Michael Sperber <sperber(a)informatik.uni-tuebingen.de> writes in xemacs-beta(a)xemacs.org:
>>>>>> "Valdis" == Valdis Kletnieks <Valdis.Kletnieks(a)vt.edu> writes:
Valdis> [1 <text/plain; us-ascii (7bit)>]
Valdis> On 05 Jan 1999 19:20:18 +0100, Michael Sperber [Mr. Preprocessor] said:
>>> From the PROBLEMS file:
>>>
>>> ** AIX
>>> *** On AIX 4.3, you must specify --with-dialogs=athena with configure
Valdis> Until such time as we track down the actual problem here, would it make
Valdis> sense for ./configure to default this, and to urinate all over the
Valdis> user if they attempt to overrride it, since we *know* it won't work?
> I *did* figure out what the problem is in a rough way: Motif pees all
> over the data area with allocated storage.
Lovely.
> My general stance is that it's a bad idea to cater to OS fuckage in
> configure, because it will alleviate pressure to get the actual bug
> fixed.
How are going to force Motif bugs to get fixed? As near as I can tell
Motif has pretty much always been broken in some way or another.
> But I know Steve has felt differently about this in the past. I'll
> leave it to him.
Death to Motif. Good riddance to bad rubbish. I think it's a serious
mistake defaulting to Motif support.
Valdis> I'm pretty sure that, unlike the 'xemacs -nw locks up' problem, this
Valdis> one isn't a problem in the IBM X libraries, since I build against
Valdis> the MIT X11R6.4 and *still* see it. I haven't looked to see if it's
Valdis> an IBM xlc compiler fuckage (possibly), or if it's an actual XEmacs bug
Valdis> (but why isn't it seen elsewhere?)
> No, it's Motif.
> BTW: We still do need a note in PROBLEMS about AIX X libXt, don't we?
That would be good.