[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Subversion 1.4.6 : 3 tests fail when locale is set to Turkish (tr_TR.UTF-8)

From: Branko Čibej <brane_at_xbc.nu>
Date: 2007-12-22 23:47:26 CET

Ismail Dönmez wrote:
> Sunday 23 December 2007 00:29:00 tarihinde Ismail Dönmez şunları yazmıştı:
>
>> Sunday 23 December 2007 00:07:07 tarihinde Branko Čibej şunları yazmıştı:
>>
>>> Oh wow. linsvn_subr/opt.c:revision_from_word does this:
>>>
>>> else if (strcasecmp(word, "committed") == 0)
>>> {
>>> revision->kind = svn_opt_revision_committed;
>>> }
>>>
>> Just do a
>>
>> const char* locale = setlocale (LC_ALL,"C") before strcasecmp and
>>
>> setlocale(LC_ALL,locale)
>>
>> after the comparisons?
>>
>
> A quick grep shows,
>
> subversion/bindings/swig/perl/native/svn_wc.c
> 2746: else if (strcasecmp(input, "WORKING") == 0)
> 2748: else if (strcasecmp(input, "COMMITTED") == 0)
>
> subversion/bindings/swig/perl/native/core.c
> 6470: else if (strcasecmp(input, "WORKING") == 0)
> 6472: else if (strcasecmp(input, "COMMITTED") == 0)
>
> Potentially also not working as expected in Turkish locale, should Subversion instead have a ascii only wrapper for these functions?
>

I saw all those, yes. The Perl wrappers have a many such compares, and
the dotless-I isn't the only worrying aspect of that.

As for ASCII-only wrappers ... I suspect we have too many specialized
string functions as it is. Or too few, depending on point of view. :)
Also I believe our friends who port Subversion to OS/400 would be a bit
unhappy if we created ASCII-specific (or rather, "7-bit Unicode
subset"-specific) string functions.

-- Brane

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Dec 22 23:47:43 2007

This is an archived mail posted to the Subversion Dev mailing list.