• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1#!/usr/bin/env python3
2#
3# Copyright The Mbed TLS Contributors
4# SPDX-License-Identifier: Apache-2.0
5#
6# Licensed under the Apache License, Version 2.0 (the "License"); you may
7# not use this file except in compliance with the License.
8# You may obtain a copy of the License at
9#
10# http://www.apache.org/licenses/LICENSE-2.0
11#
12# Unless required by applicable law or agreed to in writing, software
13# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
14# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15# See the License for the specific language governing permissions and
16# limitations under the License.
17
18"""
19This script confirms that the naming of all symbols and identifiers in Mbed TLS
20are consistent with the house style and are also self-consistent. It only runs
21on Linux and macOS since it depends on nm.
22
23It contains two major Python classes, CodeParser and NameChecker. They both have
24a comprehensive "run-all" function (comprehensive_parse() and perform_checks())
25but the individual functions can also be used for specific needs.
26
27CodeParser makes heavy use of regular expressions to parse the code, and is
28dependent on the current code formatting. Many Python C parser libraries require
29preprocessed C code, which means no macro parsing. Compiler tools are also not
30very helpful when we want the exact location in the original source (which
31becomes impossible when e.g. comments are stripped).
32
33NameChecker performs the following checks:
34
35- All exported and available symbols in the library object files, are explicitly
36  declared in the header files. This uses the nm command.
37- All macros, constants, and identifiers (function names, struct names, etc)
38  follow the required regex pattern.
39- Typo checking: All words that begin with MBED exist as macros or constants.
40
41The script returns 0 on success, 1 on test failure, and 2 if there is a script
42error. It must be run from Mbed TLS root.
43"""
44
45import abc
46import argparse
47import fnmatch
48import glob
49import textwrap
50import os
51import sys
52import traceback
53import re
54import enum
55import shutil
56import subprocess
57import logging
58
59# Naming patterns to check against. These are defined outside the NameCheck
60# class for ease of modification.
61MACRO_PATTERN = r"^(MBEDTLS|PSA)_[0-9A-Z_]*[0-9A-Z]$"
62CONSTANTS_PATTERN = MACRO_PATTERN
63IDENTIFIER_PATTERN = r"^(mbedtls|psa)_[0-9a-z_]*[0-9a-z]$"
64
65class Match(): # pylint: disable=too-few-public-methods
66    """
67    A class representing a match, together with its found position.
68
69    Fields:
70    * filename: the file that the match was in.
71    * line: the full line containing the match.
72    * line_no: the line number.
73    * pos: a tuple of (start, end) positions on the line where the match is.
74    * name: the match itself.
75    """
76    def __init__(self, filename, line, line_no, pos, name):
77        # pylint: disable=too-many-arguments
78        self.filename = filename
79        self.line = line
80        self.line_no = line_no
81        self.pos = pos
82        self.name = name
83
84    def __str__(self):
85        """
86        Return a formatted code listing representation of the erroneous line.
87        """
88        gutter = format(self.line_no, "4d")
89        underline = self.pos[0] * " " + (self.pos[1] - self.pos[0]) * "^"
90
91        return (
92            " {0} |\n".format(" " * len(gutter)) +
93            " {0} | {1}".format(gutter, self.line) +
94            " {0} | {1}\n".format(" " * len(gutter), underline)
95        )
96
97class Problem(abc.ABC): # pylint: disable=too-few-public-methods
98    """
99    An abstract parent class representing a form of static analysis error.
100    It extends an Abstract Base Class, which means it is not instantiable, and
101    it also mandates certain abstract methods to be implemented in subclasses.
102    """
103    # Class variable to control the quietness of all problems
104    quiet = False
105    def __init__(self):
106        self.textwrapper = textwrap.TextWrapper()
107        self.textwrapper.width = 80
108        self.textwrapper.initial_indent = "    > "
109        self.textwrapper.subsequent_indent = "      "
110
111    def __str__(self):
112        """
113        Unified string representation method for all Problems.
114        """
115        if self.__class__.quiet:
116            return self.quiet_output()
117        return self.verbose_output()
118
119    @abc.abstractmethod
120    def quiet_output(self):
121        """
122        The output when --quiet is enabled.
123        """
124        pass
125
126    @abc.abstractmethod
127    def verbose_output(self):
128        """
129        The default output with explanation and code snippet if appropriate.
130        """
131        pass
132
133class SymbolNotInHeader(Problem): # pylint: disable=too-few-public-methods
134    """
135    A problem that occurs when an exported/available symbol in the object file
136    is not explicitly declared in header files. Created with
137    NameCheck.check_symbols_declared_in_header()
138
139    Fields:
140    * symbol_name: the name of the symbol.
141    """
142    def __init__(self, symbol_name):
143        self.symbol_name = symbol_name
144        Problem.__init__(self)
145
146    def quiet_output(self):
147        return "{0}".format(self.symbol_name)
148
149    def verbose_output(self):
150        return self.textwrapper.fill(
151            "'{0}' was found as an available symbol in the output of nm, "
152            "however it was not declared in any header files."
153            .format(self.symbol_name))
154
155class PatternMismatch(Problem): # pylint: disable=too-few-public-methods
156    """
157    A problem that occurs when something doesn't match the expected pattern.
158    Created with NameCheck.check_match_pattern()
159
160    Fields:
161    * pattern: the expected regex pattern
162    * match: the Match object in question
163    """
164    def __init__(self, pattern, match):
165        self.pattern = pattern
166        self.match = match
167        Problem.__init__(self)
168
169
170    def quiet_output(self):
171        return (
172            "{0}:{1}:{2}"
173            .format(self.match.filename, self.match.line_no, self.match.name)
174        )
175
176    def verbose_output(self):
177        return self.textwrapper.fill(
178            "{0}:{1}: '{2}' does not match the required pattern '{3}'."
179            .format(
180                self.match.filename,
181                self.match.line_no,
182                self.match.name,
183                self.pattern
184            )
185        ) + "\n" + str(self.match)
186
187class Typo(Problem): # pylint: disable=too-few-public-methods
188    """
189    A problem that occurs when a word using MBED doesn't appear to be defined as
190    constants nor enum values. Created with NameCheck.check_for_typos()
191
192    Fields:
193    * match: the Match object of the MBED name in question.
194    """
195    def __init__(self, match):
196        self.match = match
197        Problem.__init__(self)
198
199    def quiet_output(self):
200        return (
201            "{0}:{1}:{2}"
202            .format(self.match.filename, self.match.line_no, self.match.name)
203        )
204
205    def verbose_output(self):
206        return self.textwrapper.fill(
207            "{0}:{1}: '{2}' looks like a typo. It was not found in any "
208            "macros or any enums. If this is not a typo, put "
209            "//no-check-names after it."
210            .format(self.match.filename, self.match.line_no, self.match.name)
211        ) + "\n" + str(self.match)
212
213class CodeParser():
214    """
215    Class for retrieving files and parsing the code. This can be used
216    independently of the checks that NameChecker performs, for example for
217    list_internal_identifiers.py.
218    """
219    def __init__(self, log):
220        self.log = log
221        self.check_repo_path()
222
223        # Memo for storing "glob expression": set(filepaths)
224        self.files = {}
225
226        # Globally excluded filenames.
227        # Note that "*" can match directory separators in exclude lists.
228        self.excluded_files = ["*/bn_mul", "*/compat-2.x.h"]
229
230    @staticmethod
231    def check_repo_path():
232        """
233        Check that the current working directory is the project root, and throw
234        an exception if not.
235        """
236        if not all(os.path.isdir(d) for d in ["include", "library", "tests"]):
237            raise Exception("This script must be run from Mbed TLS root")
238
239    def comprehensive_parse(self):
240        """
241        Comprehensive ("default") function to call each parsing function and
242        retrieve various elements of the code, together with the source location.
243
244        Returns a dict of parsed item key to the corresponding List of Matches.
245        """
246        self.log.info("Parsing source code...")
247        self.log.debug(
248            "The following files are excluded from the search: {}"
249            .format(str(self.excluded_files))
250        )
251
252        all_macros = self.parse_macros([
253            "include/mbedtls/*.h",
254            "include/psa/*.h",
255            "library/*.h",
256            "tests/include/test/drivers/*.h",
257            "3rdparty/everest/include/everest/everest.h",
258            "3rdparty/everest/include/everest/x25519.h"
259        ])
260        enum_consts = self.parse_enum_consts([
261            "include/mbedtls/*.h",
262            "library/*.h",
263            "3rdparty/everest/include/everest/everest.h",
264            "3rdparty/everest/include/everest/x25519.h"
265        ])
266        identifiers = self.parse_identifiers([
267            "include/mbedtls/*.h",
268            "include/psa/*.h",
269            "library/*.h",
270            "3rdparty/everest/include/everest/everest.h",
271            "3rdparty/everest/include/everest/x25519.h"
272        ])
273        mbed_words = self.parse_mbed_words([
274            "include/mbedtls/*.h",
275            "include/psa/*.h",
276            "library/*.h",
277            "3rdparty/everest/include/everest/everest.h",
278            "3rdparty/everest/include/everest/x25519.h",
279            "library/*.c",
280            "3rdparty/everest/library/everest.c",
281            "3rdparty/everest/library/x25519.c"
282        ])
283        symbols = self.parse_symbols()
284
285        # Remove identifier macros like mbedtls_printf or mbedtls_calloc
286        identifiers_justname = [x.name for x in identifiers]
287        actual_macros = []
288        for macro in all_macros:
289            if macro.name not in identifiers_justname:
290                actual_macros.append(macro)
291
292        self.log.debug("Found:")
293        # Aligns the counts on the assumption that none exceeds 4 digits
294        self.log.debug("  {:4} Total Macros".format(len(all_macros)))
295        self.log.debug("  {:4} Non-identifier Macros".format(len(actual_macros)))
296        self.log.debug("  {:4} Enum Constants".format(len(enum_consts)))
297        self.log.debug("  {:4} Identifiers".format(len(identifiers)))
298        self.log.debug("  {:4} Exported Symbols".format(len(symbols)))
299        return {
300            "macros": actual_macros,
301            "enum_consts": enum_consts,
302            "identifiers": identifiers,
303            "symbols": symbols,
304            "mbed_words": mbed_words
305        }
306
307    def is_file_excluded(self, path, exclude_wildcards):
308        """Whether the given file path is excluded."""
309        # exclude_wildcards may be None. Also, consider the global exclusions.
310        exclude_wildcards = (exclude_wildcards or []) + self.excluded_files
311        for pattern in exclude_wildcards:
312            if fnmatch.fnmatch(path, pattern):
313                return True
314        return False
315
316    def get_files(self, include_wildcards, exclude_wildcards):
317        """
318        Get all files that match any of the UNIX-style wildcards. While the
319        check_names script is designed only for use on UNIX/macOS (due to nm),
320        this function alone would work fine on Windows even with forward slashes
321        in the wildcard.
322
323        Args:
324        * include_wildcards: a List of shell-style wildcards to match filepaths.
325        * exclude_wildcards: a List of shell-style wildcards to exclude.
326
327        Returns a List of relative filepaths.
328        """
329        accumulator = set()
330
331        for include_wildcard in include_wildcards:
332            accumulator = accumulator.union(glob.iglob(include_wildcard))
333
334        return list(path for path in accumulator
335                    if not self.is_file_excluded(path, exclude_wildcards))
336
337    def parse_macros(self, include, exclude=None):
338        """
339        Parse all macros defined by #define preprocessor directives.
340
341        Args:
342        * include: A List of glob expressions to look for files through.
343        * exclude: A List of glob expressions for excluding files.
344
345        Returns a List of Match objects for the found macros.
346        """
347        macro_regex = re.compile(r"# *define +(?P<macro>\w+)")
348        exclusions = (
349            "asm", "inline", "EMIT", "_CRT_SECURE_NO_DEPRECATE", "MULADDC_"
350        )
351
352        files = self.get_files(include, exclude)
353        self.log.debug("Looking for macros in {} files".format(len(files)))
354
355        macros = []
356        for header_file in files:
357            with open(header_file, "r", encoding="utf-8") as header:
358                for line_no, line in enumerate(header):
359                    for macro in macro_regex.finditer(line):
360                        if macro.group("macro").startswith(exclusions):
361                            continue
362
363                        macros.append(Match(
364                            header_file,
365                            line,
366                            line_no,
367                            macro.span("macro"),
368                            macro.group("macro")))
369
370        return macros
371
372    def parse_mbed_words(self, include, exclude=None):
373        """
374        Parse all words in the file that begin with MBED, in and out of macros,
375        comments, anything.
376
377        Args:
378        * include: A List of glob expressions to look for files through.
379        * exclude: A List of glob expressions for excluding files.
380
381        Returns a List of Match objects for words beginning with MBED.
382        """
383        # Typos of TLS are common, hence the broader check below than MBEDTLS.
384        mbed_regex = re.compile(r"\bMBED.+?_[A-Z0-9_]*")
385        exclusions = re.compile(r"// *no-check-names|#error")
386
387        files = self.get_files(include, exclude)
388        self.log.debug("Looking for MBED words in {} files".format(len(files)))
389
390        mbed_words = []
391        for filename in files:
392            with open(filename, "r", encoding="utf-8") as fp:
393                for line_no, line in enumerate(fp):
394                    if exclusions.search(line):
395                        continue
396
397                    for name in mbed_regex.finditer(line):
398                        mbed_words.append(Match(
399                            filename,
400                            line,
401                            line_no,
402                            name.span(0),
403                            name.group(0)))
404
405        return mbed_words
406
407    def parse_enum_consts(self, include, exclude=None):
408        """
409        Parse all enum value constants that are declared.
410
411        Args:
412        * include: A List of glob expressions to look for files through.
413        * exclude: A List of glob expressions for excluding files.
414
415        Returns a List of Match objects for the findings.
416        """
417        files = self.get_files(include, exclude)
418        self.log.debug("Looking for enum consts in {} files".format(len(files)))
419
420        # Emulate a finite state machine to parse enum declarations.
421        # OUTSIDE_KEYWORD = outside the enum keyword
422        # IN_BRACES = inside enum opening braces
423        # IN_BETWEEN = between enum keyword and opening braces
424        states = enum.Enum("FSM", ["OUTSIDE_KEYWORD", "IN_BRACES", "IN_BETWEEN"])
425        enum_consts = []
426        for header_file in files:
427            state = states.OUTSIDE_KEYWORD
428            with open(header_file, "r", encoding="utf-8") as header:
429                for line_no, line in enumerate(header):
430                    # Match typedefs and brackets only when they are at the
431                    # beginning of the line -- if they are indented, they might
432                    # be sub-structures within structs, etc.
433                    if (state == states.OUTSIDE_KEYWORD and
434                            re.search(r"^(typedef +)?enum +{", line)):
435                        state = states.IN_BRACES
436                    elif (state == states.OUTSIDE_KEYWORD and
437                          re.search(r"^(typedef +)?enum", line)):
438                        state = states.IN_BETWEEN
439                    elif (state == states.IN_BETWEEN and
440                          re.search(r"^{", line)):
441                        state = states.IN_BRACES
442                    elif (state == states.IN_BRACES and
443                          re.search(r"^}", line)):
444                        state = states.OUTSIDE_KEYWORD
445                    elif (state == states.IN_BRACES and
446                          not re.search(r"^ *#", line)):
447                        enum_const = re.search(r"^ *(?P<enum_const>\w+)", line)
448                        if not enum_const:
449                            continue
450
451                        enum_consts.append(Match(
452                            header_file,
453                            line,
454                            line_no,
455                            enum_const.span("enum_const"),
456                            enum_const.group("enum_const")))
457
458        return enum_consts
459
460    IGNORED_CHUNK_REGEX = re.compile('|'.join([
461        r'/\*.*?\*/', # block comment entirely on one line
462        r'//.*', # line comment
463        r'(?P<string>")(?:[^\\\"]|\\.)*"', # string literal
464    ]))
465
466    def strip_comments_and_literals(self, line, in_block_comment):
467        """Strip comments and string literals from line.
468
469        Continuation lines are not supported.
470
471        If in_block_comment is true, assume that the line starts inside a
472        block comment.
473
474        Return updated values of (line, in_block_comment) where:
475        * Comments in line have been replaced by a space (or nothing at the
476          start or end of the line).
477        * String contents have been removed.
478        * in_block_comment indicates whether the line ends inside a block
479          comment that continues on the next line.
480        """
481
482        # Terminate current multiline comment?
483        if in_block_comment:
484            m = re.search(r"\*/", line)
485            if m:
486                in_block_comment = False
487                line = line[m.end(0):]
488            else:
489                return '', True
490
491        # Remove full comments and string literals.
492        # Do it all together to handle cases like "/*" correctly.
493        # Note that continuation lines are not supported.
494        line = re.sub(self.IGNORED_CHUNK_REGEX,
495                      lambda s: '""' if s.group('string') else ' ',
496                      line)
497
498        # Start an unfinished comment?
499        # (If `/*` was part of a complete comment, it's already been removed.)
500        m = re.search(r"/\*", line)
501        if m:
502            in_block_comment = True
503            line = line[:m.start(0)]
504
505        return line, in_block_comment
506
507    IDENTIFIER_REGEX = re.compile('|'.join([
508        # Match " something(a" or " *something(a". Functions.
509        # Assumptions:
510        # - function definition from return type to one of its arguments is
511        #   all on one line
512        # - function definition line only contains alphanumeric, asterisk,
513        #   underscore, and open bracket
514        r".* \**(\w+) *\( *\w",
515        # Match "(*something)(".
516        r".*\( *\* *(\w+) *\) *\(",
517        # Match names of named data structures.
518        r"(?:typedef +)?(?:struct|union|enum) +(\w+)(?: *{)?$",
519        # Match names of typedef instances, after closing bracket.
520        r"}? *(\w+)[;[].*",
521    ]))
522    # The regex below is indented for clarity.
523    EXCLUSION_LINES = re.compile("|".join([
524        r"extern +\"C\"",
525        r"(typedef +)?(struct|union|enum)( *{)?$",
526        r"} *;?$",
527        r"$",
528        r"//",
529        r"#",
530    ]))
531
532    def parse_identifiers_in_file(self, header_file, identifiers):
533        """
534        Parse all lines of a header where a function/enum/struct/union/typedef
535        identifier is declared, based on some regex and heuristics. Highly
536        dependent on formatting style.
537
538        Append found matches to the list ``identifiers``.
539        """
540
541        with open(header_file, "r", encoding="utf-8") as header:
542            in_block_comment = False
543            # The previous line variable is used for concatenating lines
544            # when identifiers are formatted and spread across multiple
545            # lines.
546            previous_line = ""
547
548            for line_no, line in enumerate(header):
549                line, in_block_comment = \
550                    self.strip_comments_and_literals(line, in_block_comment)
551
552                if self.EXCLUSION_LINES.match(line):
553                    previous_line = ""
554                    continue
555
556                # If the line contains only space-separated alphanumeric
557                # characters (or underscore, asterisk, or open parenthesis),
558                # and nothing else, high chance it's a declaration that
559                # continues on the next line
560                if re.search(r"^([\w\*\(]+\s+)+$", line):
561                    previous_line += line
562                    continue
563
564                # If previous line seemed to start an unfinished declaration
565                # (as above), concat and treat them as one.
566                if previous_line:
567                    line = previous_line.strip() + " " + line.strip() + "\n"
568                    previous_line = ""
569
570                # Skip parsing if line has a space in front = heuristic to
571                # skip function argument lines (highly subject to formatting
572                # changes)
573                if line[0] == " ":
574                    continue
575
576                identifier = self.IDENTIFIER_REGEX.search(line)
577
578                if not identifier:
579                    continue
580
581                # Find the group that matched, and append it
582                for group in identifier.groups():
583                    if not group:
584                        continue
585
586                    identifiers.append(Match(
587                        header_file,
588                        line,
589                        line_no,
590                        identifier.span(),
591                        group))
592
593    def parse_identifiers(self, include, exclude=None):
594        """
595        Parse all lines of a header where a function/enum/struct/union/typedef
596        identifier is declared, based on some regex and heuristics. Highly
597        dependent on formatting style.
598
599        Args:
600        * include: A List of glob expressions to look for files through.
601        * exclude: A List of glob expressions for excluding files.
602
603        Returns a List of Match objects with identifiers.
604        """
605
606        files = self.get_files(include, exclude)
607        self.log.debug("Looking for identifiers in {} files".format(len(files)))
608
609        identifiers = []
610        for header_file in files:
611            self.parse_identifiers_in_file(header_file, identifiers)
612
613        return identifiers
614
615    def parse_symbols(self):
616        """
617        Compile the Mbed TLS libraries, and parse the TLS, Crypto, and x509
618        object files using nm to retrieve the list of referenced symbols.
619        Exceptions thrown here are rethrown because they would be critical
620        errors that void several tests, and thus needs to halt the program. This
621        is explicitly done for clarity.
622
623        Returns a List of unique symbols defined and used in the libraries.
624        """
625        self.log.info("Compiling...")
626        symbols = []
627
628        # Back up the config and atomically compile with the full configratuion.
629        shutil.copy(
630            "include/mbedtls/mbedtls_config.h",
631            "include/mbedtls/mbedtls_config.h.bak"
632        )
633        try:
634            # Use check=True in all subprocess calls so that failures are raised
635            # as exceptions and logged.
636            subprocess.run(
637                ["python3", "scripts/config.py", "full"],
638                universal_newlines=True,
639                check=True
640            )
641            my_environment = os.environ.copy()
642            my_environment["CFLAGS"] = "-fno-asynchronous-unwind-tables"
643            # Run make clean separately to lib to prevent unwanted behavior when
644            # make is invoked with parallelism.
645            subprocess.run(
646                ["make", "clean"],
647                universal_newlines=True,
648                check=True
649            )
650            subprocess.run(
651                ["make", "lib"],
652                env=my_environment,
653                universal_newlines=True,
654                stdout=subprocess.PIPE,
655                stderr=subprocess.STDOUT,
656                check=True
657            )
658
659            # Perform object file analysis using nm
660            symbols = self.parse_symbols_from_nm([
661                "library/libmbedcrypto.a",
662                "library/libmbedtls.a",
663                "library/libmbedx509.a"
664            ])
665
666            subprocess.run(
667                ["make", "clean"],
668                universal_newlines=True,
669                check=True
670            )
671        except subprocess.CalledProcessError as error:
672            self.log.debug(error.output)
673            raise error
674        finally:
675            # Put back the original config regardless of there being errors.
676            # Works also for keyboard interrupts.
677            shutil.move(
678                "include/mbedtls/mbedtls_config.h.bak",
679                "include/mbedtls/mbedtls_config.h"
680            )
681
682        return symbols
683
684    def parse_symbols_from_nm(self, object_files):
685        """
686        Run nm to retrieve the list of referenced symbols in each object file.
687        Does not return the position data since it is of no use.
688
689        Args:
690        * object_files: a List of compiled object filepaths to search through.
691
692        Returns a List of unique symbols defined and used in any of the object
693        files.
694        """
695        nm_undefined_regex = re.compile(r"^\S+: +U |^$|^\S+:$")
696        nm_valid_regex = re.compile(r"^\S+( [0-9A-Fa-f]+)* . _*(?P<symbol>\w+)")
697        exclusions = ("FStar", "Hacl")
698
699        symbols = []
700
701        # Gather all outputs of nm
702        nm_output = ""
703        for lib in object_files:
704            nm_output += subprocess.run(
705                ["nm", "-og", lib],
706                universal_newlines=True,
707                stdout=subprocess.PIPE,
708                stderr=subprocess.STDOUT,
709                check=True
710            ).stdout
711
712        for line in nm_output.splitlines():
713            if not nm_undefined_regex.search(line):
714                symbol = nm_valid_regex.search(line)
715                if (symbol and not symbol.group("symbol").startswith(exclusions)):
716                    symbols.append(symbol.group("symbol"))
717                else:
718                    self.log.error(line)
719
720        return symbols
721
722class NameChecker():
723    """
724    Representation of the core name checking operation performed by this script.
725    """
726    def __init__(self, parse_result, log):
727        self.parse_result = parse_result
728        self.log = log
729
730    def perform_checks(self, quiet=False):
731        """
732        A comprehensive checker that performs each check in order, and outputs
733        a final verdict.
734
735        Args:
736        * quiet: whether to hide detailed problem explanation.
737        """
738        self.log.info("=============")
739        Problem.quiet = quiet
740        problems = 0
741        problems += self.check_symbols_declared_in_header()
742
743        pattern_checks = [
744            ("macros", MACRO_PATTERN),
745            ("enum_consts", CONSTANTS_PATTERN),
746            ("identifiers", IDENTIFIER_PATTERN)
747        ]
748        for group, check_pattern in pattern_checks:
749            problems += self.check_match_pattern(group, check_pattern)
750
751        problems += self.check_for_typos()
752
753        self.log.info("=============")
754        if problems > 0:
755            self.log.info("FAIL: {0} problem(s) to fix".format(str(problems)))
756            if quiet:
757                self.log.info("Remove --quiet to see explanations.")
758            else:
759                self.log.info("Use --quiet for minimal output.")
760            return 1
761        else:
762            self.log.info("PASS")
763            return 0
764
765    def check_symbols_declared_in_header(self):
766        """
767        Perform a check that all detected symbols in the library object files
768        are properly declared in headers.
769        Assumes parse_names_in_source() was called before this.
770
771        Returns the number of problems that need fixing.
772        """
773        problems = []
774
775        for symbol in self.parse_result["symbols"]:
776            found_symbol_declared = False
777            for identifier_match in self.parse_result["identifiers"]:
778                if symbol == identifier_match.name:
779                    found_symbol_declared = True
780                    break
781
782            if not found_symbol_declared:
783                problems.append(SymbolNotInHeader(symbol))
784
785        self.output_check_result("All symbols in header", problems)
786        return len(problems)
787
788    def check_match_pattern(self, group_to_check, check_pattern):
789        """
790        Perform a check that all items of a group conform to a regex pattern.
791        Assumes parse_names_in_source() was called before this.
792
793        Args:
794        * group_to_check: string key to index into self.parse_result.
795        * check_pattern: the regex to check against.
796
797        Returns the number of problems that need fixing.
798        """
799        problems = []
800
801        for item_match in self.parse_result[group_to_check]:
802            if not re.search(check_pattern, item_match.name):
803                problems.append(PatternMismatch(check_pattern, item_match))
804            # Double underscore should not be used for names
805            if re.search(r".*__.*", item_match.name):
806                problems.append(
807                    PatternMismatch("no double underscore allowed", item_match))
808
809        self.output_check_result(
810            "Naming patterns of {}".format(group_to_check),
811            problems)
812        return len(problems)
813
814    def check_for_typos(self):
815        """
816        Perform a check that all words in the soure code beginning with MBED are
817        either defined as macros, or as enum constants.
818        Assumes parse_names_in_source() was called before this.
819
820        Returns the number of problems that need fixing.
821        """
822        problems = []
823
824        # Set comprehension, equivalent to a list comprehension wrapped by set()
825        all_caps_names = {
826            match.name
827            for match
828            in self.parse_result["macros"] + self.parse_result["enum_consts"]}
829        typo_exclusion = re.compile(r"XXX|__|_$|^MBEDTLS_.*CONFIG_FILE$|"
830                                    r"MBEDTLS_TEST_LIBTESTDRIVER*")
831
832        for name_match in self.parse_result["mbed_words"]:
833            found = name_match.name in all_caps_names
834
835            # Since MBEDTLS_PSA_ACCEL_XXX defines are defined by the
836            # PSA driver, they will not exist as macros. However, they
837            # should still be checked for typos using the equivalent
838            # BUILTINs that exist.
839            if "MBEDTLS_PSA_ACCEL_" in name_match.name:
840                found = name_match.name.replace(
841                    "MBEDTLS_PSA_ACCEL_",
842                    "MBEDTLS_PSA_BUILTIN_") in all_caps_names
843
844            if not found and not typo_exclusion.search(name_match.name):
845                problems.append(Typo(name_match))
846
847        self.output_check_result("Likely typos", problems)
848        return len(problems)
849
850    def output_check_result(self, name, problems):
851        """
852        Write out the PASS/FAIL status of a performed check depending on whether
853        there were problems.
854
855        Args:
856        * name: the name of the test
857        * problems: a List of encountered Problems
858        """
859        if problems:
860            self.log.info("{}: FAIL\n".format(name))
861            for problem in problems:
862                self.log.warning(str(problem))
863        else:
864            self.log.info("{}: PASS".format(name))
865
866def main():
867    """
868    Perform argument parsing, and create an instance of CodeParser and
869    NameChecker to begin the core operation.
870    """
871    parser = argparse.ArgumentParser(
872        formatter_class=argparse.RawDescriptionHelpFormatter,
873        description=(
874            "This script confirms that the naming of all symbols and identifiers "
875            "in Mbed TLS are consistent with the house style and are also "
876            "self-consistent.\n\n"
877            "Expected to be run from the MbedTLS root directory.")
878    )
879    parser.add_argument(
880        "-v", "--verbose",
881        action="store_true",
882        help="show parse results"
883    )
884    parser.add_argument(
885        "-q", "--quiet",
886        action="store_true",
887        help="hide unnecessary text, explanations, and highlighs"
888    )
889
890    args = parser.parse_args()
891
892    # Configure the global logger, which is then passed to the classes below
893    log = logging.getLogger()
894    log.setLevel(logging.DEBUG if args.verbose else logging.INFO)
895    log.addHandler(logging.StreamHandler())
896
897    try:
898        code_parser = CodeParser(log)
899        parse_result = code_parser.comprehensive_parse()
900    except Exception: # pylint: disable=broad-except
901        traceback.print_exc()
902        sys.exit(2)
903
904    name_checker = NameChecker(parse_result, log)
905    return_code = name_checker.perform_checks(quiet=args.quiet)
906
907    sys.exit(return_code)
908
909if __name__ == "__main__":
910    main()
911