Bugzilla – Bug 2867
Alsa-driver fails on Pentium 4 machines
Last modified: 2007-04-01 01:09:15 UTC
The line sedit "s:\-O2:${CFLAGS}:" configure && causes pcm-oss.o to fail to build.
Ahhh....so that's why it fails on my g/f's P4 ('course she has a faster computer than me :(). Should we have BUILD check the -mcpu line of CFLAGS, and if it's "pentium4", then don't run sedit, otherwise do? if ! "cat ${CFLAGS}|grep -q -mcpu" = "-mcpu=pentium4" then sedit "s:\-O2:${CFLAGS}:" configure && fi Should work, haven't tested the above, but I did remove the sedit line and alsa-driver now compiles. :)
This fix is easy, available and works, is it going to happen?
I applied the fix to devel, try casting it tomorrow (from devel) after the grimoire updates and let me know if it works.
Seems there's a syntax error somewhere in that fix. Using gcc version: 3.2.2 Installing in dir: /usr/src/alsa-driver-0.9.2 alsa-driver 0.9.2 /tmp/casting.18621: /var/lib/sorcery/codex/devel/audio/alsa-driver/BUILD: line 31: syntax error near unexpected token `fi' /tmp/casting.18621: /var/lib/sorcery/codex/devel/audio/alsa-driver/BUILD: line 31: `fi' ! Problem Detected ! Creating /var/log/sorcery/compile/alsa-driver-0.9.2.bz2
Fixed the syntax in devel.
Build is still broken on p4's check the fix
What's the error you get? With this fix it compiled fine on my g/f's PIV-1.8GHz.
This is second person information, I installed my friend's machine and submitted this bug based on his system, its a PIV 2.2GHz, I'm getting this info from him, perhpas its the awk error that I've been expereienceing but continuing the compile works, I'll have to talk to him again.
Which grimoire is he using? Does he have the syntax error? If so, have him do a 'scribe update' and see if that fixes it, if not, set this bug to the appropriate grimoire and I'll pull it through, though I believe the fix has been pulled to test, at least.
Figured out what the error is, its a bug in gcc3.2 (http://gcc.gnu.org/PR10077) its fixed in 3.3 so we just need to wait for the merge.
Seems this is fixed now
Should we remove the pentium4 fix code from BUILD then (and alsa-driver-virmidi)?
When I say fixed I mean that the fix implimented is good, it should not be removed.
gcc 3.3 didn't fix it? I know the patch fixed it, but gcc 3.3 was supposed to be fixed so that we wouldn't need to change our flags to compile anymore on a pentium4 (it fixed the mozilla bug).
Whoops yeah forgot thats what the fix was for in the first place. Tested it on my friend's machine, works fine. So to change my answer YES the fix can be removed since GCC 3.3 fixes the compile bug.
Done, thanks. :)
Closure....
reassign to sm-grimoire-bugs