Re: The current libc/locale/toupper.c is mistaken

From: Kohji Okuno <okuno.kohji_at_jp.panasonic.com>
Date: Fri, 11 May 2012 10:58:05 +0900 (JST)
Hi,

I'm sorry. I forgot to attach a file.

Regards,
 Kohji Okuno

> Hi David,
> 
> From: David Chisnall <theraven_at_FreeBSD.org>
> 
>> If you have a test case, I can commit it to the libc++ test suite.
>> 
>> David
> 
> I attached my test source.
> This test program shoud output as below.
> 
> towupper_l
> 0049, 0049
> 0131, 0049
> 0130, 0130
> 0069, 0049
> 
> towlower_l
> 0049, 0069
> 0131, 0131
> 0130, 0069
> 0069, 0069
> 
> But, when I use original toupper.c, this proguram output as below.
> 
> towupper_l
> 0049, 0049
> 0131, 0131
> 0130, 0069
> 0069, 0049
> 
> towlower_l
> 0049, 0069
> 0131, 0131
> 0130, 0069
> 0069, 0069
> 
> Regards,
>  Kohji Okuno
> 
>> 
>> On 10 May 2012, at 21:42, Kohji Okuno wrote:
>> 
>>> Hi Eric,
>>> 
>>>> I'm left wondering how this was not caught by the libc++ test
>>>> suite. The current toupper.c shouldn't pass
>>>> http://llvm.org/svn/llvm-project/libcxx/trunk/test/localization/locales/locale.convenience/conversions/conversions.character/toupper.pass.cpp
>>> 
>>> This test suite checks only popular characters.
>>> __mapupper_ext is used in case of special characters.
>>> 
>>> For example, Turkish 'i' (0x0131) should convert 'I' (0x49).
>>> 
>>> Regards,
>>> Kohji Okuno
>>> 
>>> 
>>>> Den 10/05/2012 kl. 12.03 skrev Dimitry Andric:
>>>> 
>>>>> On 2012-05-10 11:02, Kohji Okuno wrote:
>>>>>> I think that libc/locale/toupper.c is mistaken.
>>>>>> Could you check it?
>>>>>> 
>>>>>> _at__at_ -51,7 +51,7 _at__at_ ___toupper_l(c, l)
>>>>>> {
>>>>>>       size_t lim;
>>>>>>       FIX_LOCALE(l);
>>>>>> -       _RuneRange *rr = &XLOCALE_CTYPE(l)->runes->__maplower_ext;
>>>>>> +       _RuneRange *rr = &XLOCALE_CTYPE(l)->runes->__mapupper_ext;
>>>>>>       _RuneEntry *base, *re;
>>>>>> 
>>>>>>       if (c < 0 || c == EOF)
>>>>> 
>>>>> Yes, this definitely looks like a copy/paste error, introduced here:
>>>>> 
>>>>> http://svnweb.freebsd.org/base/head/lib/libc/locale/toupper.c?r1=165903&r2=227753
>>>>> 
>>>>> I'll commit the fix tonight (CEST), if David isn't faster than me. :)
>>>> 
>>>> I'm left wondering how this was not caught by the libc++ test suite. The current toupper.c shouldn't pass http://llvm.org/svn/llvm-project/libcxx/trunk/test/localization/locales/locale.convenience/conversions/conversions.character/toupper.pass.cpp
>>>> 
>>>> Thanks,
>>>> Erik_______________________________________________
>>>> freebsd-current_at_freebsd.org mailing list
>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>>>> To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
>> 
>> _______________________________________________
>> freebsd-current_at_freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>> To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"

#include <stdlib.h>
#include <stdio.h>
#include <locale.h>
#include <wctype.h>
#include <xlocale.h>

int
main()
{
	int i;
	wint_t test[] = {0x0049, 0x0131, 0x0130, 0x0069};
	wint_t ret;
	locale_t x;
	x = newlocale(0x1f, "tr_TR", NULL);

	printf("towupper_l\n");
	for (i = 0; i < sizeof(test)/sizeof(wint_t); i++) {
		ret = towupper_l(test[i], x);
		printf("%04x, %04x\n", test[i], ret);
	}
	printf("\n");

	printf("towlower_l\n");
	for (i = 0; i < sizeof(test)/sizeof(wint_t); i++) {
		ret = towlower_l(test[i], x);
		printf("%04x, %04x\n", test[i], ret);
	}

	exit(0);
}
Received on Thu May 10 2012 - 23:58:14 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:26 UTC