scripts/check-obsolete-constructs.py

   1 #! /usr/bin/python3
   2 # Copyright (C) 2019-2024 Free Software Foundation, Inc.
   3 # This file is part of the GNU C Library.
   4 #
   5 # The GNU C Library is free software; you can redistribute it and/or
   6 # modify it under the terms of the GNU Lesser General Public
   7 # License as published by the Free Software Foundation; either
   8 # version 2.1 of the License, or (at your option) any later version.
   9 #
  10 # The GNU C Library is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13 # Lesser General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU Lesser General Public
  16 # License along with the GNU C Library; if not, see
  17 # <https://www.gnu.org/licenses/>.
  18
  19 """Verifies that installed headers do not use any obsolete constructs:
  20  * legacy BSD typedefs superseded by <stdint.h>:
  21    ushort uint ulong u_char u_short u_int u_long u_intNN_t quad_t u_quad_t
  22    (sys/types.h is allowed to _define_ these types, but not to use them
  23     to define anything else).
  24 """
  25
  26 import argparse
  27 import os
  28 import re
  29 import sys
  30
  31 # Make available glibc Python modules.
  32 sys.path.append(os.path.dirname(os.path.realpath(__file__)))
  33
  34 import glibcpp
  35
  36 #
  37 # Base and generic classes for individual checks.
  38 #
  39
  40 class ConstructChecker:
  41     """Scan a stream of C preprocessing tokens and possibly report
  42        problems with them.  The REPORTER object passed to __init__ has
  43        one method, reporter.error(token, message), which should be
  44        called to indicate a problem detected at the position of TOKEN.
  45        If MESSAGE contains the four-character sequence '{!r}' then that
  46        will be replaced with a textual representation of TOKEN.
  47     """
  48     def __init__(self, reporter):
  49         self.reporter = reporter
  50
  51     def examine(self, tok):
  52         """Called once for each token in a header file.
  53            Call self.reporter.error if a problem is detected.
  54         """
  55         raise NotImplementedError
  56
  57     def eof(self):
  58         """Called once at the end of the stream.  Subclasses need only
  59            override this if it might have something to do."""
  60         pass
  61
  62 class NoCheck(ConstructChecker):
  63     """Generic checker class which doesn't do anything.  Substitute this
  64        class for a real checker when a particular check should be skipped
  65        for some file."""
  66
  67     def examine(self, tok):
  68         pass
  69
  70 #
  71 # Check for obsolete type names.
  72 #
  73
  74 # The obsolete type names we're looking for:
  75 OBSOLETE_TYPE_RE_ = re.compile(r"""\A
  76   (__)?
  77   (   quad_t
  78     | u(?: short | int | long
  79          | _(?: char | short | int(?:[0-9]+_t)? | long | quad_t )))
  80 \Z""", re.VERBOSE)
  81
  82 class ObsoleteNotAllowed(ConstructChecker):
  83     """Don't allow any use of the obsolete typedefs."""
  84     def examine(self, tok):
  85         if OBSOLETE_TYPE_RE_.match(tok.text):
  86             self.reporter.error(tok, "use of {!r}")
  87
  88 class ObsoletePrivateDefinitionsAllowed(ConstructChecker):
  89     """Allow definitions of the private versions of the
  90        obsolete typedefs; that is, 'typedef [anything] __obsolete;'
  91     """
  92     def __init__(self, reporter):
  93         super().__init__(reporter)
  94         self.in_typedef = False
  95         self.prev_token = None
  96
  97     def examine(self, tok):
  98         # bits/types.h hides 'typedef' in a macro sometimes.
  99         if (tok.kind == "IDENT"
 100             and tok.text in ("typedef", "__STD_TYPE")
 101             and tok.context is None):
 102             self.in_typedef = True
 103         elif tok.kind == "PUNCTUATOR" and tok.text == ";" and self.in_typedef:
 104             self.in_typedef = False
 105             if self.prev_token.kind == "IDENT":
 106                 m = OBSOLETE_TYPE_RE_.match(self.prev_token.text)
 107                 if m and m.group(1) != "__":
 108                     self.reporter.error(self.prev_token, "use of {!r}")
 109             self.prev_token = None
 110         else:
 111             self._check_prev()
 112
 113         self.prev_token = tok
 114
 115     def eof(self):
 116         self._check_prev()
 117
 118     def _check_prev(self):
 119         if (self.prev_token is not None
 120             and self.prev_token.kind == "IDENT"
 121             and OBSOLETE_TYPE_RE_.match(self.prev_token.text)):
 122             self.reporter.error(self.prev_token, "use of {!r}")
 123
 124 class ObsoletePublicDefinitionsAllowed(ConstructChecker):
 125     """Allow definitions of the public versions of the obsolete
 126        typedefs.  Only specific forms of definition are allowed:
 127
 128            typedef __obsolete obsolete;  // identifiers must agree
 129            typedef __uintN_t u_intN_t;   // N must agree
 130            typedef unsigned long int ulong;
 131            typedef unsigned short int ushort;
 132            typedef unsigned int uint;
 133     """
 134     def __init__(self, reporter):
 135         super().__init__(reporter)
 136         self.typedef_tokens = []
 137
 138     def examine(self, tok):
 139         if tok.kind in ("WHITESPACE", "BLOCK_COMMENT",
 140                         "LINE_COMMENT", "NL", "ESCNL"):
 141             pass
 142
 143         elif (tok.kind == "IDENT" and tok.text == "typedef"
 144               and tok.context is None):
 145             if self.typedef_tokens:
 146                 self.reporter.error(tok, "typedef inside typedef")
 147                 self._reset()
 148             self.typedef_tokens.append(tok)
 149
 150         elif tok.kind == "PUNCTUATOR" and tok.text == ";":
 151             self._finish()
 152
 153         elif self.typedef_tokens:
 154             self.typedef_tokens.append(tok)
 155
 156     def eof(self):
 157         self._reset()
 158
 159     def _reset(self):
 160         while self.typedef_tokens:
 161             tok = self.typedef_tokens.pop(0)
 162             if tok.kind == "IDENT" and OBSOLETE_TYPE_RE_.match(tok.text):
 163                 self.reporter.error(tok, "use of {!r}")
 164
 165     def _finish(self):
 166         if not self.typedef_tokens: return
 167         if self.typedef_tokens[-1].kind == "IDENT":
 168             m = OBSOLETE_TYPE_RE_.match(self.typedef_tokens[-1].text)
 169             if m:
 170                 if self._permissible_public_definition(m):
 171                     self.typedef_tokens.clear()
 172         self._reset()
 173
 174     def _permissible_public_definition(self, m):
 175         if m.group(1) == "__": return False
 176         name = m.group(2)
 177         toks = self.typedef_tokens
 178         ntok = len(toks)
 179         if ntok == 3 and toks[1].kind == "IDENT":
 180             defn = toks[1].text
 181             n = OBSOLETE_TYPE_RE_.match(defn)
 182             if n and n.group(1) == "__" and n.group(2) == name:
 183                 return True
 184
 185             if (name[:5] == "u_int" and name[-2:] == "_t"
 186                 and defn[:6] == "__uint" and defn[-2:] == "_t"
 187                 and name[5:-2] == defn[6:-2]):
 188                 return True
 189
 190             return False
 191
 192         if (name == "ulong" and ntok == 5
 193             and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
 194             and toks[2].kind == "IDENT" and toks[2].text == "long"
 195             and toks[3].kind == "IDENT" and toks[3].text == "int"):
 196             return True
 197
 198         if (name == "ushort" and ntok == 5
 199             and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
 200             and toks[2].kind == "IDENT" and toks[2].text == "short"
 201             and toks[3].kind == "IDENT" and toks[3].text == "int"):
 202             return True
 203
 204         if (name == "uint" and ntok == 4
 205             and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
 206             and toks[2].kind == "IDENT" and toks[2].text == "int"):
 207             return True
 208
 209         return False
 210
 211 def ObsoleteTypedefChecker(reporter, fname):
 212     """Factory: produce an instance of the appropriate
 213        obsolete-typedef checker for FNAME."""
 214
 215     # The obsolete rpc/ and rpcsvc/ headers are allowed to use the
 216     # obsolete types, because it would be more trouble than it's
 217     # worth to remove them from headers that we intend to stop
 218     # installing eventually anyway.
 219     if (fname.startswith("rpc/")
 220         or fname.startswith("rpcsvc/")
 221         or "/rpc/" in fname
 222         or "/rpcsvc/" in fname):
 223         return NoCheck(reporter)
 224
 225     # bits/types.h is allowed to define the __-versions of the
 226     # obsolete types.
 227     if (fname == "bits/types.h"
 228         or fname.endswith("/bits/types.h")):
 229         return ObsoletePrivateDefinitionsAllowed(reporter)
 230
 231     # sys/types.h is allowed to use the __-versions of the
 232     # obsolete types, but only to define the unprefixed versions.
 233     if (fname == "sys/types.h"
 234         or fname.endswith("/sys/types.h")):
 235         return ObsoletePublicDefinitionsAllowed(reporter)
 236
 237     return ObsoleteNotAllowed(reporter)
 238
 239 #
 240 # Master control
 241 #
 242
 243 class HeaderChecker:
 244     """Perform all of the checks on each header.  This is also the
 245        "reporter" object expected by tokenize_c and ConstructChecker.
 246     """
 247     def __init__(self):
 248         self.fname = None
 249         self.status = 0
 250
 251     def error(self, tok, message):
 252         self.status = 1
 253         if '{!r}' in message:
 254             message = message.format(tok.text)
 255         sys.stderr.write("{}:{}:{}: error: {}\n".format(
 256             self.fname, tok.line, tok.column, message))
 257
 258     def check(self, fname):
 259         self.fname = fname
 260         try:
 261             with open(fname, "rt", encoding="utf-8") as fp:
 262                 contents = fp.read()
 263         except OSError as e:
 264             sys.stderr.write("{}: {}\n".format(fname, e.strerror))
 265             self.status = 1
 266             return
 267
 268         typedef_checker = ObsoleteTypedefChecker(self, self.fname)
 269
 270         for tok in glibcpp.tokenize_c(contents, self):
 271             typedef_checker.examine(tok)
 272
 273 def main():
 274     ap = argparse.ArgumentParser(description=__doc__)
 275     ap.add_argument("headers", metavar="header", nargs="+",
 276                     help="one or more headers to scan for obsolete constructs")
 277     args = ap.parse_args()
 278
 279     checker = HeaderChecker()
 280     for fname in args.headers:
 281         # Headers whose installed name begins with "finclude/" contain
 282         # Fortran, not C, and this program should completely ignore them.
 283         if not (fname.startswith("finclude/") or "/finclude/" in fname):
 284             checker.check(fname)
 285     sys.exit(checker.status)
 286
 287 main()