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 7use warnings; 8use strict; 9use 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. 14my %false_positives = ( 15 "Documentation/scsi/scsi_mid_low_api.rst" => "Documentation/Configure.help", 16 "drivers/vhost/vhost.c" => "Documentation/virtual/lguest/lguest.c", 17); 18 19my $scriptname = $0; 20$scriptname =~ s,.*/([^/]+/),$1,; 21 22# Parse arguments 23my $help = 0; 24my $fix = 0; 25my $warn = 0; 26 27if (! -e ".git") { 28 printf "Warning: can't check if file exists, as this is not a git tree\n"; 29 exit 0; 30} 31 32GetOptions( 33 'fix' => \$fix, 34 'warn' => \$warn, 35 'h|help|usage' => \$help, 36); 37 38if ($help != 0) { 39 print "$scriptname [--help] [--fix]\n"; 40 exit -1; 41} 42 43# Step 1: find broken references 44print "Finding broken references. This may take a while... " if ($fix); 45 46my %broken_ref; 47 48my $doc_fix = 0; 49 50open IN, "git grep ':doc:\`' Documentation/|" 51 or die "Failed to run git grep"; 52while (<IN>) { 53 next if (!m,^([^:]+):.*\:doc\:\`([^\`]+)\`,); 54 next if (m,sphinx/,); 55 56 my $file = $1; 57 my $d = $1; 58 my $doc_ref = $2; 59 60 my $f = $doc_ref; 61 62 $d =~ s,(.*/).*,$1,; 63 $f =~ s,.*\<([^\>]+)\>,$1,; 64 65 if ($f =~ m,^/,) { 66 $f = "$f.rst"; 67 $f =~ s,^/,Documentation/,; 68 } else { 69 $f = "$d$f.rst"; 70 } 71 72 next if (grep -e, glob("$f")); 73 74 if ($fix && !$doc_fix) { 75 print STDERR "\nWARNING: Currently, can't fix broken :doc:`` fields\n"; 76 } 77 $doc_fix++; 78 79 print STDERR "$file: :doc:`$doc_ref`\n"; 80} 81close IN; 82 83open IN, "git grep 'Documentation/'|" 84 or die "Failed to run git grep"; 85while (<IN>) { 86 next if (!m/^([^:]+):(.*)/); 87 88 my $f = $1; 89 my $ln = $2; 90 91 # On linux-next, discard the Next/ directory 92 next if ($f =~ m,^Next/,); 93 94 # Makefiles and scripts contain nasty expressions to parse docs 95 next if ($f =~ m/Makefile/ || $f =~ m/\.sh$/); 96 97 # Skip this script 98 next if ($f eq $scriptname); 99 100 # Ignore the dir where documentation will be built 101 next if ($ln =~ m,\b(\S*)Documentation/output,); 102 103 if ($ln =~ m,\b(\S*)(Documentation/[A-Za-z0-9\_\.\,\~/\*\[\]\?+-]*)(.*),) { 104 my $prefix = $1; 105 my $ref = $2; 106 my $base = $2; 107 my $extra = $3; 108 109 # some file references are like: 110 # /usr/src/linux/Documentation/DMA-{API,mapping}.txt 111 # For now, ignore them 112 next if ($extra =~ m/^{/); 113 114 # Remove footnotes at the end like: 115 # Documentation/devicetree/dt-object-internal.txt[1] 116 $ref =~ s/(txt|rst)\[\d+]$/$1/; 117 118 # Remove ending ']' without any '[' 119 $ref =~ s/\].*// if (!($ref =~ m/\[/)); 120 121 # Remove puntuation marks at the end 122 $ref =~ s/[\,\.]+$//; 123 124 my $fulref = "$prefix$ref"; 125 126 $fulref =~ s/^(\<file|ref)://; 127 $fulref =~ s/^[\'\`]+//; 128 $fulref =~ s,^\$\(.*\)/,,; 129 $base =~ s,.*/,,; 130 131 # Remove URL false-positives 132 next if ($fulref =~ m/^http/); 133 134 # Remove sched-pelt false-positive 135 next if ($fulref =~ m,^Documentation/scheduler/sched-pelt$,); 136 137 # Discard some build examples from Documentation/target/tcm_mod_builder.rst 138 next if ($fulref =~ m,mnt/sdb/lio-core-2.6.git/Documentation/target,); 139 140 # Check if exists, evaluating wildcards 141 next if (grep -e, glob("$ref $fulref")); 142 143 # Accept relative Documentation patches for tools/ 144 if ($f =~ m/tools/) { 145 my $path = $f; 146 $path =~ s,(.*)/.*,$1,; 147 next if (grep -e, glob("$path/$ref $path/../$ref $path/$fulref")); 148 } 149 150 # Discard known false-positives 151 if (defined($false_positives{$f})) { 152 next if ($false_positives{$f} eq $fulref); 153 } 154 155 if ($fix) { 156 if (!($ref =~ m/(scripts|Kconfig|Kbuild)/)) { 157 $broken_ref{$ref}++; 158 } 159 } elsif ($warn) { 160 print STDERR "Warning: $f references a file that doesn't exist: $fulref\n"; 161 } else { 162 print STDERR "$f: $fulref\n"; 163 } 164 } 165} 166close IN; 167 168exit 0 if (!$fix); 169 170# Step 2: Seek for file name alternatives 171print "Auto-fixing broken references. Please double-check the results\n"; 172 173foreach my $ref (keys %broken_ref) { 174 my $new =$ref; 175 176 my $basedir = "."; 177 # On translations, only seek inside the translations directory 178 $basedir = $1 if ($ref =~ m,(Documentation/translations/[^/]+),); 179 180 # get just the basename 181 $new =~ s,.*/,,; 182 183 my $f=""; 184 185 # usual reason for breakage: DT file moved around 186 if ($ref =~ /devicetree/) { 187 # usual reason for breakage: DT file renamed to .yaml 188 if (!$f) { 189 my $new_ref = $ref; 190 $new_ref =~ s/\.txt$/.yaml/; 191 $f=$new_ref if (-f $new_ref); 192 } 193 194 if (!$f) { 195 my $search = $new; 196 $search =~ s,^.*/,,; 197 $f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search); 198 if (!$f) { 199 # Manufacturer name may have changed 200 $search =~ s/^.*,//; 201 $f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search); 202 } 203 } 204 } 205 206 # usual reason for breakage: file renamed to .rst 207 if (!$f) { 208 $new =~ s/\.txt$/.rst/; 209 $f=qx(find $basedir -iname $new) if ($new); 210 } 211 212 # usual reason for breakage: use dash or underline 213 if (!$f) { 214 $new =~ s/[-_]/[-_]/g; 215 $f=qx(find $basedir -iname $new) if ($new); 216 } 217 218 # Wild guess: seek for the same name on another place 219 if (!$f) { 220 $f = qx(find $basedir -iname $new) if ($new); 221 } 222 223 my @find = split /\s+/, $f; 224 225 if (!$f) { 226 print STDERR "ERROR: Didn't find a replacement for $ref\n"; 227 } elsif (scalar(@find) > 1) { 228 print STDERR "WARNING: Won't auto-replace, as found multiple files close to $ref:\n"; 229 foreach my $j (@find) { 230 $j =~ s,^./,,; 231 print STDERR " $j\n"; 232 } 233 } else { 234 $f = $find[0]; 235 $f =~ s,^./,,; 236 print "INFO: Replacing $ref to $f\n"; 237 foreach my $j (qx(git grep -l $ref)) { 238 qx(sed "s\@$ref\@$f\@g" -i $j); 239 } 240 } 241} 242