offhand, i doubt that this is related to the recent makepkg config changes; because that was three weeks ago, and everything seemed fine last week i have been compiling kernels for about 10 days - the last builds were very slow for all arches - i noticed because the ARM build is taking an absurdly long time - i compared the build times for this release to the ones from last week, and the stats are striking
the difference could be related to the kernel version - that is the only known variable; but it is hard to believe such a difference - maybe something in the config is accounting for this - or maybe beefcake just needs a reboot after spending so much time in swap-land the kernel is maybe more suspect; but its hard to know for sure, because of the build order - on the 13th, the ARM build completed (in a typical time-frame) and i started the x86_64 build of the other release - that x86_64 build took much longer than usual, nearly the same amount of time as the ARM build, when all things being equal, there is normally an order of magnitude difference - OTOH, it could be that the ARM build put beefcake in a bad state whatever the cause, it is a remarkable performance difference - i dont want to stop the arm build, just in case it completes, and i dont want to start any more in case it needs all of the RAM; so i wont be able to try other package until then - iceweasel is ready in my queue now though - that would be a good test linux-libre 6.9.3 x86_64 | june 7th | 3h 30m i686 | june 10th | 4h 40m armv7h | june 13th | 16h 40m linux-libre 6.6.32 x86_64 | june 13th | 16h i686 | june 13th | 8h (failed - did not complete) armv7h | june 14th | >72h (is still running) just a heads up to luke - maybe you have some ideas - i probably will want to reboot beefcake tomorrow (whenever the ARM build completes or fails); but maybe an iceweasel experiment would be informative first _______________________________________________ Dev mailing list [email protected] https://lists.parabola.nu/mailman/listinfo/dev
