but it compiles fine with MAKEOPTS="-j2". Can the ebuild filter out higher than 2?
no, you'll have to provide more information as to what's going on here.
Created attachment 38827 [details] log of failed compile Sorry, thought I'd attached something similar to this already...
I think I'd like to re-open this. I have a similar phenomenon. I haven't attempted to narrow the parameters down yet, but MAKEOPTS="-j1" emerge -v kvirc built kvirc-3.0.1-r1, while MAKEOPTS="-j6" (using distcc across three machines) failed to compile. I didn't try with -j2. This particular machine has kvirc in /etc/portage/package.keywords, and thus built the latest unstable, 3.0.1-r1. Tom Wesley -- if you're still following this bug, did you have "distcc" enabled, or were you simply trying to run three jobs concurrently on a single machine? Meanwhile, I'll try some other combinations, like -j2, -j3, etc., with distcc off to see if I can isolate what's happening.
If it's parallel make's fault, I don't see how using -jN with different N can help narrow down the problem. Rather, run it with -j (no N == unlimited parallel processes) to maximize the chance of encountering the issue, and post the failed build's log here.
Sounds like a plan ... should I disable "distcc" as well? I still don't know if the original poster had it enabled.
distcc shouldn't matter much with -j IMHO.
Created attachment 50298 [details] Log of attempted emerge with MAKEOPTS="-j" ... failed to compile
Created attachment 50299 [details] "emerge -v info" after failed emerge of kvirc
Here you are -- I left "distcc" enabled and a number of error messages from "distcc" are present in the log. Please let me know if you want me to try this without distcc.
