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		# Check if exists, evaluating wildcards
79		next if (grep -e, glob("$ref $fulref"));
80
81		if ($fix) {
82			if (!($ref =~ m/(scripts|Kconfig|Kbuild)/)) {
83				$broken_ref{$ref}++;
84			}
85		} else {
86			print STDERR "$f: $fulref\n";
87		}
88	}
89}
90
91exit 0 if (!$fix);
92
93# Step 2: Seek for file name alternatives
94print "Auto-fixing broken references. Please double-check the results\n";
95
96foreach my $ref (keys %broken_ref) {
97	my $new =$ref;
98
99	# get just the basename
100	$new =~ s,.*/,,;
101
102	my $f="";
103
104	# usual reason for breakage: DT file moved around
105	if ($ref =~ /devicetree/) {
106		my $search = $new;
107		$search =~ s,^.*/,,;
108		$f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
109		if (!$f) {
110			# Manufacturer name may have changed
111			$search =~ s/^.*,//;
112			$f = qx(find Documentation/devicetree/ -iname "*$search*") if ($search);
113		}
114	}
115
116	# usual reason for breakage: file renamed to .rst
117	if (!$f) {
118		$new =~ s/\.txt$/.rst/;
119		$f=qx(find . -iname $new) if ($new);
120	}
121
122	# usual reason for breakage: use dash or underline
123	if (!$f) {
124		$new =~ s/[-_]/[-_]/g;
125		$f=qx(find . -iname $new) if ($new);
126	}
127
128	# Wild guess: seek for the same name on another place
129	if (!$f) {
130		$f = qx(find . -iname $new) if ($new);
131	}
132
133	my @find = split /\s+/, $f;
134
135	if (!$f) {
136		print STDERR "ERROR: Didn't find a replacement for $ref\n";
137	} elsif (scalar(@find) > 1) {
138		print STDERR "WARNING: Won't auto-replace, as found multiple files close to $ref:\n";
139		foreach my $j (@find) {
140			$j =~ s,^./,,;
141			print STDERR "    $j\n";
142		}
143	} else {
144		$f = $find[0];
145		$f =~ s,^./,,;
146		print "INFO: Replacing $ref to $f\n";
147		foreach my $j (qx(git grep -l $ref)) {
148			qx(sed "s\@$ref\@$f\@g" -i $j);
149		}
150	}
151}
152