check-module

   1 #!/usr/bin/perl -w
   2 # Check a gnulib module.
   3
   4 # Copyright (C) 2005-2007, 2009-2021 Free Software Foundation, Inc.
   5
   6 # This file is free software: you can redistribute it and/or modify it
   7 # under the terms of the GNU General Public License as published by
   8 # the Free Software Foundation; either version 3 of the License, or
   9 # (at your option) any later version.
  10 #
  11 # This program is distributed in the hope that it will be useful,
  12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 # GNU General Public License for more details.
  15 #
  16 # You should have received a copy of the GNU General Public License
  17 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
  18
  19
  20 # Read a module description file and derive the set of files
  21 # included directly by any .c or .h file listed in the 'Files:' section.
  22 # Take the union of all such sets for any dependent modules.
  23 # Then, compare that set with the set derived from the names
  24 # listed in the various Files: sections.
  25
  26 # This script makes no attempt to diagnose invalid or empty
  27 # module-description files.
  28
  29 # Written by Jim Meyering
  30
  31 # FIXME:
  32 # for each .m4 file listed in the Files: section(s)
  33 # parse it for AC_LIBSOURCES directives, and accumulate the set
  34 # of files "required" via all AC_LIBSOURCES.
  35 # If this set is not empty, ensure that it contains
  36 # the same (.c and .h only?) files as are listed in the Files: sections.
  37
  38 use strict;
  39 use Getopt::Long;
  40 use File::Basename;
  41 #use Coda;
  42
  43 my $COPYRIGHT_NOTICE = "Copyright (C) 2012 Free Software Foundation, Inc.\n".
  44 "This is free software.  You may redistribute copies of it under the terms of\n".
  45 "the GNU General Public License <https://www.gnu.org/licenses/gpl.html>.\n".
  46 "There is NO WARRANTY, to the extent permitted by law.\n";
  47
  48 (my $VERSION = '$Revision: 1.8 $ ') =~ tr/[0-9].//cd;
  49 (my $ME = $0) =~ s|.*/||;
  50
  51 use constant ST_INIT => 1;
  52 use constant ST_FILES => 2;
  53 use constant ST_DEPENDENTS => 3;
  54
  55 # Parse a module file (returning list of Files: names and
  56 # list of dependent-modules.
  57 # my ($file, $dep) = parse_module_file $module_file;
  58 sub parse_module_file ($)
  59 {
  60   my ($module_file) = @_;
  61
  62   open FH, '<', $module_file
  63     or die "$ME: can't open '$module_file' for reading: $!\n";
  64
  65   my %file_set;
  66   my %dep_set;
  67
  68   my $state = ST_INIT;
  69   while (defined (my $line = <FH>))
  70     {
  71       if ($state eq ST_INIT)
  72         {
  73           if ($line =~ /^Files:$/)
  74             {
  75               $state = ST_FILES;
  76             }
  77           elsif ($line =~ /^Depends-on:$/)
  78             {
  79               $state = ST_DEPENDENTS;
  80             }
  81         }
  82       else
  83         {
  84           chomp $line;
  85           $line =~ s/^\s+//;
  86           $line =~ s/\s+$//;
  87           if ( ! $line)
  88             {
  89               $state = ST_INIT;
  90               next;
  91             }
  92
  93           if ($state eq ST_FILES)
  94             {
  95               $file_set{$line} = 1;
  96             }
  97           elsif ($state eq ST_DEPENDENTS)
  98             {
  99               $dep_set{$line} = 1;
 100               (my $base = $module_file) =~ s,.*/,,;
 101               $line eq $base
 102                 and die "$ME: module $module_file depends on itself\n";
 103             }
 104         }
 105     }
 106   close FH;
 107
 108   # my @t = sort keys %file_set;
 109   # print "files: @t\n";
 110   # my @u = sort keys %dep_set;
 111   # print "dependents: @u\n";
 112
 113   return (\%file_set, \%dep_set);
 114 }
 115
 116 # Extract the set of files required for this module, including
 117 # those required via dependent modules.
 118
 119 # Files:
 120 # lib/stat.c
 121 # m4/stat.m4
 122 # lib/foo.h
 123 #
 124 # Depends-on:
 125 # some-other-module
 126
 127 sub usage ($)
 128 {
 129   my ($exit_code) = @_;
 130   my $STREAM = ($exit_code == 0 ? *STDOUT : *STDERR);
 131   if ($exit_code != 0)
 132     {
 133       print $STREAM "Try '$ME --help' for more information.\n";
 134     }
 135   else
 136     {
 137       print $STREAM <<EOF;
 138 Usage: $ME [OPTIONS] FILE...
 139
 140 Read a module description file and derive the set of files
 141 included directly by any .c or .h file listed in the 'Files:' section.
 142 Take the union of all such sets for any dependent modules.
 143 Then, compare that set with the set derived from the names
 144 listed in the various Files: sections.
 145
 146 OPTIONS:
 147
 148    --help             display this help and exit
 149    --version          output version information and exit
 150
 151 EOF
 152     }
 153   exit $exit_code;
 154 }
 155
 156 sub find_included_lib_files ($)
 157 {
 158   my ($file) = @_;
 159
 160   # Special cases...
 161   my %special_non_dup = ( 'fnmatch_loop.c' => 1,
 162                           'regex.c' => 1, 'at-func.c' => 1,
 163                           'vasnprintf.c' => 1
 164                         );
 165   my %dup_include_ok;
 166   $dup_include_ok{'vasnprintf.c'}{'isnand-nolibm.h'} = 1;
 167   $dup_include_ok{'vasnprintf.c'}{'isnanl-nolibm.h'} = 1;
 168   $dup_include_ok{'vasnprintf.c'}{'fpucw.h'} = 1;
 169   $dup_include_ok{'gen-uni-tables.c'}{'3level.h'} = 1;
 170   $dup_include_ok{'csharpexec.c'}{'classpath.c'} = 1;
 171   $dup_include_ok{'csharpexec.c'}{'classpath.h'} = 1;
 172
 173   my %inc;
 174   open FH, '<', $file
 175     or die "$ME: can't open '$file' for reading: $!\n";
 176
 177   while (defined (my $line = <FH>))
 178     {
 179       # Ignore test-driver code at end of file.
 180       $line =~ m!^\#if(def)? TEST_!
 181         and last;
 182
 183       $line =~ m!^\s*\#\s*include\s+"!
 184         or next;
 185       $line =~ s///;
 186       chomp $line;
 187       $line =~ s/".*//;
 188       exists $inc{$line} && ! exists $special_non_dup{$line}
 189           && ! exists $dup_include_ok{basename $file}{$line}
 190         and warn "$ME: $file: duplicate inclusion of $line\n";
 191
 192       $inc{$line} = 1;
 193     }
 194   close FH;
 195
 196   return \%inc;
 197 }
 198
 199 my %exempt_header =
 200   (
 201    # Exempt headers like unlocked-io.h that are '#include'd
 202    # but not necessarily used.
 203    'unlocked-io.h' => 1,
 204
 205    # Give gettext.h a free pass only when included from lib/error.c,
 206    # since we've made that exception solely to make the error
 207    # module easier to use -- at RMS's request.
 208    'lib/error.c:gettext.h' => 1,
 209
 210    # The full-read module shares code with the full-write module.
 211    'lib/full-write.c:full-read.h' => 1,
 212
 213    # The safe-write module shares code with the safe-read module.
 214    'lib/safe-read.c:safe-write.h' => 1,
 215
 216    # The use of obstack.h in the hash module is conditional, off by default.
 217    'lib/hash.c:obstack.h' => 1,
 218
 219    # C files in the gc module have conditional includes.
 220    'lib/gc-gnulib.c:des.h' => 1,
 221    'lib/gc-gnulib.c:arcfour.h' => 1,
 222    'lib/gc-gnulib.c:arctwo.h' => 1,
 223    'lib/gc-gnulib.c:md2.h' => 1,
 224    'lib/gc-gnulib.c:md4.h' => 1,
 225    'lib/gc-gnulib.c:md5.h' => 1,
 226    'lib/gc-gnulib.c:rijndael.h' => 1,
 227    'lib/gc-gnulib.c:sha1.h' => 1,
 228    'lib/gc-gnulib.c:rijndael-api-fst.h' => 1,
 229    'lib/gc-gnulib.c:hmac.h' => 1,
 230    'lib/gc-libgcrypt.c:md2.h' => 1,
 231   );
 232
 233 sub check_module ($)
 234 {
 235   my @m = @_;
 236
 237   my %file;
 238   my %module_all_files;
 239   my %dep;
 240   my %seen_module;
 241
 242   while (@m)
 243     {
 244       my $m = pop @m;
 245       # warn "M: $m\n";
 246       exists $seen_module{$m}
 247         and next;
 248       $seen_module{$m} = 1;
 249       my ($file, $dep) = parse_module_file $m;
 250       push @m, keys %$dep;
 251       foreach my $f (keys %$file)
 252         {
 253           $module_all_files{$f} = 1;
 254         }
 255     }
 256
 257   my @t = sort keys %module_all_files;
 258   # warn "ALL files: @t\n";
 259
 260   # Derive from %module_all_files (by parsing the .c and .h files therein),
 261   # the list of all #include'd files that reside in lib/.
 262   foreach my $f (keys %module_all_files)
 263     {
 264       $f =~ /\.[ch]$/
 265         or next;
 266       # FIXME: this is too naive
 267       my $inc = find_included_lib_files "../$f";
 268       foreach my $i (sort keys %$inc)
 269         {
 270           my $lib_file = "lib/$i";
 271           exists $exempt_header{"$f:$i"}
 272             || exists $exempt_header{$i}
 273               and next;
 274           !exists $module_all_files{$lib_file} && -f "../lib/$i"
 275             and warn "$f: $i is '#include'd, but not "
 276               . "listed in module's Files: section\n";
 277         }
 278       #my @t = sort keys %$inc;
 279       #print "** $f: @t\n";
 280     }
 281 }
 282
 283 {
 284   GetOptions
 285     (
 286      help => sub { usage 0 },
 287      version => sub { print "$ME version $VERSION\n$COPYRIGHT_NOTICE"; exit },
 288     ) or usage 1;
 289
 290   @ARGV < 1
 291     and (warn "$ME: missing FILE argument\n"), usage 1;
 292
 293   foreach my $module (@ARGV)
 294     {
 295       check_module $module;
 296     }
 297
 298   exit 0;
 299 }