archive/php-src
1
0
Fork 0
mirror of https://github.com/php/php-src.git synced 2025-08-15 21:48:51 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Fix legacy conversion filter for CP50220

Browse source
This commit is contained in:
Alex Dowad 2022-07-05 08:20:19 +02:00
parent 1662f7f79f
commit 6938e35122
2 changed files with 11 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ext/mbstring/libmbfl/filters/mbfilter_cp5022x.c
View file

@ -338,7 +338,11 @@ static int mbfl_filt_conv_wchar_cp50220(int c, mbfl_convert_filter *filter)
if (filter->cache) {
int s = mbfl_convert_kana(filter->cache, c, &consumed, NULL, mode);
filter->cache = consumed ? 0 : c;
/* Terrible hack to get CP50220 to emit error markers in the proper
* position, not reordering them with subsequent characters */
filter->filter_function = mbfl_filt_conv_wchar_cp50221;
mbfl_filt_conv_wchar_cp50221(s, filter);
filter->filter_function = mbfl_filt_conv_wchar_cp50220;
} else if (c == 0) {
/* This case has to be handled separately, since `filter->cache == 0` means
* no codepoint is cached */

7
ext/mbstring/tests/cp5022x_encoding.phpt
View file

@ -375,6 +375,13 @@ $converted = mb_convert_encoding("ab\x00", 'UTF-16BE', 'CP50220');
if ($converted !== "\x00a\x00b\x00\x00")
die("Bad handling of trailing null byte (got " . bin2hex($converted) . ")");
// Previously, the CP50220 implementation would reorder error markers with
// subsequent characters
mb_substitute_character(0x3F);
$converted = mb_convert_encoding("\xff\xff\x00&", 'CP50220', 'UTF-16BE');
if ($converted !== '?&')
die("Bad handling of erroneous codepoint followed by good one (got " . bin2hex($converted) . ")");
?>
--EXPECT--
ASCII support OK