scripts/documentation-file-ref-check

   1 #!/usr/bin/env perl
   2 # SPDX-License-Identifier: GPL-2.0
   3 #
   4 # Treewide grep for references to files under Documentation, and report
   5 # non-existing files in stderr.
   6
   7 use warnings;
   8 use strict;
   9 use Getopt::Long qw(:config no_auto_abbrev);
  10
  11 # NOTE: only add things here when the file was gone, but the text wants
  12 # to mention a past documentation file, for example, to give credits for
  13 # the original work.
  14 my %false_positives = (
  15         "Documentation/scsi/scsi_mid_low_api.txt" => "Documentation/Configure.help",
  16         "drivers/vhost/vhost.c" => "Documentation/virtual/lguest/lguest.c",
  17 );
  18
  19 my $scriptname = $0;
  20 $scriptname =~ s,.*/([^/]+/),$1,;
  21
  22 # Parse arguments
  23 my $help = 0;
  24 my $fix = 0;
  25 my $warn = 0;
  26
  27 if (! -d ".git") {
  28         printf "Warning: can't check if file exists, as this is not a git tree";
  29         exit 0;
  30 }
  31
  32 GetOptions(
  33         'fix' => \$fix,
  34         'warn' => \$warn,
  35         'h|help|usage' => \$help,
  36 );
  37
  38 if ($help != 0) {
  39     print "$scriptname [--help] [--fix]\n";
  40     exit -1;
  41 }
  42
  43 # Step 1: find broken references
  44 print "Finding broken references. This may take a while...  " if ($fix);
  45
  46 my %broken_ref;
  47
  48 my $doc_fix = 0;
  49
  50 open IN, "git grep ':doc:\`' Documentation/|"
  51      or die "Failed to run git grep";
  52 while (<IN>) {
  53         next if (!m,^([^:]+):.*\:doc\:\`([^\`]+)\`,);
  54
  55         my $d = $1;
  56         my $doc_ref = $2;
  57
  58         my $f = $doc_ref;
  59
  60         $d =~ s,(.*/).*,$1,;
  61         $f =~ s,.*\<([^\>]+)\>,$1,;
  62
  63         $f ="$d$f.rst";
  64
  65         next if (grep -e, glob("$f"));
  66
  67         if ($fix && !$doc_fix) {
  68                 print STDERR "\nWARNING: Currently, can't fix broken :doc:`` fields\n";
  69         }
  70         $doc_fix++;
  71
  72         print STDERR "$f: :doc:`$doc_ref`\n";
  73 }
  74 close IN;
  75
  76 open IN, "git grep 'Documentation/'|"
  77      or die "Failed to run git grep";
  78 while (<IN>) {
  79         next if (!m/^([^:]+):(.*)/);
  80
  81         my $f = $1;
  82         my $ln = $2;
  83
  84         # On linux-next, discard the Next/ directory
  85         next if ($f =~ m,^Next/,);
  86
  87         # Makefiles and scripts contain nasty expressions to parse docs
  88         next if ($f =~ m/Makefile/ || $f =~ m/\.sh$/);
  89
  90         # Skip this script
  91         next if ($f eq $scriptname);
  92
  93         # Ignore the dir where documentation will be built
  94         next if ($ln =~ m,\b(\S*)Documentation/output,);
  95
  96         if ($ln =~ m,\b(\S*)(Documentation/[A-Za-z0-9\_\.\,\~/\*\[\]\?+-]*)(.*),) {
  97                 my $prefix = $1;
  98                 my $ref = $2;
  99                 my $base = $2;
 100                 my $extra = $3;
 101
 102                 # some file references are like:
 103                 # /usr/src/linux/Documentation/DMA-{API,mapping}.txt
 104                 # For now, ignore them
 105                 next if ($extra =~ m/^{/);
 106
 107                 # Remove footnotes at the end like:
 108                 # Documentation/devicetree/dt-object-internal.txt[1]
 109                 $ref =~ s/(txt|rst)\[\d+]$/$1/;
 110
 111                 # Remove ending ']' without any '['
 112                 $ref =~ s/\].*// if (!($ref =~ m/\[/));
 113
 114                 # Remove puntuation marks at the end
 115                 $ref =~ s/[\,\.]+$//;
 116
 117                 my $fulref = "$prefix$ref";
 118
 119                 $fulref =~ s/^(\<file|ref)://;
 120                 $fulref =~ s/^[\'\`]+//;
 121                 $fulref =~ s,^\$\(.*\)/,,;
 122                 $base =~ s,.*/,,;
 123
 124                 # Remove URL false-positives
 125                 next if ($fulref =~ m/^http/);
 126
 127                 # Remove sched-pelt false-positive
 128                 next if ($fulref =~ m,^Documentation/scheduler/sched-pelt$,);
 129
 130                 # Discard some build examples from Documentation/target/tcm_mod_builder.rst
 131                 next if ($fulref =~ m,mnt/sdb/lio-core-2.6.git/Documentation/target,);
 132
 133                 # Check if exists, evaluating wildcards
 134                 next if (grep -e, glob("$ref $fulref"));
 135
 136                 # Accept relative Documentation patches for tools/
 137                 if ($f =~ m/tools/) {
 138                         my $path = $f;
 139                         $path =~ s,(.*)/.*,$1,;
 140                         next if (grep -e, glob("$path/$ref $path/../$ref $path/$fulref"));
 141                 }
 142
 143                 # Discard known false-positives
 144                 if (defined($false_positives{$f})) {
 145                         next if ($false_positives{$f} eq $fulref);
 146                 }
 147
 148                 if ($fix) {
 149                         if (!($ref =~ m/(scripts|Kconfig|Kbuild)/)) {
 150                                 $broken_ref{$ref}++;
 151                         }
 152                 } elsif ($warn) {
 153                         print STDERR "Warning: $f references a file that doesn't exist: $fulref\n";
 154                 } else {
 155                         print STDERR "$f: $fulref\n";
 156                 }
 157         }
 158 }
 159 close IN;
 160
 161 exit 0 if (!$fix);
 162
 163 # Step 2: Seek for file name alternatives
 164 print "Auto-fixing broken references. Please double-check the results\n";
 165
 166 foreach my $ref (keys %broken_ref) {
 167         my $new =$ref;
 168
 169         my $basedir = ".";
 170         # On translations, only seek inside the translations directory
 171         $basedir  = $1 if ($ref =~ m,(Documentation/translations/[^/]+),);
 172
 173         # get just the basename
 174         $new =~ s,.*/,,;
 175
 176         my $f="";
 177
 178         # usual reason for breakage: DT file moved around
 179         if ($ref =~ /devicetree/) {
 180                 # usual reason for breakage: DT file renamed to .yaml
 181                 if (!$f) {
 182                         my $new_ref = $ref;
 183                         $new_ref =~ s/\.txt$/.yaml/;
 184                         $f=$new_ref if (-f $new_ref);
 185                 }
 186
 187                 if (!$f) {
 188                         my $search = $new;
 189                         $search =~ s,^.*/,,;
 190                         $f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
 191                         if (!$f) {
 192                                 # Manufacturer name may have changed
 193                                 $search =~ s/^.*,//;
 194                                 $f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
 195                         }
 196                 }
 197         }
 198
 199         # usual reason for breakage: file renamed to .rst
 200         if (!$f) {
 201                 $new =~ s/\.txt$/.rst/;
 202                 $f=qx(find $basedir -iname $new) if ($new);
 203         }
 204
 205         # usual reason for breakage: use dash or underline
 206         if (!$f) {
 207                 $new =~ s/[-_]/[-_]/g;
 208                 $f=qx(find $basedir -iname $new) if ($new);
 209         }
 210
 211         # Wild guess: seek for the same name on another place
 212         if (!$f) {
 213                 $f = qx(find $basedir -iname $new) if ($new);
 214         }
 215
 216         my @find = split /\s+/, $f;
 217
 218         if (!$f) {
 219                 print STDERR "ERROR: Didn't find a replacement for $ref\n";
 220         } elsif (scalar(@find) > 1) {
 221                 print STDERR "WARNING: Won't auto-replace, as found multiple files close to $ref:\n";
 222                 foreach my $j (@find) {
 223                         $j =~ s,^./,,;
 224                         print STDERR "    $j\n";
 225                 }
 226         } else {
 227                 $f = $find[0];
 228                 $f =~ s,^./,,;
 229                 print "INFO: Replacing $ref to $f\n";
 230                 foreach my $j (qx(git grep -l $ref)) {
 231                         qx(sed "s\@$ref\@$f\@g" -i $j);
 232                 }
 233         }
 234 }