Hi, On 18/08/07, Mark Pustjens <[EMAIL PROTECTED]> wrote: > On Fri, 17 Aug 2007, Anibal Avelar wrote: > > On 8/17/07, Sascha Vogt <[EMAIL PROTECTED]> wrote: > >> About the broken utf8 support, i dunno wether this can be fixed quickly? > > > > Yes, it problem between libncurses5 and libncursesw5 (wide support for > > UTF8). I think it was easy but I need to think the solution. How do > > you see a special compile option called --with-utf8? I think it would > > be the better and clean solution. > > > > This cannot be easily fixed, as I pointed out multiple times on this list. > Non-ascii languages will only work if the locale used encodes all > characters in one byte. > > Multibyte encodings will probably not work, as CenterIM does not use any > functions of ncursesw which can render wide character strings.
The current centerim package already compiles with libncursesw when available, libncurses otherwise. The only advantage in doing this is that UTF-8 characters can be correctly visualized, but this is far from UTF-8 support. Actually thinking over and over about the issue we could hack the code to have decent UTF-8 support, we'd need to have a correct conversion between protocol encodings and local encoding and we should adjust the way text buffers are handled (which is poor design anyway) -- Stéphane _______________________________________________ Centerim-devel mailing list [email protected] http://centerim.org/mailman/listinfo/centerim-devel
