Paul Eggert wrote:
> + if (setlocale (LC_ALL, "tr_TR.UTF-8") && really_utf8 ())
> +{
> + re_set_syntax (RE_SYNTAX_GREP | RE_ICASE);
> + if (re_compile_pattern ("i", 1, ®ex))
> +result |= 1;
> + else
> +{
> + /* UTF-8 encoding of U+0130 LATIN CAPITAL LETT
* lib/regex_internal.c (build_wcs_upper_buffer):
Do not assume that converting single-byte character to upper
yields a single-byte character. This is not true for Turkish,
where towupper (L'i') yields L'İ', which is not single-byte.
* tests/test-regex.c (main): Test for this bug.
---
ChangeLog
