isinstance(node, ast.Name): + if current == node.id: + return True + return False + + @classmethod + def compare_name(cls, value, node): + """Convenience function for the _compare_node method, which + can accept a string (which is split by '.' for you), or an + iterable of strings, in which case it checks to see if any of + them match, similar to isinstance. + """ + + if isinstance(value, basestring): + return cls._compare_name(tuple(reversed(value.split("."))), + node) + else: + return any(cls.compare_name(item, node) for item in value) + + def __init__(self, value): + self.var_references = set() + self.var_execs = set() + self.direct_func_calls = set() + self.var_expands = set() + self.value = value + + @classmethod + def warn(cls, func, arg): + """Warn about calls of bitbake APIs which pass a non-literal + argument for the variable name, as we're not able to track such + a reference. + """ + + try: + funcstr = codegen.to_source(func) + argstr = codegen.to_source(arg) + except TypeError: + msg.debug(2, None, "Failed to convert function and argument to source form") + else: + msg.debug(1, None, "Warning: in call to '%s', argument '%s' is not a literal" % + (funcstr, argstr)) + + def visit_Call(self, node): + if self.compare_name(self.getvars, node.func): + if isinstance(node.args[0], ast.Str): + self.var_references.add(node.args[0].s) + else: + self.warn(node.func, node.args[0]) + elif self.compare_name(self.expands, node.func): + if isinstance(node.args[0], ast.Str): + self.warn(node.func, node.args[0]) + self.var_expands.update(node.args[0].s) + elif isinstance(node.args[0], ast.Call) and \ + self.compare_name(self.getvars, node.args[0].func): + pass + else: + self.warn(node.func, node.args[0]) + elif self.compare_name(self.execs, node.func): + if isinstance(node.args[0], ast.Str): + self.var_execs.add(node.args[0].s) + else: + self.warn(node.func, node.args[0]) + elif isinstance(node.func, ast.Name): + self.direct_func_calls.add(node.func.id) + elif isinstance(node.func, ast.Attribute): + # We must have a qualified name. Therefore we need + # to walk the chain of 'Attribute' nodes to determine + # the qualification. + attr_node = node.func.value + identifier = node.func.attr + while isinstance(attr_node, ast.Attribute): + identifier = attr_node.attr + "." + identifier + attr_node = attr_node.value + if isinstance(attr_node, ast.Name): + identifier = attr_node.id + "." + identifier + self.direct_func_calls.add(identifier) + + def __init__(self): + #self.funcdefs = set() + self.execs = set() + #self.external_cmds = set() + self.references = set() + + def parse_python(self, node): + + if node in pythonparsecache: + self.references = pythonparsecache[node].references + self.execs = pythonparsecache[node].execs + return + + code = compile(check_indent(str(node)), "", "exec", + ast.PyCF_ONLY_AST) + + visitor = self.ValueVisitor(code) + for n in ast.walk(code): + if n.__class__.__name__ == "Call": + visitor.visit_Call(n) + + self.references.update(visitor.var_references) + self.references.update(visitor.var_execs) + self.execs = visitor.direct_func_calls + + pythonparsecache[node] = self + + +shellparsecache = {} + +class ShellParser(): + def __init__(self): + self.funcdefs = set() + self.allexecs = set() + self.execs = set() + + def parse_shell(self, value): + """Parse the supplied shell code in a string, returning the external + commands it executes. + """ + + if value in pythonparsecache: + self.execs = shellparsecache[value].execs + return + + try: + tokens, _ = pyshyacc.parse(value, eof=True, debug=False) + except pyshlex.NeedMore: + raise ShellSyntaxError("Unexpected EOF") + + for token in tokens: + self.process_tokens(token) + self.execs = set(cmd for cmd in self.allexecs if cmd not in self.funcdefs) + + shellparsecache[value] = self + + return self.execs + + def process_tokens(self, tokens): + """Process a supplied portion of the syntax tree as returned by + pyshyacc.parse. + """ + + def function_definition(value): + self.funcdefs.add(value.name) + return [value.body], None + + def case_clause(value): + # Element 0 of each item in the case is the list of patterns, and + # Element 1 of each item in the case is the list of commands to be + # executed when that pattern matches. + words = chain(*[item[0] for item in value.items]) + cmds = chain(*[item[1] for item in value.items]) + return cmds, words + + def if_clause(value): + main = chain(value.cond, value.if_cmds) + rest = value.else_cmds + if isinstance(rest, tuple) and rest[0] == "elif": + return chain(main, if_clause(rest[1])) + else: + return chain(main, rest) + + def simple_command(value): + return None, chain(value.words, (assign[1] for assign in value.assigns)) + + token_handlers = { + "and_or": lambda x: ((x.left, x.right), None), + "async": lambda x: ([x], None), + "brace_group": lambda x: (x.cmds, None), + "for_clause": lambda x: (x.cmds, x.items), + "function_definition": function_definition, + "if_clause": lambda x: (if_clause(x), None), + "pipeline": lambda x: (x.commands, None), + "redirect_list": lambda x: ([x.cmd], None), + "subshell": lambda x: (x.cmds, None), + "while_clause": lambda x: (chain(x.condition, x.cmds), None), + "until_clause": lambda x: (chain(x.condition, x.cmds), None), + "simple_command": simple_command, + "case_clause": case_clause, + } + + for token in tokens: + name, value = token + try: + more_tokens, words = token_handlers[name](value) + except KeyError: + raise NotImplementedError("Unsupported token type " + name) + + if more_tokens: + self.process_tokens(more_tokens) + + if words: + self.process_words(words) + + def process_words(self, words): + """Process a set of 'words' in pyshyacc parlance, which includes + extraction of executed commands from $() blocks, as well as grabbing + the command name argument. + """ + + words = list(words) + for word in list(words): + wtree = pyshlex.make_wordtree(word[1]) + for part in wtree: + if not isinstance(part, list): + continue + + if part[0] in ('`', '$('): + command = pyshlex.wordtree_as_string(part[1:-1]) + self.parse_shell(command) + + if word[0] in ("cmd_name", "cmd_word"): + if word in words: + words.remove(word) + + usetoken = False + for word in words: + if word[0] in ("cmd_name", "cmd_word") or \ + (usetoken and word[0] == "TOKEN"): + if "=" in word[1]: + usetoken = True + continue + + cmd = word[1] + if cmd.startswith("$"): + msg.debug(1, None, "Warning: execution of non-literal command '%s'" % cmd) + elif cmd == "eval": + command = " ".join(word for _, word in words[1:]) + self.parse_shell(command) + else: + self.allexecs.add(cmd) + break + diff --git a/bitbake/lib/bb/data_smart.py b/bitbake/lib/bb/data_smart.py index 1ed04d50c..b9d9476fd 100644 --- a/bitbake/lib/bb/data_smart.py +++ b/bitbake/lib/bb/data_smart.py @@ -46,7 +46,7 @@ class VariableParse: self.value = val self.references = set() - self.funcrefs = set() + self.execs = set() def var_sub(self, match): key = match.group()[2:-1] @@ -64,10 +64,10 @@ class VariableParse: code = match.group()[3:-1] codeobj = compile(code.strip(), self.varname or "", "eval") - parser = bb.rptest.PythonParser() + parser = bb.codeparser.PythonParser() parser.parse_python(code) self.references |= parser.references - self.funcrefs |= parser.execs + self.execs |= parser.execs value = utils.better_eval(codeobj, {"d": self.d}) return str(value) -- cgit v1.2.3