File: check-obsolete-constructs.py

package info (click to toggle)
glibc 2.29-3
  • links: PTS, VCS
  • area: main
  • in suites: bullseye, sid
  • size: 271,436 kB
  • sloc: ansic: 1,013,441; asm: 261,303; makefile: 11,428; sh: 10,300; python: 8,944; cpp: 5,062; awk: 1,956; perl: 514; yacc: 290; pascal: 182; sed: 73
file content (466 lines) | stat: -rw-r--r-- 16,913 bytes parent folder | download | duplicates (6)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
#! /usr/bin/python3
# Copyright (C) 2019 Free Software Foundation, Inc.
# This file is part of the GNU C Library.
#
# The GNU C Library is free software; you can redistribute it and/or
# modify it under the terms of the GNU Lesser General Public
# License as published by the Free Software Foundation; either
# version 2.1 of the License, or (at your option) any later version.
#
# The GNU C Library is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public
# License along with the GNU C Library; if not, see
# <http://www.gnu.org/licenses/>.

"""Verifies that installed headers do not use any obsolete constructs:
 * legacy BSD typedefs superseded by <stdint.h>:
   ushort uint ulong u_char u_short u_int u_long u_intNN_t quad_t u_quad_t
   (sys/types.h is allowed to _define_ these types, but not to use them
    to define anything else).
"""

import argparse
import collections
import re
import sys

# Simplified lexical analyzer for C preprocessing tokens.
# Does not implement trigraphs.
# Does not implement backslash-newline in the middle of any lexical
#   item other than a string literal.
# Does not implement universal-character-names in identifiers.
# Treats prefixed strings (e.g. L"...") as two tokens (L and "...")
# Accepts non-ASCII characters only within comments and strings.

# Caution: The order of the outermost alternation matters.
# STRING must be before BAD_STRING, CHARCONST before BAD_CHARCONST,
# BLOCK_COMMENT before BAD_BLOCK_COM before PUNCTUATOR, and OTHER must
# be last.
# Caution: There should be no capturing groups other than the named
# captures in the outermost alternation.

# For reference, these are all of the C punctuators as of C11:
#   [ ] ( ) { } , ; ? ~
#   ! != * *= / /= ^ ^= = ==
#   # ##
#   % %= %> %: %:%:
#   & &= &&
#   | |= ||
#   + += ++
#   - -= -- ->
#   . ...
#   : :>
#   < <% <: << <<= <=
#   > >= >> >>=

# The BAD_* tokens are not part of the official definition of pp-tokens;
# they match unclosed strings, character constants, and block comments,
# so that the regex engine doesn't have to backtrack all the way to the
# beginning of a broken construct and then emit dozens of junk tokens.

PP_TOKEN_RE_ = re.compile(r"""
    (?P<STRING>        \"(?:[^\"\\\r\n]|\\(?:[\r\n -~]|\r\n))*\")
   |(?P<BAD_STRING>    \"(?:[^\"\\\r\n]|\\[ -~])*)
   |(?P<CHARCONST>     \'(?:[^\'\\\r\n]|\\(?:[\r\n -~]|\r\n))*\')
   |(?P<BAD_CHARCONST> \'(?:[^\'\\\r\n]|\\[ -~])*)
   |(?P<BLOCK_COMMENT> /\*(?:\*(?!/)|[^*])*\*/)
   |(?P<BAD_BLOCK_COM> /\*(?:\*(?!/)|[^*])*\*?)
   |(?P<LINE_COMMENT>  //[^\r\n]*)
   |(?P<IDENT>         [_a-zA-Z][_a-zA-Z0-9]*)
   |(?P<PP_NUMBER>     \.?[0-9](?:[0-9a-df-oq-zA-DF-OQ-Z_.]|[eEpP][+-]?)*)
   |(?P<PUNCTUATOR>
       [,;?~(){}\[\]]
     | [!*/^=]=?
     | \#\#?
     | %(?:[=>]|:(?:%:)?)?
     | &[=&]?
     |\|[=|]?
     |\+[=+]?
     | -[=->]?
     |\.(?:\.\.)?
     | :>?
     | <(?:[%:]|<(?:=|<=?)?)?
     | >(?:=|>=?)?)
   |(?P<ESCNL>         \\(?:\r|\n|\r\n))
   |(?P<WHITESPACE>    [ \t\n\r\v\f]+)
   |(?P<OTHER>         .)
""", re.DOTALL | re.VERBOSE)

HEADER_NAME_RE_ = re.compile(r"""
    < [^>\r\n]+ >
  | " [^"\r\n]+ "
""", re.DOTALL | re.VERBOSE)

ENDLINE_RE_ = re.compile(r"""\r|\n|\r\n""")

# based on the sample code in the Python re documentation
Token_ = collections.namedtuple("Token", (
    "kind", "text", "line", "column", "context"))
Token_.__doc__ = """
   One C preprocessing token, comment, or chunk of whitespace.
   'kind' identifies the token type, which will be one of:
       STRING, CHARCONST, BLOCK_COMMENT, LINE_COMMENT, IDENT,
       PP_NUMBER, PUNCTUATOR, ESCNL, WHITESPACE, HEADER_NAME,
       or OTHER.  The BAD_* alternatives in PP_TOKEN_RE_ are
       handled within tokenize_c, below.

   'text' is the sequence of source characters making up the token;
       no decoding whatsoever is performed.

   'line' and 'column' give the position of the first character of the
      token within the source file.  They are both 1-based.

   'context' indicates whether or not this token occurred within a
      preprocessing directive; it will be None for running text,
      '<null>' for the leading '#' of a directive line (because '#'
      all by itself on a line is a "null directive"), or the name of
      the directive for tokens within a directive line, starting with
      the IDENT for the name itself.
"""

def tokenize_c(file_contents, reporter):
    """Yield a series of Token objects, one for each preprocessing
       token, comment, or chunk of whitespace within FILE_CONTENTS.
       The REPORTER object is expected to have one method,
       reporter.error(token, message), which will be called to
       indicate a lexical error at the position of TOKEN.
       If MESSAGE contains the four-character sequence '{!r}', that
       is expected to be replaced by repr(token.text).
    """

    Token = Token_
    PP_TOKEN_RE = PP_TOKEN_RE_
    ENDLINE_RE = ENDLINE_RE_
    HEADER_NAME_RE = HEADER_NAME_RE_

    line_num = 1
    line_start = 0
    pos = 0
    limit = len(file_contents)
    directive = None
    at_bol = True
    while pos < limit:
        if directive == "include":
            mo = HEADER_NAME_RE.match(file_contents, pos)
            if mo:
                kind = "HEADER_NAME"
                directive = "after_include"
            else:
                mo = PP_TOKEN_RE.match(file_contents, pos)
                kind = mo.lastgroup
                if kind != "WHITESPACE":
                    directive = "after_include"
        else:
            mo = PP_TOKEN_RE.match(file_contents, pos)
            kind = mo.lastgroup

        text = mo.group()
        line = line_num
        column = mo.start() - line_start
        adj_line_start = 0
        # only these kinds can contain a newline
        if kind in ("WHITESPACE", "BLOCK_COMMENT", "LINE_COMMENT",
                    "STRING", "CHARCONST", "BAD_BLOCK_COM", "ESCNL"):
            for tmo in ENDLINE_RE.finditer(text):
                line_num += 1
                adj_line_start = tmo.end()
            if adj_line_start:
                line_start = mo.start() + adj_line_start

        # Track whether or not we are scanning a preprocessing directive.
        if kind == "LINE_COMMENT" or (kind == "WHITESPACE" and adj_line_start):
            at_bol = True
            directive = None
        else:
            if kind == "PUNCTUATOR" and text == "#" and at_bol:
                directive = "<null>"
            elif kind == "IDENT" and directive == "<null>":
                directive = text
            at_bol = False

        # Report ill-formed tokens and rewrite them as their well-formed
        # equivalents, so downstream processing doesn't have to know about them.
        # (Rewriting instead of discarding provides better error recovery.)
        if kind == "BAD_BLOCK_COM":
            reporter.error(Token("BAD_BLOCK_COM", "", line, column+1, ""),
                           "unclosed block comment")
            text += "*/"
            kind = "BLOCK_COMMENT"
        elif kind == "BAD_STRING":
            reporter.error(Token("BAD_STRING", "", line, column+1, ""),
                           "unclosed string")
            text += "\""
            kind = "STRING"
        elif kind == "BAD_CHARCONST":
            reporter.error(Token("BAD_CHARCONST", "", line, column+1, ""),
                           "unclosed char constant")
            text += "'"
            kind = "CHARCONST"

        tok = Token(kind, text, line, column+1,
                    "include" if directive == "after_include" else directive)
        # Do not complain about OTHER tokens inside macro definitions.
        # $ and @ appear in macros defined by headers intended to be
        # included from assembly language, e.g. sysdeps/mips/sys/asm.h.
        if kind == "OTHER" and directive != "define":
            self.error(tok, "stray {!r} in program")

        yield tok
        pos = mo.end()

#
# Base and generic classes for individual checks.
#

class ConstructChecker:
    """Scan a stream of C preprocessing tokens and possibly report
       problems with them.  The REPORTER object passed to __init__ has
       one method, reporter.error(token, message), which should be
       called to indicate a problem detected at the position of TOKEN.
       If MESSAGE contains the four-character sequence '{!r}' then that
       will be replaced with a textual representation of TOKEN.
    """
    def __init__(self, reporter):
        self.reporter = reporter

    def examine(self, tok):
        """Called once for each token in a header file.
           Call self.reporter.error if a problem is detected.
        """
        raise NotImplementedError

    def eof(self):
        """Called once at the end of the stream.  Subclasses need only
           override this if it might have something to do."""
        pass

class NoCheck(ConstructChecker):
    """Generic checker class which doesn't do anything.  Substitute this
       class for a real checker when a particular check should be skipped
       for some file."""

    def examine(self, tok):
        pass

#
# Check for obsolete type names.
#

# The obsolete type names we're looking for:
OBSOLETE_TYPE_RE_ = re.compile(r"""\A
  (__)?
  (   quad_t
    | u(?: short | int | long
         | _(?: char | short | int(?:[0-9]+_t)? | long | quad_t )))
\Z""", re.VERBOSE)

class ObsoleteNotAllowed(ConstructChecker):
    """Don't allow any use of the obsolete typedefs."""
    def examine(self, tok):
        if OBSOLETE_TYPE_RE_.match(tok.text):
            self.reporter.error(tok, "use of {!r}")

class ObsoletePrivateDefinitionsAllowed(ConstructChecker):
    """Allow definitions of the private versions of the
       obsolete typedefs; that is, 'typedef [anything] __obsolete;'
    """
    def __init__(self, reporter):
        super().__init__(reporter)
        self.in_typedef = False
        self.prev_token = None

    def examine(self, tok):
        # bits/types.h hides 'typedef' in a macro sometimes.
        if (tok.kind == "IDENT"
            and tok.text in ("typedef", "__STD_TYPE")
            and tok.context is None):
            self.in_typedef = True
        elif tok.kind == "PUNCTUATOR" and tok.text == ";" and self.in_typedef:
            self.in_typedef = False
            if self.prev_token.kind == "IDENT":
                m = OBSOLETE_TYPE_RE_.match(self.prev_token.text)
                if m and m.group(1) != "__":
                    self.reporter.error(self.prev_token, "use of {!r}")
            self.prev_token = None
        else:
            self._check_prev()

        self.prev_token = tok

    def eof(self):
        self._check_prev()

    def _check_prev(self):
        if (self.prev_token is not None
            and self.prev_token.kind == "IDENT"
            and OBSOLETE_TYPE_RE_.match(self.prev_token.text)):
            self.reporter.error(self.prev_token, "use of {!r}")

class ObsoletePublicDefinitionsAllowed(ConstructChecker):
    """Allow definitions of the public versions of the obsolete
       typedefs.  Only specific forms of definition are allowed:

           typedef __obsolete obsolete;  // identifiers must agree
           typedef __uintN_t u_intN_t;   // N must agree
           typedef unsigned long int ulong;
           typedef unsigned short int ushort;
           typedef unsigned int uint;
    """
    def __init__(self, reporter):
        super().__init__(reporter)
        self.typedef_tokens = []

    def examine(self, tok):
        if tok.kind in ("WHITESPACE", "BLOCK_COMMENT",
                        "LINE_COMMENT", "NL", "ESCNL"):
            pass

        elif (tok.kind == "IDENT" and tok.text == "typedef"
              and tok.context is None):
            if self.typedef_tokens:
                self.reporter.error(tok, "typedef inside typedef")
                self._reset()
            self.typedef_tokens.append(tok)

        elif tok.kind == "PUNCTUATOR" and tok.text == ";":
            self._finish()

        elif self.typedef_tokens:
            self.typedef_tokens.append(tok)

    def eof(self):
        self._reset()

    def _reset(self):
        while self.typedef_tokens:
            tok = self.typedef_tokens.pop(0)
            if tok.kind == "IDENT" and OBSOLETE_TYPE_RE_.match(tok.text):
                self.reporter.error(tok, "use of {!r}")

    def _finish(self):
        if not self.typedef_tokens: return
        if self.typedef_tokens[-1].kind == "IDENT":
            m = OBSOLETE_TYPE_RE_.match(self.typedef_tokens[-1].text)
            if m:
                if self._permissible_public_definition(m):
                    self.typedef_tokens.clear()
        self._reset()

    def _permissible_public_definition(self, m):
        if m.group(1) == "__": return False
        name = m.group(2)
        toks = self.typedef_tokens
        ntok = len(toks)
        if ntok == 3 and toks[1].kind == "IDENT":
            defn = toks[1].text
            n = OBSOLETE_TYPE_RE_.match(defn)
            if n and n.group(1) == "__" and n.group(2) == name:
                return True

            if (name[:5] == "u_int" and name[-2:] == "_t"
                and defn[:6] == "__uint" and defn[-2:] == "_t"
                and name[5:-2] == defn[6:-2]):
                return True

            return False

        if (name == "ulong" and ntok == 5
            and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
            and toks[2].kind == "IDENT" and toks[2].text == "long"
            and toks[3].kind == "IDENT" and toks[3].text == "int"):
            return True

        if (name == "ushort" and ntok == 5
            and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
            and toks[2].kind == "IDENT" and toks[2].text == "short"
            and toks[3].kind == "IDENT" and toks[3].text == "int"):
            return True

        if (name == "uint" and ntok == 4
            and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
            and toks[2].kind == "IDENT" and toks[2].text == "int"):
            return True

        return False

def ObsoleteTypedefChecker(reporter, fname):
    """Factory: produce an instance of the appropriate
       obsolete-typedef checker for FNAME."""

    # The obsolete rpc/ and rpcsvc/ headers are allowed to use the
    # obsolete types, because it would be more trouble than it's
    # worth to remove them from headers that we intend to stop
    # installing eventually anyway.
    if (fname.startswith("rpc/")
        or fname.startswith("rpcsvc/")
        or "/rpc/" in fname
        or "/rpcsvc/" in fname):
        return NoCheck(reporter)

    # bits/types.h is allowed to define the __-versions of the
    # obsolete types.
    if (fname == "bits/types.h"
        or fname.endswith("/bits/types.h")):
        return ObsoletePrivateDefinitionsAllowed(reporter)

    # sys/types.h is allowed to use the __-versions of the
    # obsolete types, but only to define the unprefixed versions.
    if (fname == "sys/types.h"
        or fname.endswith("/sys/types.h")):
        return ObsoletePublicDefinitionsAllowed(reporter)

    return ObsoleteNotAllowed(reporter)

#
# Master control
#

class HeaderChecker:
    """Perform all of the checks on each header.  This is also the
       "reporter" object expected by tokenize_c and ConstructChecker.
    """
    def __init__(self):
        self.fname = None
        self.status = 0

    def error(self, tok, message):
        self.status = 1
        if '{!r}' in message:
            message = message.format(tok.text)
        sys.stderr.write("{}:{}:{}: error: {}\n".format(
            self.fname, tok.line, tok.column, message))

    def check(self, fname):
        self.fname = fname
        try:
            with open(fname, "rt") as fp:
                contents = fp.read()
        except OSError as e:
            sys.stderr.write("{}: {}\n".format(fname, e.strerror))
            self.status = 1
            return

        typedef_checker = ObsoleteTypedefChecker(self, self.fname)

        for tok in tokenize_c(contents, self):
            typedef_checker.examine(tok)

def main():
    ap = argparse.ArgumentParser(description=__doc__)
    ap.add_argument("headers", metavar="header", nargs="+",
                    help="one or more headers to scan for obsolete constructs")
    args = ap.parse_args()

    checker = HeaderChecker()
    for fname in args.headers:
        # Headers whose installed name begins with "finclude/" contain
        # Fortran, not C, and this program should completely ignore them.
        if not (fname.startswith("finclude/") or "/finclude/" in fname):
            checker.check(fname)
    sys.exit(checker.status)

main()