check_style_c.py

re_ifndef = re.compile("^\s*#\s*ifndef\s+([A-z0-9_]+).*$")
re_define = re.compile("^\s*#\s*define\s+([A-z0-9_]+).*$")


def quick_check_include_guard(lines):
    # found = 0
    def_value = ""
    ok = False

    def fn_as_guard(fn):
        name = os.path.basename(fn).upper().replace(".", "_").replace("-", "_")
        return "__%s__" % name

    for i, l in enumerate(lines):
        ndef_match = re_ifndef.match(l)
        if ndef_match:
            ndef_value = ndef_match.group(1).strip()
            for j in range(i + 1, len(lines)):
                l_next = lines[j]
                def_match = re_define.match(l_next)
                if def_match:
                    def_value = def_match.group(1).strip()
                    if def_value == ndef_value:
                        ok = True
                        break
                elif l_next.strip():
                    # print(filepath)
                    # found non empty non ndef line. quit
                    break
                else:
                    # allow blank lines
                    pass
            break

    guard = fn_as_guard(filepath)

    if ok:
        # print("found:", def_value, "->", filepath)
        if def_value != guard:
            # print("%s: %s -> %s" % (filepath, def_value, guard))
            warning_lineonly("E147", "non-conforming include guard (found %r, expected %r)" %
                             (def_value, guard), i + 1)
    else:
        warning_lineonly("E148", "missing include guard %r" % guard, 1)


def quick_check_source(fp, code, args):

    global filepath

    is_header = fp.endswith((".h", ".hxx", ".hpp"))

    filepath = fp

    lines = code.split("\n")

    if is_header:
        quick_check_include_guard(lines)

    quick_check_indentation(lines)


def scan_source(fp, code, args):
    # print("scanning: %r" % fp)

    global filepath

    is_cpp = fp.endswith((".cpp", ".cxx"))

    filepath = fp

    # if "displist.c" not in filepath:
    #     return

    filepath_base = os.path.basename(filepath)
    filepath_split = filepath.split(os.sep)

    # print(highlight(code, CLexer(), RawTokenFormatter()).decode('utf-8'))

    del tokens[:]
    line = 1

    for ttype, text in lex(code, CLexer()):
        if text:
            tokens.append(TokStore(ttype, text, line))
            line += text.count("\n")

    col = 0  # track line length
    index_line_start = 0

    for i, tok in enumerate(tokens):
        # print(tok.type, tok.text)
        if tok.type == Token.Keyword:
            if tok.text in {"switch", "while", "if", "for"}:
                item_range = extract_statement_if(i)
                if item_range is not None:
                    blender_check_kw_if(item_range[0], i, item_range[1])
                if tok.text == "switch":
                    blender_check_kw_switch(item_range[0], i, item_range[1])
            elif tok.text == "else":
                blender_check_kw_else(i)
            elif tok.text == "sizeof":
                blender_check_kw_sizeof(i)
        elif tok.type == Token.Punctuation:
            if tok.text == ",":
                blender_check_comma(i)
            elif tok.text == ".":
                blender_check_period(i)
            elif tok.text == "[":
                # note, we're quite relaxed about this but
                # disallow 'foo ['
                if tokens[i - 1].text.isspace():
                    if is_cpp and tokens[i + 1].text == "]":
                        # c++ can do delete []
                        pass
                    else:
                        warning("E149", "space before '['", i, i)
            elif tok.text == "(":
                # check if this is a cast, eg:
                #  (char), (char **), (float (*)[3])
                item_range = extract_cast(i)
                if item_range is not None:
                    blender_check_cast(item_range[0], item_range[1])
            elif tok.text == "{":
                # check matching brace is indented correctly (slow!)
                blender_check_brace_indent(i)

                # check previous character is either a '{' or whitespace.
                if (tokens[i - 1].line == tok.line) and not (tokens[i - 1].text.isspace() or tokens[i - 1].text == "{"):
                    warning("E150", "no space before '{'", i, i)

                blender_check_function_definition(i)

        elif tok.type == Token.Operator:
            # we check these in pairs, only want first
            if tokens[i - 1].type != Token.Operator:
                op, index_kw_end = extract_operator(i)
                blender_check_operator(i, index_kw_end, op, is_cpp)
        elif tok.type in Token.Comment:
            doxyfn = None
            if "\\file" in tok.text:
                doxyfn = tok.text.split("\\file", 1)[1].strip().split()[0]
            elif "@file" in tok.text:
                doxyfn = tok.text.split("@file", 1)[1].strip().split()[0]

            if doxyfn is not None:
                doxyfn_base = os.path.basename(doxyfn)
                if doxyfn_base != filepath_base:
                    warning("E151", "doxygen filename mismatch %s != %s" % (doxyfn_base, filepath_base), i, i)
                doxyfn_split = doxyfn.split("/")
                if len(doxyfn_split) > 1:
                    fp_split = filepath_split[-len(doxyfn_split):]
                    if doxyfn_split != fp_split:
                        warning("E151", "doxygen filepath mismatch %s != %s" % (doxyfn, "/".join(fp_split)), i, i)
                    del fp_split
                del doxyfn_base, doxyfn_split
            del doxyfn

        # ensure line length
        if (not args.no_length_check) and tok.type == Token.Text and tok.text == "\n":
            # check line len
            blender_check_linelength(index_line_start, i - 1, col)

            col = 0
            index_line_start = i + 1
        else:
            col += len(tok.text.expandtabs(TAB_SIZE))

        #elif tok.type == Token.Name:
        #    print(tok.text)

        #print(ttype, type(ttype))
        #print((ttype, value))

    #for ttype, value in la:
    #    #print(value, end="")


def scan_source_filepath(filepath, args):
    # for quick tests
    #~ if not filepath.endswith("creator.c"):
    #~     return

    code = open(filepath, 'r', encoding="utf-8").read()

    # fast checks which don't require full parsing
    quick_check_source(filepath, code, args)

    # use lexer
    scan_source(filepath, code, args)


def scan_source_recursive(dirpath, args):
    import os
    from os.path import join, splitext

    def source_list(path, filename_check=None):
        for dirpath, dirnames, filenames in os.walk(path):

            # skip '.svn'
            if dirpath.startswith("."):
                continue

            for filename in filenames:
                filepath = join(dirpath, filename)
                if filename_check is None or filename_check(filepath):
                    yield filepath

    def is_source(filename):
        ext = splitext(filename)[1]
        return (ext in {".c", ".inl", ".cpp", ".cxx", ".cc", ".hpp", ".hxx", ".h", ".hh", ".osl"})

    for filepath in sorted(source_list(dirpath, is_source)):
        if is_ignore(filepath):
            continue

        scan_source_filepath(filepath, args)


def create_parser():
    parser = argparse.ArgumentParser(
            description=(
            "Check C/C++ code for conformance with blenders style guide:\n"
            "http://wiki.blender.org/index.php/Dev:Doc/CodeStyle)")
            )
    parser.add_argument(
            "paths",
            nargs='+',
            help="list of files or directories to check",
            )
    parser.add_argument(
            "-l",
            "--no-length-check",
            action="store_true",
            help="skip warnings for long lines",
            )
    return parser


def main(argv=None):
    import sys
    import os

    if argv is None:
        argv = sys.argv[1:]

    parser = create_parser()
    args = parser.parse_args(argv)
    del argv

    print("Scanning:", SOURCE_DIR)

    if 0:
        # SOURCE_DIR = os.path.normpath(os.path.abspath(os.path.normpath(os.path.join(os.path.dirname(__file__), "..", "..", ".."))))
        # scan_source_recursive(os.path.join(SOURCE_DIR, "source", "blender", "bmesh"))
        scan_source_recursive(os.path.join(SOURCE_DIR, "source/blender/makesrna/intern"), args)
        sys.exit(0)

    for filepath in args.paths:
        if os.path.isdir(filepath):
            # recursive search
            scan_source_recursive(filepath, args)
        else:
            # single file
            scan_source_filepath(filepath, args)


if __name__ == "__main__":
    main()