Re: libXt, libXmu, libXaw, libXtst, libdps

"Mike A. Harris" <[email protected]> Fri, 12 Dec 2003 13:50:04 -0500 (EST)
Newsgroups gmane.comp.freedesktop.xlibs.general
Message-ID <[email protected]>
On Thu, 11 Dec 2003, Jeremy C. Reed wrote:

>I have been using cvs for a couple years, but I still have lots to learn.
>I assume this is different than cvs add or cvs commit, because when I have
>commited code before with pre-existing CVS/Entries files, it appears that
>the revision numbers were reset for new commit.

Right, they would be.

>So would that be done by a CVS admin on the freedesktop.org CVS server
>itself?

Anyone with root access or project write access should be able to 
do it, carefully of course.


>Can one of the CVS admins copy in the Xmu, Xaw, Xtst, and dps
>directories?


>(Maybe Xaw should be done differently since there are alternatives to
>choose from)

I disagree, because there are many alternative forks of Xaw, but
none of them are in widespread use, and their quality is unknown.  
Xaw itself we know more or less "just works", and a few new bug 
fixes trickle in over time as they're found.

If we were to replace Xaw with any of the alternative Xaw forks, 
we would be bringing in code of unknown quality, which has not 
been tested on anywhere near the number of different OS/arch 
platforms as the existing code, and we would essentially be 
commiting to fixing any problems people find that are caused by 
this.  The alternative libraries might possibly change the way 
certain applications work out there in the wild in ways deemed 
negative to the userbase.

Personally, I consider Xaw to be a matured library which "just
works" more or less, and that it is included strictly for legacy
compatibility with ancient applications.  It's use in new
application development should be discouraged IMHO.  I think
doing anything to Xaw that changes it drastically just drops more
work on our laps unnecessarily, and will create massive problems
for various users out there, the majority of whom have probably
never even tried to compile or use their important legacy
applications with alternative Xaw implementations.

I would prefer to just see Xaw just slowly die into obsolesense 
the way that PEX and XIE did.


>Once those new modules are in place, I will check them out and
>start working on autotooling them. Then I can provide back
>patches to this list for review before I can commit myself.

Sounds good.

>Also, Mike mentioned that the CVS metadata needs to be cleaned up; I can
>fix all the CVS/Root files.

It might or might not need cleaning up, depends on where the 
files are obtained from and what their contents are.

Take care,
TTYL


-- 
Mike A. Harris