commit: b54f286110c85c5808594b0ced0cb8a5f08994b5
Author: Kerin Millar <kfm <AT> plushkava <DOT> net>
AuthorDate: Sat Aug 9 19:37:41 2025 +0000
Commit: Kerin Millar <kfm <AT> plushkava <DOT> net>
CommitDate: Sat Aug 9 19:37:41 2025 +0000
URL: https://gitweb.gentoo.org/proj/locale-gen.git/commit/?id=b54f2861
Disambiguate a mismatching codeset/charmap as a class of error
Consider the following two locale definitions, both of which are
erroneous.
# Invalid because the codeset part mismatches the charmap
zh_CN.GB18030 UTF-8
# Invalid because the charmap does not exist
zh_CN GB2319
Presently, locale-gen(8) will raise the same diagnostic for both cases,
complaining of an "invalid/mismatching charmap". Render the
parse_config() subroutine able to distinguish between both cases so that
the diagnostics can be improved.
Mismatching codeset/charmap at /etc/locale.gen[2]: "zh_CN.GB18030 UTF-8"
Invalid charmap at /etc/locale.gen[5]: "zh_CN GB2319"
Signed-off-by: Kerin Millar <kfm <AT> plushkava.net>
locale-gen | 16 ++++++++--------
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/locale-gen b/locale-gen
index cb54ed5..fb1a9bf 100755
--- a/locale-gen
+++ b/locale-gen
@@ -328,13 +328,15 @@ sub parse_config ($fh, $path, $locale_by, $charmap_by) {
# Extract the specified locale and character map. Upon success,
# a canonicalised representation of the locale is also returned.
- my ($locale, $charmap, $canonical) = parse_entry(@fields);
+ my ($locale, $codeset, $charmap, $canonical) =
parse_entry(@fields);
# Validate both locale and character map before accepting.
if (! $locale_by->{$locale}) {
$thrower->('Invalid locale', $line);
+ } elsif (defined $codeset && $codeset ne $charmap) {
+ $thrower->('Mismatching codeset/charmap', $line);
} elsif (! $charmap_by->{$charmap}) {
- $thrower->('Invalid/mismatching charmap', $line);
+ $thrower->('Invalid charmap', $line);
} else {
push @locales, [ $locale, $charmap, $canonical ];
}
@@ -345,21 +347,19 @@ sub parse_config ($fh, $path, $locale_by, $charmap_by) {
sub parse_entry ($locale, $charmap) {
my $canonical;
+ my $codeset;
if (2 == (my @fields = split /@/, $locale, 3)) {
# de_DE@euro ISO-8859-15 => de_DE.ISO-8859-15@euro
$canonical = sprintf '%s.%s@%s', $fields[0], $charmap,
$fields[1];
} elsif (2 == (@fields = split /\./, $locale, 3)) {
# en_US.UTF-8 UTF-8 => en_US.UTF-8
- $locale = $fields[0];
- $canonical = "$locale.$charmap";
- if ($fields[1] ne $charmap) {
- $charmap = '';
- }
+ ($locale, $codeset) = @fields;
+ $canonical = "$locale.$codeset";
} elsif (1 == @fields) {
# en_US ISO-8859-1 => en_US.ISO-8859-1
$canonical = "$locale.$charmap";
}
- return $locale, $charmap, $canonical;
+ return $locale, $codeset, $charmap, $canonical;
}
sub check_archive_dir ($prefix, $locale_dir) {
