[cfe-dev] [cfe-commits] [PATCH] pretty display of unprintable source and correct caret/range/fixup alignment

[Seth Cantrell]

Okay. The current version of the patch with a generic implementation
does not make this assumption.

 - Seth

On Apr 16, 2012, at 6:41 AM, Joerg Sonnenberger <joerg at britannica.bec.de> wrote:

> On Sun, Apr 15, 2012 at 07:56:12PM -0400, Seth Cantrell wrote:
>> But I do have to convert from UTF-8 to the host's wchar_t encoding in
>> order to use these functions. Currently I'm creating a locale where
>> the narrow character encoding is UTF-8 so I can convert from UTF-8 to
>> the wchar_t encoding. And since I go ahead and use that for the other
>> locale based functions. I could assume that the wchar_t is UTF-16 or
>> UTF-32 depending on the sizeof(wchar_t), and use the ConvertUTF8ToUTFXX
>> functions. As far as I know this holds for the platforms we care about.
>
> It is not a portable assumption. At least NetBSD explicitly does not
> make any such promise, since it can lose information. You can assume it
> iff __STDC_ISO_10646__ is defined.
>
> Joerg
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev