=== morphis9 is now known as morphis | ||
b-rad | hi i have a strange build issue, where 32bit failed but 64bit completed. The 32bit failure was due to "stray ‘\1’ in program", but I don't see any such character in a hex editor | 16:51 |
---|---|---|
cjwatson | Have you tried reproducing the bug locally in e.g. sbuild? | 16:53 |
cjwatson | It's very unlikely to be a Launchpad-specific problem. | 16:53 |
b-rad | i'll look into setting up sbuild, i just do standard sanity builds for 64bit | 16:53 |
b-rad | just weird that for the same code 32bit pukes, but not 64bit | 16:53 |
b-rad | and unfortunately where build log says there's a \1 there isn't | 16:54 |
cjwatson | Can't say any more without a link to the failure | 16:55 |
b-rad | https://launchpadlibrarian.net/435131782/buildlog_ubuntu-bionic-i386.linux_4.15.0-55201907291620.0+mediatree+hauppauge_BUILDING.txt.gz | 16:55 |
cjwatson | I suppose stray memory corruption is always a possibility (retrying would confirm whether it's just that), but would be surprising | 16:56 |
cjwatson | If it persists beyond that then you probably need to consult a kernel developer for help | 16:56 |
b-rad | i guess i'll push rebuild and see, was hesitant to do that without inquiring if this had been seen before | 16:57 |
b-rad | as well as get a sbuild docker setup | 16:57 |
cjwatson | It's not a familiar problem, no | 17:00 |
b-rad | cjwatson, must have been a stray cosmic ray collision, the rebuild succeeded | 17:57 |
b-rad | lcy01-amd64-024 was the machine | 17:57 |
cjwatson | b-rad: OK, I guess with enough builds going through small probabilities add up | 21:36 |
b-rad | i shall be that statistical anomaly | 21:38 |
cjwatson | (The machines are ephemeral VMs, so there's unlikely to be much trace left) | 21:39 |
b-rad | i saved the build log in case it was desired, but i doubt it would be of any in sight besides the host machine | 21:39 |
cjwatson | No, I don't think so, indeed | 21:43 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!