=== sem2peie- is now known as sem2peie === happyaron is now known as Guest35098 === _hc is now known as Guest95017 [20:16] hey, isc-dhcp fails to migrate because its regression w/ network-manager; cf: https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#isc-dhcp [20:16] but the autopkgtest is only failing on ppc64el and I tried to figure out but no luck. [20:17] it's reproducible as well (I did so on canonistack). [20:18] if anybody has any clue about this, that'd be really helpful! TIA! :) [22:33] utkarsh2102: hmm i wonder if anyone has ever put a wifi card in a ppc64el system :) [22:34] mwhudson: heh :P [22:34] these tests have been failing for a while now [22:34] blocking isc-dhcp :( [22:34] yeah just looking at the history [22:35] does seem to be a genuine regression in that they pretty much always pass with the older isc-dhcp and always fail with the new [22:36] indeed! [22:36] but it's weird that only ppc64el fails [22:37] yes [22:37] it seems to fail with 4.4.1-2.2ubuntu1 [22:38] which comes with this diff: http://launchpadlibrarian.net/523135764/isc-dhcp_4.4.1-2.1ubuntu9_4.4.1-2.2ubuntu1.diff.gz [22:38] and it passes with 4.4.1-2.1ubuntu9 [22:39] so um,, [22:40] although the changlog for ubuntu9 makes me go "hmm" https://launchpad.net/ubuntu/+source/isc-dhcp/4.4.1-2.1ubuntu9 [22:41] oooh! [22:42] er so the gcc 10 fixes break ppc64el, get reverted in ubuntu, nmued into debian and get into ubuntu again via merge? [22:42] and break ppc64el again [22:42] hahahaha [22:42] xnox: hello [22:43] although that said [22:43] it can't be the gcc-10 fixes that break this [22:43] they're trivial, must be just gcc-10 vs gcc-9? [22:44] i wonder if disabling -O3 will make this work... [22:44] could this be "it": http://launchpadlibrarian.net/493173027/isc-dhcp_4.4.1-2.1ubuntu8_4.4.1-2.1ubuntu9.diff.gz? [22:44] utkarsh2102: yeah but i'm curious now :) [22:45] utkarsh2102: do you still have a set up to reproduce this? [22:45] no, but it should be fast to shoot up a VM at canonistack [22:46] utkarsh2102: two things to try would be building with gcc-9 vs default gcc [22:47] and something like the rules change from http://launchpadlibrarian.net/516062485/w1retap_1.4.4-4ubuntu1_1.4.4-4ubuntu2.diff.gz [22:47] sigh why doesn't "nova list" include arch :) one of these is ppc64el i think... [22:48] yeah, openstack image list, ftw :) [22:48] it's 4:18 AM here atm so I won't be troubleshooting now at least [22:49] utkarsh2102: i'll have a poke, go to bed :) [22:49] awesome, thank you! <3 [22:50] let me know if case you find anything [22:50] interesting or otherwise :) [22:50] utkarsh2102: will do [22:50] \o/ [22:59] hm how do you run a packages autopkgtests locally and satisfy depends on another source packages packages from the debs you just built? apt-ftparchive games? [23:00] You can just give autopkgtest extra .debs on its command line, I think [23:00] oh right! [23:01] Pretty sure they don't have to be from the source package under test or anything [23:01] i thought it was easy, just failed reading comprehension apparently [23:08] yeah, just passing debs should be enough [23:09] eg of what I use: `autopkgtest -B ../*.deb ../../faye-websocket/ruby-faye-websocket_0.11.0-1_all.deb -- schroot unstable-amd64-sbuild` [23:11] utkarsh2102: go to bed! [23:11] araragh [23:11] wpa-dhclient SKIP Test requires machine-level isolation but testbed does not provide that [23:11] i guess i shouldn't be surprised really [23:13] you can update from bionic to hirsute with sed -e s/bionic/hirsute/ /etc/apt/sources.list right? :) [23:17] mwhudson: eee, yeah, just winded up everything o/ [23:18] last comment though: you can install the locally built deb and then from the source of network-manager, run debian/tests/wpa-dhclient [23:18] at least that's how I reproduced easily :) [23:19] see you tomorrow and thanks a bunch for taking a look! \o/