Re: [LINUX5250] Minor updates in CVS

[Scott Klement <klemscot@xxxxxxxxxxxx>]

On Tue, 19 Mar 2002, Jay 'Eraserhead' Felice wrote:
> >
> > > One small problem, though.  b0-16 does not use glib.  (Only HEAD uses 
>glib)
> > >
> > > That's why the Windows support only exists in 0-16, because of glib.
> > > (Yes, I know they make glib for Windows, but it's beta-only, and I haven't
> > > had a chance to download and play with it.)
>
> It's not beta!  I used it _years_ ago, when gtk+ on Windows was not even
> alpha software, and glib was solid as a rock with Tor Lillqvist's changes
> folded back into the main tree (it's not that I feel the need to give credit,
> I just love the dude's name).  In fact, separating glib from gtk+ was
> partly motivated by the fact that glib was useful and more portable than
> gtk+.

It was listed on Tor's page as "beta" prior to March 10th.   The new
version is not listed as beta, and has already been integrated into the
Windows port of Tn5250.

>
> I think making the Windows port use glib would probably make things easier.
>

Jay, I'm not speculating here.   I wrote the non-Glib version 2 months
ago.  I wrote the GLib version last weekend.   Glib did NOT make it
easier.   GLib was a complication.  An extra hurdle that I had to jump
over in order to make it work. (Which I did)

The problem is, it didn't gain me anything.  The build process is now a
little more complicated.   The install now requires 3 more DLL's than it
did before.   And for what?  TN5250 works exactly the same as it did in
the non-glib version.

> >
> > Because of this I assume that scs2pdf won't work under Windows (scs2pdf
> > uses glib)?
>
> I wonder if glib has some sort of pipe or other inter-process communication
> feature?  It seems likely, as Tor was running into pipes not working those
> million years ago-- he was porting the Gimp, and Gimp plugins use pipes for
> communication.  I think I suggested using TCP sockets, as they pretty much
> work accross platforms.
>

Pipes work, they're just slow.  TCP sockets would work, but would require
changing the way scs2ascii & friends work, which is what James appears
to be protesting against.