diff options
Diffstat (limited to 'Tools/c-analyzer/c_symbols')
| -rw-r--r-- | Tools/c-analyzer/c_symbols/__init__.py | 0 | ||||
| -rw-r--r-- | Tools/c-analyzer/c_symbols/binary.py | 157 | ||||
| -rw-r--r-- | Tools/c-analyzer/c_symbols/info.py | 51 | ||||
| -rw-r--r-- | Tools/c-analyzer/c_symbols/resolve.py | 147 | ||||
| -rw-r--r-- | Tools/c-analyzer/c_symbols/source.py | 58 |
5 files changed, 0 insertions, 413 deletions
diff --git a/Tools/c-analyzer/c_symbols/__init__.py b/Tools/c-analyzer/c_symbols/__init__.py deleted file mode 100644 index e69de29bb2..0000000000 --- a/Tools/c-analyzer/c_symbols/__init__.py +++ /dev/null diff --git a/Tools/c-analyzer/c_symbols/binary.py b/Tools/c-analyzer/c_symbols/binary.py deleted file mode 100644 index e125dbd5b5..0000000000 --- a/Tools/c-analyzer/c_symbols/binary.py +++ /dev/null @@ -1,157 +0,0 @@ -import os -import os.path -import shutil -import sys - -from c_analyzer_common import util, info -from . import source -from .info import Symbol - - -#PYTHON = os.path.join(REPO_ROOT, 'python') -PYTHON = sys.executable - - -def iter_symbols(binary=PYTHON, dirnames=None, *, - # Alternately, use look_up_known_symbol() - # from c_globals.supported. - find_local_symbol=source.find_symbol, - _file_exists=os.path.exists, - _iter_symbols_nm=(lambda b, *a: _iter_symbols_nm(b, *a)), - ): - """Yield a Symbol for each symbol found in the binary.""" - if not _file_exists(binary): - raise Exception('executable missing (need to build it first?)') - - if find_local_symbol: - cache = {} - def find_local_symbol(name, *, _find=find_local_symbol): - return _find(name, dirnames, _perfilecache=cache) - else: - find_local_symbol = None - - if os.name == 'nt': - # XXX Support this. - raise NotImplementedError - else: - yield from _iter_symbols_nm(binary, find_local_symbol) - - -############################# -# binary format (e.g. ELF) - -SPECIAL_SYMBOLS = { - '__bss_start', - '__data_start', - '__dso_handle', - '_DYNAMIC', - '_edata', - '_end', - '__environ@@GLIBC_2.2.5', - '_GLOBAL_OFFSET_TABLE_', - '__JCR_END__', - '__JCR_LIST__', - '__TMC_END__', - } - - -def _is_special_symbol(name): - if name in SPECIAL_SYMBOLS: - return True - if '@@GLIBC' in name: - return True - return False - - -############################# -# "nm" - -NM_KINDS = { - 'b': Symbol.KIND.VARIABLE, # uninitialized - 'd': Symbol.KIND.VARIABLE, # initialized - #'g': Symbol.KIND.VARIABLE, # uninitialized - #'s': Symbol.KIND.VARIABLE, # initialized - 't': Symbol.KIND.FUNCTION, - } - - -def _iter_symbols_nm(binary, find_local_symbol=None, - *, - _which=shutil.which, - _run=util.run_cmd, - ): - nm = _which('nm') - if not nm: - raise NotImplementedError - argv = [nm, - '--line-numbers', - binary, - ] - try: - output = _run(argv) - except Exception: - if nm is None: - # XXX Use dumpbin.exe /SYMBOLS on Windows. - raise NotImplementedError - raise - for line in output.splitlines(): - (name, kind, external, filename, funcname, vartype, - ) = _parse_nm_line(line, - _find_local_symbol=find_local_symbol, - ) - if kind != Symbol.KIND.VARIABLE: - continue - elif _is_special_symbol(name): - continue - assert vartype is None - yield Symbol( - id=(filename, funcname, name), - kind=kind, - external=external, - ) - - -def _parse_nm_line(line, *, _find_local_symbol=None): - _origline = line - _, _, line = line.partition(' ') # strip off the address - line = line.strip() - - kind, _, line = line.partition(' ') - line = line.strip() - external = kind.isupper() - kind = NM_KINDS.get(kind.lower(), Symbol.KIND.OTHER) - - name, _, filename = line.partition('\t') - name = name.strip() - if filename: - filename = os.path.relpath(filename.partition(':')[0]) - else: - filename = info.UNKNOWN - - vartype = None - name, islocal = _parse_nm_name(name, kind) - if islocal: - funcname = info.UNKNOWN - if _find_local_symbol is not None: - filename, funcname, vartype = _find_local_symbol(name) - filename = filename or info.UNKNOWN - funcname = funcname or info.UNKNOWN - else: - funcname = None - # XXX fine filename and vartype? - return name, kind, external, filename, funcname, vartype - - -def _parse_nm_name(name, kind): - if kind != Symbol.KIND.VARIABLE: - return name, None - if _is_special_symbol(name): - return name, None - - actual, sep, digits = name.partition('.') - if not sep: - return name, False - - if not digits.isdigit(): - raise Exception(f'got bogus name {name}') - return actual, True diff --git a/Tools/c-analyzer/c_symbols/info.py b/Tools/c-analyzer/c_symbols/info.py deleted file mode 100644 index f6ed52c8f0..0000000000 --- a/Tools/c-analyzer/c_symbols/info.py +++ /dev/null @@ -1,51 +0,0 @@ -from collections import namedtuple - -from c_analyzer_common.info import ID -from c_analyzer_common.util import classonly, _NTBase - - -class Symbol(_NTBase, namedtuple('Symbol', 'id kind external')): - """Info for a single compilation symbol.""" - - __slots__ = () - - class KIND: - VARIABLE = 'variable' - FUNCTION = 'function' - OTHER = 'other' - - @classonly - def from_name(cls, name, filename=None, kind=KIND.VARIABLE, external=None): - """Return a new symbol based on the given name.""" - id = ID(filename, None, name) - return cls(id, kind, external) - - def __new__(cls, id, kind=KIND.VARIABLE, external=None): - self = super().__new__( - cls, - id=ID.from_raw(id), - kind=str(kind) if kind else None, - external=bool(external) if external is not None else None, - ) - return self - - def __hash__(self): - return hash(self.id) - - def __getattr__(self, name): - return getattr(self.id, name) - - def validate(self): - """Fail if the object is invalid (i.e. init with bad data).""" - if not self.id: - raise TypeError('missing id') - else: - self.id.validate() - - if not self.kind: - raise TypeError('missing kind') - elif self.kind not in vars(self.KIND).values(): - raise ValueError(f'unsupported kind {self.kind}') - - if self.external is None: - raise TypeError('missing external') diff --git a/Tools/c-analyzer/c_symbols/resolve.py b/Tools/c-analyzer/c_symbols/resolve.py deleted file mode 100644 index 56210cefeb..0000000000 --- a/Tools/c-analyzer/c_symbols/resolve.py +++ /dev/null @@ -1,147 +0,0 @@ -import os.path - -from c_analyzer_common import files -from c_analyzer_common.info import UNKNOWN -from c_parser import declarations, info -from .info import Symbol -from .source import _find_symbol - - -# XXX need tests: -# * look_up_known_symbol() -# * symbol_from_source() -# * get_resolver() -# * symbols_to_variables() - -def look_up_known_symbol(symbol, knownvars, *, - match_files=(lambda f1, f2: f1 == f2), - ): - """Return the known variable matching the given symbol. - - "knownvars" is a mapping of common.ID to parser.Variable. - - "match_files" is used to verify if two filenames point to - the same file. - """ - if not knownvars: - return None - - if symbol.funcname == UNKNOWN: - if not symbol.filename or symbol.filename == UNKNOWN: - for varid in knownvars: - if not varid.funcname: - continue - if varid.name == symbol.name: - return knownvars[varid] - else: - return None - else: - for varid in knownvars: - if not varid.funcname: - continue - if not match_files(varid.filename, symbol.filename): - continue - if varid.name == symbol.name: - return knownvars[varid] - else: - return None - elif not symbol.filename or symbol.filename == UNKNOWN: - raise NotImplementedError - else: - return knownvars.get(symbol.id) - - -def find_in_source(symbol, dirnames, *, - _perfilecache={}, - _find_symbol=_find_symbol, - _iter_files=files.iter_files_by_suffix, - ): - """Return the Variable matching the given Symbol. - - If there is no match then return None. - """ - if symbol.filename and symbol.filename != UNKNOWN: - filenames = [symbol.filename] - else: - filenames = _iter_files(dirnames, ('.c', '.h')) - - if symbol.funcname and symbol.funcname != UNKNOWN: - raise NotImplementedError - - (filename, funcname, decl - ) = _find_symbol(symbol.name, filenames, _perfilecache) - if filename == UNKNOWN: - return None - return info.Variable.from_parts(filename, funcname, symbol.name, decl) - - -def get_resolver(knownvars=None, dirnames=None, *, - _look_up_known=look_up_known_symbol, - _from_source=find_in_source, - ): - """Return a "resolver" func for the given known vars and dirnames. - - The func takes a single Symbol and returns a corresponding Variable. - If the symbol was located then the variable will be valid, populated - with the corresponding information. Otherwise None is returned. - """ - if knownvars: - knownvars = dict(knownvars) # a copy - def resolve_known(symbol): - found = _look_up_known(symbol, knownvars) - if found is None: - return None - elif symbol.funcname == UNKNOWN: - knownvars.pop(found.id) - elif not symbol.filename or symbol.filename == UNKNOWN: - knownvars.pop(found.id) - return found - if dirnames: - def resolve(symbol): - found = resolve_known(symbol) - if found is None: - return None - #return _from_source(symbol, dirnames) - else: - for dirname in dirnames: - if not dirname.endswith(os.path.sep): - dirname += os.path.sep - if found.filename.startswith(dirname): - break - else: - return None - return found - else: - resolve = resolve_known - elif dirnames: - def resolve(symbol): - return _from_source(symbol, dirnames) - else: - def resolve(symbol): - return None - return resolve - - -def symbols_to_variables(symbols, *, - resolve=(lambda s: look_up_known_symbol(s, None)), - ): - """Yield the variable the matches each given symbol. - - Use get_resolver() for a "resolve" func to use. - """ - for symbol in symbols: - if isinstance(symbol, info.Variable): - # XXX validate? - yield symbol - continue - if symbol.kind != Symbol.KIND.VARIABLE: - continue - resolved = resolve(symbol) - if resolved is None: - #raise NotImplementedError(symbol) - resolved = info.Variable( - id=symbol.id, - storage=UNKNOWN, - vartype=UNKNOWN, - ) - yield resolved diff --git a/Tools/c-analyzer/c_symbols/source.py b/Tools/c-analyzer/c_symbols/source.py deleted file mode 100644 index a7248104c9..0000000000 --- a/Tools/c-analyzer/c_symbols/source.py +++ /dev/null @@ -1,58 +0,0 @@ -from c_analyzer_common import files -from c_analyzer_common.info import UNKNOWN -from c_parser import declarations - - -# XXX need tests: -# * find_symbol() - -def find_symbol(name, dirnames, *, - _perfilecache, - _iter_files=files.iter_files_by_suffix, - **kwargs - ): - """Return (filename, funcname, vartype) for the matching Symbol.""" - filenames = _iter_files(dirnames, ('.c', '.h')) - return _find_symbol(name, filenames, _perfilecache, **kwargs) - - -def _get_symbols(filename, *, - _iter_variables=declarations.iter_variables, - ): - """Return the list of Symbols found in the given file.""" - symbols = {} - for funcname, name, vartype in _iter_variables(filename): - if not funcname: - continue - try: - instances = symbols[name] - except KeyError: - instances = symbols[name] = [] - instances.append((funcname, vartype)) - return symbols - - -def _find_symbol(name, filenames, _perfilecache, *, - _get_local_symbols=_get_symbols, - ): - for filename in filenames: - try: - symbols = _perfilecache[filename] - except KeyError: - symbols = _perfilecache[filename] = _get_local_symbols(filename) - - try: - instances = symbols[name] - except KeyError: - continue - - funcname, vartype = instances.pop(0) - if not instances: - symbols.pop(name) - return filename, funcname, vartype - else: - return UNKNOWN, UNKNOWN, UNKNOWN - - -def iter_symbols(): - raise NotImplementedError |