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
11my $scriptname = $0;
12$scriptname =~ s,.*/([^/]+/),$1,;
13
14# Parse arguments
15my $help = 0;
16my $fix = 0;
17
18GetOptions(
19	'fix' => \$fix,
20	'h|help|usage' => \$help,
21);
22
23if ($help != 0) {
24    print "$scriptname [--help] [--fix]\n";
25    exit -1;
26}
27
28# Step 1: find broken references
29print "Finding broken references. This may take a while...  " if ($fix);
30
31my %broken_ref;
32
33open IN, "git grep 'Documentation/'|"
34     or die "Failed to run git grep";
35while (<IN>) {
36	next if (!m/^([^:]+):(.*)/);
37
38	my $f = $1;
39	my $ln = $2;
40
41	# Makefiles and scripts contain nasty expressions to parse docs
42	next if ($f =~ m/Makefile/ || $f =~ m/\.sh$/);
43
44	# Skip this script
45	next if ($f eq $scriptname);
46
47	if ($ln =~ m,\b(\S*)(Documentation/[A-Za-z0-9\_\.\,\~/\*\[\]\?+-]*)(.*),) {
48		my $prefix = $1;
49		my $ref = $2;
50		my $base = $2;
51		my $extra = $3;
52
53		# some file references are like:
54		# /usr/src/linux/Documentation/DMA-{API,mapping}.txt
55		# For now, ignore them
56		next if ($extra =~ m/^{/);
57
58		# Remove footnotes at the end like:
59		# Documentation/devicetree/dt-object-internal.txt[1]
60		$ref =~ s/(txt|rst)\[\d+]$/$1/;
61
62		# Remove ending ']' without any '['
63		$ref =~ s/\].*// if (!($ref =~ m/\[/));
64
65		# Remove puntuation marks at the end
66		$ref =~ s/[\,\.]+$//;
67
68		my $fulref = "$prefix$ref";
69
70		$fulref =~ s/^(\<file|ref)://;
71		$fulref =~ s/^[\'\`]+//;
72		$fulref =~ s,^\$\(.*\)/,,;
73		$base =~ s,.*/,,;
74
75		# Remove URL false-positives
76		next if ($fulref =~ m/^http/);
77
78		# Remove sched-pelt false-positive
79		next if ($fulref =~ m,^Documentation/scheduler/sched-pelt$,);
80
81		# Discard some build examples from Documentation/target/tcm_mod_builder.txt
82		next if ($fulref =~ m,mnt/sdb/lio-core-2.6.git/Documentation/target,);
83
84		# Check if exists, evaluating wildcards
85		next if (grep -e, glob("$ref $fulref"));
86
87		# Accept relative Documentation patches for tools/
88		if ($f =~ m/tools/) {
89			my $path = $f;
90			$path =~ s,(.*)/.*,$1,;
91			next if (grep -e, glob("$path/$ref $path/$fulref"));
92		}
93
94		if ($fix) {
95			if (!($ref =~ m/(scripts|Kconfig|Kbuild)/)) {
96				$broken_ref{$ref}++;
97			}
98		} else {
99			print STDERR "$f: $fulref\n";
100		}
101	}
102}
103
104exit 0 if (!$fix);
105
106# Step 2: Seek for file name alternatives
107print "Auto-fixing broken references. Please double-check the results\n";
108
109foreach my $ref (keys %broken_ref) {
110	my $new =$ref;
111
112	# get just the basename
113	$new =~ s,.*/,,;
114
115	my $f="";
116
117	# usual reason for breakage: DT file moved around
118	if ($ref =~ /devicetree/) {
119		my $search = $new;
120		$search =~ s,^.*/,,;
121		$f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
122		if (!$f) {
123			# Manufacturer name may have changed
124			$search =~ s/^.*,//;
125			$f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
126		}
127	}
128
129	# usual reason for breakage: file renamed to .rst
130	if (!$f) {
131		$new =~ s/\.txt$/.rst/;
132		$f=qx(find . -iname $new) if ($new);
133	}
134
135	# usual reason for breakage: use dash or underline
136	if (!$f) {
137		$new =~ s/[-_]/[-_]/g;
138		$f=qx(find . -iname $new) if ($new);
139	}
140
141	# Wild guess: seek for the same name on another place
142	if (!$f) {
143		$f = qx(find . -iname $new) if ($new);
144	}
145
146	my @find = split /\s+/, $f;
147
148	if (!$f) {
149		print STDERR "ERROR: Didn't find a replacement for $ref\n";
150	} elsif (scalar(@find) > 1) {
151		print STDERR "WARNING: Won't auto-replace, as found multiple files close to $ref:\n";
152		foreach my $j (@find) {
153			$j =~ s,^./,,;
154			print STDERR "    $j\n";
155		}
156	} else {
157		$f = $find[0];
158		$f =~ s,^./,,;
159		print "INFO: Replacing $ref to $f\n";
160		foreach my $j (qx(git grep -l $ref)) {
161			qx(sed "s\@$ref\@$f\@g" -i $j);
162		}
163	}
164}
165