Re: e_iconv_* replaced with g_iconv_*

On Mon, 2006-11-13 at 19:31 +0100, Philip Van Hoof wrote:
> On Mon, 2006-11-13 at 12:15 -0500, Jeffrey Stedfast wrote:
> > Because giconv doesn't know the mappings for charsets that eiconv does.
> > it's not about the libc apis they wrap, it's about using the correct
> > to/from charset strings in *_iconv_open() (and about the caching -
> > you'll notice that glib has a copy of e-iconv's cache, but at the
> > gconvert level rather than at the giconv level)
> 
> Aha, so a patch that would replace all the iconv functionality with
> gconvert could work then?

no, I just meant that the cache was implemented at the gconvert level.
that still doesn't solve the charset names issue.

> 
> It's indeed a little bit more work. Not undoable. 
> 
> Anyway, my only goal was/is to get rid of the iconv-detect.c thingy in
> the configure script, which is making cross compilation a little bit
> harder (a burden I'd rather don't put on packagers who'll target
> different architectures).
> 
> Other than perhaps removing code, it doesn't look like GIConv adds a lot
> to e_iconv_*
> 
> 
-- 
Jeffrey Stedfast
Desktop Hacker - Novell, Inc.
[email protected]  - www.novell.com