~inkscape.dev/inkscape/RELEASE_0_91_BRANCH

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
#! /usr/bin/perl -w
# Try to detect markup errors in translations.

# Author: Peter Moulder <pmoulder@mail.csse.monash.edu.au>
# Copyright (C) 2004 Monash University
# License: GNU GPL v2 or (at your option) any later version.

# Initial egrep version:
#mydir=`dirname "$0"`
#egrep '<b>[^<>]*(>|<([^/]|/([^b"]|b[^>])))' "$mydir"/*.po
# Somewhat simplified by use of negative lookahead in perl.
# (The egrep version as written can't detect problems that span a line,
# e.g. unterminated `<b>'.  One way of doing the s/"\n"//g thing would be with
# tr and sed, but that requires a sed that allows arbitrary line lengths, which
# many non-GNU seds don't.)

use strict;

my $com = qr/(?:\#[^\n]*\n)/;
my $str = qr/(?:"(?:[^"\\]|\\.)*")/;
my $attrsRE = qr/(?: +[^<>]*)?/;
my $span_attr = qr/(?:\ +(?:font_(?:desc|family)|face|size|style|weight|variant|stretch|(?:fore|back)ground|underline|rise|strikethrough|fallback|lang)\=\\\"[^\\\"]*\\\")/;

my $rc = 0;

sub po_error ($) {
    my ($msg) = @_;
    my $name = $ARGV;
    $name =~ s,.*/,,;
    print "$name: $msg:\n$_";
    $rc = 1;
}

# Returns true iff successful.
sub check_str ($$) {
    my ($certainly_pango_str, $str) = @_;

    $str =~ s/\A"// or die "Bug: No leading `\"' in `$str'";
    $str =~ s/"\Z// or die "Bug: No trailing `\"' in `$str'";

    if ($str =~ /\AProject-Id-Version:.*POT-Creation-Date/
        or $str =~ /\A<[^<>]*>\Z/) {
	# Not a Pango string.
	return 1;
    }

    my $is_xml = 0;

    # Remove valid sequences.
    while ($str =~ s{<([bisu]|big|su[bp]|small|tt|span)(${attrsRE})>[^<>]*</\1>}{}) {
	$is_xml = 1;
	my ($tag, $attrs) = ($1, $2);
	if ($tag eq 'span') {
	    $attrs =~ s/${span_attr}*//g;
	    if ($attrs ne '') {
		$attrs =~ s/\A *//;
		$attrs =~ s/\\"/"/g;
		po_error("Unexpected <span> attributes `$attrs'");
		return 0;
	    }
	} else {
	    if ($attrs !~ /\A *\Z/) {
		po_error("<$tag> can't have attributes in Pango");
		return 0;
	    }
	}
    }

    if (($str =~ m{&#[^0-9]+;}) or ($str =~ m{&#x[^0-9a-fA-F]+;})) {
	po_error("Entity declaration error (must look like '&#123;' or '&#x40;' and be in ASCII)");
	return 0;
    }

    if (($str =~ m{&#[^0-9]+}) or ($str =~ m{&#x[^0-9a-fA-F]+})) {
	po_error("Entity declaration error 2 (must look like '&#123;' or '&#x40;' and be in ASCII)");
	return 0;
    }

    if (($str =~ m{&#(?![0-9]{2,4};)}) or ($str =~ m{&#x(?![0-9a-fA-F]{2,4};)})) {
	po_error("Entity declaration error 3 (must look like '&#123;' or '&#x40;' and be in ASCII)");
	return 0;
    }

    # Check for attributes etc. in non-<span> element.
    if ($str =~ m{<([bisu]|big|su[bp]|small|tt)\b(?! *)>}) {
	po_error("Unexpected characters in <$1> tag");
	return 0;
    }

    if ($str =~ m{<([bisu]|big|su[bp]|small|span|tt)${attrsRE}>}) {
	po_error("unclosed <$1>");
	return 0;
    }

    if ($str =~ m{</\ *([bisu]|big|su[bp]|small|span|tt)\ *>}) {
	po_error("Unmatched closing </$1>");
	return 0;
    }

    if (!$is_xml and !$certainly_pango_str) {
	$str =~ s/<(?:defs|image|rect|svg)>//g;
	$str =~ s/<[ 01]//g;
	$str =~ s/\A>+//;
	$str =~ s/<+\Z//;
	$str =~ s/\([<>][01]\)//g;
	$str =~ s/ -> //g;

	# Quoting.
	$str =~ s/\[[<>]\]//g;
	$str =~ s/\\"[<>]\\"//g;
	$str =~ s/\xe2\x80\x9e[<>]\xe2\x80\x9c//g;
	$str =~ s/\xc2\xab[<>]\xc2\xbb//;
    }

    $str =~ s/\A[^<>]*//;
    $str =~ s/[^<>]*\Z//;

    if ($str =~ /\A([<>])\Z/) {
	if ($is_xml or $certainly_pango_str) {
	    po_error("Unescaped `$1'");
	    return 0;
	} else {
	    return 1;
	}
    }

    if ($str ne '') {
	po_error("parsing error for `$str'");
	return 0;
    }

    return 1;
}

sub check_strs ($@) {
    my $is_pango_str = shift(@_);
    if ($#_ < 1) {
	die "check_strs: expecting >= 2 strings";
    }
    if ((($_[0] eq '""') && ($_[1] =~ /Project-Id-Version:.*POT-Creation-Date:/s))
    or ($_[0] eq '"> and < _scale by:"')
    or $is_pango_str == 0)
    {
	# Not a Pango string.
	return 1;
    }
    foreach my $str (@_) {
	$str eq '""' or check_str($is_pango_str - 1, $str) or return 0;
    }
    return 1;
}

$/ = '';

# Reference for the markup language:
# http://developer.gnome.org/doc/API/2.0/pango/PangoMarkupFormat.html
# (though not all translation strings will be pango markup strings).
ENTRY: while(<>) {
	if (m{\A${com}*\Z}) {
	    next ENTRY;
	}

	# Concatenate multi-line string literals.
	s/"\n\s*"//g;

	if (!m{\A${com}*
	       (?:msgctxt[^\n]*\n)?
	       msgid[^\n]*\n
	       ${com}*
	       (?:
	           msgstr[^\n]*\n
		   ${com}*
	         | msgid_plural[^\n]*\n
		   ${com}*
		   (?:msgstr\[[^\n]*\n${com}*)+
	        )\Z}x)
	{
	    po_error('Not in msg format');
	    next ENTRY;
	}
	if (!m{
		\A${com}*
	        (?:msgctxt\ ${str}\s*\n)?
		msgid\ ${str}\s*\n
		${com}*
		(?:
		    msgstr\ ${str}\s*\n
		    ${com}*
		  | msgid_plural\ ${str}\s*\n
		    ${com}*
		    (?:msgstr\[\d+\]\ ${str}\s*\n${com}*)+
		)\Z}x)
	{
	    po_error('Mismatched quotes');
	    next ENTRY;
	}

	if (m{\n\#(?:,\ [-a-z0-9]+)*,\ fuzzy}) {
	    # Fuzzy entries aren't used, so ignore them.
	    # (This prevents warnings about mismatching <>/ pattern.)
	    next ENTRY;
	}

	# 0 for known not pango format, 2 for known pango format.
	my $is_pango_format = 1;
    	if (m{\n\#\.\ .*\bxgettext:(no-)?pango-format\s}) {
            $is_pango_format = ( defined($1) ? 0 : 2 );
	}
	if (m{\n\#(?:,\ [-a-z0-9+])*,\ (no-)?pango-format[,\n]}) {
	    $is_pango_format = ( defined($1) ? 0 : 2 );
	}
	if (m{\n\#:\ \.\./share/extensions/[-a-zA-Z0-9_]+\.(?:inx|py)(?:\.h)?:}) {
	    $is_pango_format = 0;
	}

	if (m{\A
	        ${com}*
	        (?:msgctxt\ ${str}\s*\n)?
	        msgid\ (${str})\n
	        ${com}*
	        msgstr\ (${str})\n
	        ${com}*
	      \Z}x)
	{
	    check_strs($is_pango_format, $1, $2) or next ENTRY;
	}
	elsif (m{\A
	           ${com}*
	           (?:msgctxt\ ${str}\s*\n)?
		   msgid\ (${str})\n
		   ${com}*
		   msgid_plural\ (${str})\n
		   ((?:${com}*msgstr\[\d+\]\ ${str}\n${com}*)+)
		 \Z}x)
	{
	    my ($s1, $s2, $rest) = ($1, $2, $3);
	    my @strs = ($s1, $s2);
	    while ($rest =~ s/\A${com}*msgstr\[\d+\]\ (${str})\n${com}*//) {
		push @strs, ($1);
	    }
	    $rest eq '' or die "BUG: unparsed plural entries `$rest'";
	    check_strs($is_pango_format, @strs) or next ENTRY;
	}
	elsif (m{$str[ \t]}) {
	    po_error('Trailing whitespace');
	    next ENTRY;
	} else {
	    po_error("parse error; may be a bug in po/check-markup");
	}
}

# Some makefiles (currently the top-level Makefile.am) expect this script to
# exit 1 if any problems found.
exit $rc;


# vi: set autoindent shiftwidth=4 tabstop=8 encoding=utf-8 softtabstop=4 filetype=perl :