Re: [gnet-dev] last patch

[Ian Peters <itp ximian com>]

To replicate the very sane behavior from glib 1.2, rather than the
obnoxious glib 2.0 defaults, you need

channel = g_io_channel_unix_new (fd);
g_io_channel_set_encoding (channel, NULL, NULL);
g_io_channel_set_buffered (channel, FALSE);

By default, glib 2.0 channels are utf-8 encoded and buffered, and only
channels with a NULL encoding may be unbuffered, so it must be set in
that order.  The third parameter to _set_encoding is for a GError, and
the ignored return value is a GIOStatus, but setting the encoding is
guaranteed to work for a newly created channel.

Thanks for the library, btw, you've helped me keep a safe distance from
the always-enjoyable world of unix network programming.  Hope this
helps.

Ian

On Mon, 2002-09-16 at 10:27, David Helder wrote:
> GNet is built against 1.2 or 2.0, so we could either have a #define or
> function.  It requires a variable, so a function would be best
> (like gnet_private_iochannel_new()).
> 
> Does calling g_io_channel_set_encoding() change the semantics any?  That
> is, will the iochannel in 2.0 behave just like it does in 1.2.
> 
> David
> 
> 
> On 13 Sep 2002, Ian Peters wrote:
> > I don't know how I convinced myself that last patch would help without
> > fixing GNET_SOCKET_IOCHANNEL_NEW in gnet-private.h, but it's rather more
> > difficult with it as a macro to set the encoding in that call.  It's
> > rather ugly to set the encoding everywhere it's called, though.
> >
> > Of course, I also forgot people could still be using glib-1.2.
> 
> 
> -- 
>       __          _    __ David Helder - dhelder umich edu
>   ___/ /__ __  __(_)__/ / <http://www.eecs.umich.edu/~dhelder>
>  / _  / _ `/ |/ / / _  /  Jungle Monkey: <http://www.junglemonkey.net>
>  |_,_/|_,_/|___/_/|_,_/   Paper CD Case: <http://www.papercdcase.com>
>