On Fri, May 15, 2026 at 11:23 AM Tomasz Kaminski <[email protected]>
wrote:

>
>
> On Fri, May 15, 2026 at 11:15 AM Jakub Jelinek <[email protected]> wrote:
>
>> On Fri, May 15, 2026 at 11:07:40AM +0200, Tomasz Kamiński wrote:
>> > --- /dev/null
>> > +++ b/libstdc++-v3/testsuite/std/format/debug_nonunicode_neg.cc
>> > @@ -0,0 +1,18 @@
>> > +// { dg-options "-fexec-charset=ISO8859-1" }
>>
>> I think this should be ISO-8859-1 instead
>>
> We use the above name as an "exec-charset" and locale name in a few other
> tests.
> and they seem to work as expected. I think this is because  “Charset Alias
> Matching” algorithm described
>  in the Unicode Technical Standard (we use that in text_encoding) ignores
> them,
> but I will fix it.
>
IANA database uses  ISO-8859-1 as name for this encoding.

>
>> > +// { dg-do compile { target c++26 } }
>> > +// { dg-require-effective-target cxx11_abi }
>>
>> I think this is missing
>> // { dg-require-iconv "ISO-8859-1" }
>> Not all target support all -fexec-charset= arguments.
>>
> Thanks will add that, and do separate commit also  fix other files.
>
>>
>>         Jakub
>>
>>

Reply via email to