[fab] Re: openmotif

[Ralf Corsepius]

On Wed, 2006-08-30 at 15:18 -0500, Tom 'spot' Callaway wrote:
> On Wed, 2006-08-30 at 10:49 -0500, Tom 'spot' Callaway wrote:
> 
> > Actually, I think I can fix these things by Tuesday. Stay tuned.
> 
> OK, here's how we can do this:
> 
> 1. Lesstif is approved. It needs to incorporate my fixes (see BZ 203274)
> and be rebuilt.
Has the package made installable in parallel to OpenMotif?

If not, very bad decision

> 2. libGLw needs to be approved. I posted a fixed SRPM (see BZ 188974).
> It needs to be committed and rebuilt into FE.
> 3. lesstif needs to be rebuilt one more time after libGLw is built.
> Hooray for circular deps!
There is no circular dep. libGLw depends on Motif.

libGLw depends on Motif, that's all.

> 4. mesa without libGLw needs to be rebuilt in Core.
> 5. ddd and xpdf need to be moved to Fedora Extras,
Bummer, FC dropping the primary GUI-debugger - MASSIVE REGRESSION. 

>  their BR changed from
> openmotif-devel to lesstif-devel.
> 6. tetex-xdvi can pass --with-xdvi-x-toolkit=xaw to configure, and no
> longer depend on motif. It needs to do this and be rebuilt in Core.
> 7. At this point, we can destroy openmotif from orbit (with our newly
> repaired ORBITAL LASER).
> 8. cernlib, cmucl, geomview, gpsd, grace, Inventor, nedit, and xlockmore

+ OpenSceneGraph, it depends on Inventor.

> (all maintainers CC'd) need to change their BR from openmotif-devel to
> lesstif-devel. I tested all of these apps and they work correctly
> against lesstif.
Are you serious about this?

Motif is a GUI-toolkit and you can't have tested all apps's
functionality.

> 9. fbida is the only one in Fedora Core or Fedora Extras that doesn't
> work with lesstif, because it needs an actual implementation for
> XmeDropSink() (at a minimum, there may be more). Specifically, it is the
> "ida" binary in this package that needs openmotif, so I altered the
> fbida package to not create the -ida subpackage anymore:
> http://www.auroralinux.org/people/spot/review/MOTIF/fbida-2.06-1.src.rpm
> It needs to be rebuilt and the fbida-ida package nuked (Note: I didn't
> make the new package Obsolete/Provide fbida-ida...)
> 
> See? Less than 10 steps even? :)

11. Me filing "critical/highest priority" PRs, should I notice any
functional regression due to moving to lesstif and assigning them to
Pertusus and Callaway.

PS: To me, this mover is a massive regression and fault.

Ralf