Re: a few test results
"David A. van Leeuwen" <david@tm.tno.nl> writes:
> Hi,
>
> It seems sid/alpha is now in slink. Good. The Packages files seem to
> refer to unstable a lot, it seems. Nasty is you have an incomplete
> mirror, e.g., only slink/binary-al{pha,l}.
Ups, I have to watch my mirror script then, it will download
everything again, because its moved. :(
> I am running a limited set of packages, in a math/networking
> environment.
>
> I upgraded just this morning.
>
> `man' seems to give `unimplemented OSF mmap flags 0201' a lot.
I have to agree and disagree there. pre2.0.36 gives "unimplemented OSF
mmap" warningsm, but 2.1.128 works without glitch. There seems to be
some stuff missing in the stable kernels for alpha that man uses.
> `XF86_SVGA', `identd' give unaligned traps.
Hmm, I use the Mach64 driver.
> `ypbind' stops working correctly after an hour or so, producing
> enourmous amounts of error messages. "Alpha kernel: ypbind: memory
> violation at pc=00000000 rp=1555590112c (bad address = 00000000)"
ypbind seems to go berserk after some time on any system. Maybe on
alpha it shows up as a memory violation. Please try to debug ypbind
(attach gdb to it while running and wait for a crash). You probably
have to recompile ypbind first to get debug infos into it. Watch that
you have "-O -g" as cflags.
> `emacs20' hangs at random moments, claiming 100% cpu time. Especially
> in text mode. Emacs19 seems OK.
Yep, thats very anoying. xemacs latest beta works fine for me. I heard
that the emacs hangs are probably a gcc optimisation bug, maybe "-O
-g" helps.
> System is a pc164sx, kernel 2.0.35+patch0.2
Try a 2.1.128 or better kernel. You find a very small patch for
2.1.129 (I think), that fixes two missing include files, at:
ftp://mirjam.informatik.uni-tuebingen.de/pub/alpha/
May the Source be with you.
Goswin
Reply to: