Re: [OK] 21.0 "Somali-pre2" (beta42) on MkLinux
[For faster response, if you have a problem with a build, the report
should be directed to xemacs-beta].
Stephen R Anderson <sra(a)bloch.ling.yale.edu> writes to xemacs-build-reports:
Built from new tarballs. Compiler = egcs-1.02
The matherr problem in src/floatfns.c still exists.
Can you expand a little on that?
After commenting out the definition of HAVE_MATHERR in src/config.h,
xemacs builds without incident. I realize that this problem results
from the fact that the MkLinux development environment is out of
date, but apparently there's nothing we can do about that for the
present. If someone could eliminate the matherr problem, it would be
a nice service to this community.
The fact that HAVE_MATHERR is defined means that configure is
detecting it. Can you send in the portion of config.log where this is
occurring? Grep for `checking for matherr' and snip the output from
there to the next `checking for %s' message.
Just how old is your environment? The last time somebody checked, I
think XEmacs was buildable on Linux/a.out with proper versions of the
libraries.
Beta42 runs fine, but I have one problem with sound, which I had not
noticed before. When xemacs wants to signal errors, I get a message
in the minibuffer:
audio: /dev/dsp, No such device.
I installed the sounds.au package. Why is xemacs trying to write to
a dsp?
Because it's supposed to? /dev/dsp is where the sound code outputs
(successfully) to, at least on Intel/Linux.
This isn't a NeXT machine....