IRC logs for #buildstream for Tuesday, 2020-03-10

*** narispo has quit IRC00:58
*** narispo has joined #buildstream00:59
*** hasebastian has joined #buildstream04:11
*** hasebastian has quit IRC07:03
*** hasebastian has joined #buildstream08:12
*** traveltissues has joined #buildstream08:15
*** hasebastian has quit IRC08:22
*** narispo has quit IRC08:52
*** narispo has joined #buildstream08:53
*** santi has joined #buildstream09:30
*** rdale has joined #buildstream09:35
*** narispo has quit IRC09:35
*** narispo has joined #buildstream09:36
*** tme5 has joined #buildstream09:40
*** jonathanmaw has joined #buildstream10:18
*** traveltissues has quit IRC11:50
valentindjuergbi, If there are commits we can backport from master to support 3.8, this would be great. But when I was looking for fixes in master I did not find what was related to fixing running on 3.8.11:57
*** traveltissues has joined #buildstream11:57
juergbimaybe not all commits were clearly marked as related, some might have been general cleanup/improvement that happens to be important for 3.811:58
juergbivalentind: I've linked a few commits in the MR, though11:58
valentindNot easy to backport with the changed paths.12:01
*** traveltissues has quit IRC12:03
*** traveltissues has joined #buildstream12:04
juergbivalentind: if that's the main issue, a possible workaround is to edit/sed the `src/` out of patch files before cherry-picking them12:07
juergbigit rebase --onto might also work better than cherry-pick across renames12:07
valentindrebase did not work, I had to use patch -p212:08
valentindAnd the code is quite changed.12:08
juergbithe black reformatting is probably a culprit as well12:11
valentindOK, managed to port them. Now running the tests.12:22
valentindBut I am not sure it is enough.12:22
valentindBecause after these 4 commits it still calls add_child_handler outside of the loop.12:22
valentindSo far the tests work.12:25
valentindAh maybe the fix is fc6b3f37f173fd19e5a2593b7bf8cce40f03c61512:26
gitlab-br-botvalentindavid opened MR !1830 (valentindavid/bst-1/python3.8-with-backports->bst-1: [BuildStream 1.4] Backport changes to enable Python 3.8) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/183012:31
valentindIt works. Just one test fails but this is because of the same bug in Python that triggers only during testing because of coverage.12:50
valentindIt is missing some fixes. Otherwise it prints too much in the console after terminating13:00
valentindI wonder if master has the issue.13:01
*** traveltissues has quit IRC13:08
*** traveltissues has joined #buildstream13:08
valentindIt seems we receive the SIGTERM we send to the child processes back in the parent13:27
*** CTtpollard is now known as tpollard13:30
tpollardthat rings a bell13:30
juergbivalentind: sounds like https://gitlab.com/BuildStream/buildstream/commit/278d010b873b59aa5eed577cb92dd1342a6f6f3413:45
valentindI think it inherits the signal handler at fork.13:45
juergbiyes, that commit should fix it13:46
valentindSeems to work better yes13:58
*** santi has quit IRC14:00
*** santi has joined #buildstream14:00
gitlab-br-botjuergbi approved MR !1830 (valentindavid/bst-1/python3.8-with-backports->bst-1: [BuildStream 1.4] Backport changes to enable Python 3.8) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/183014:19
*** narispo has quit IRC14:21
*** traveltissues has quit IRC14:31
valentindjuergbi, thank you.14:34
gitlab-br-botvalentindavid merged MR !1830 (valentindavid/bst-1/python3.8-with-backports->bst-1: [BuildStream 1.4] Backport changes to enable Python 3.8) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/183014:35
gitlab-br-botvalentindavid closed MR !1829 (valentindavid/bst-1/python3.8->bst-1: [Buildstream 1.4] Add support for Python 3.8) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/182914:35
*** traveltissues has joined #buildstream14:45
*** santi has quit IRC14:56
*** santi has joined #buildstream14:57
*** pulsar has joined #buildstream15:44
*** hasebastian has joined #buildstream15:45
*** pulsar has quit IRC15:47
*** phildawson_ has quit IRC16:35
*** phildawson_ has joined #buildstream16:37
*** phildawson has joined #buildstream16:38
*** tme5 has quit IRC17:32
*** traveltissues has quit IRC17:54
*** santi has quit IRC17:58
*** jonathanmaw has quit IRC18:10
*** phildawson_ has joined #buildstream18:16
*** phildawson has quit IRC18:17
*** hasebastian has quit IRC18:17
*** juergbi has quit IRC18:17
*** flatmush has quit IRC18:17
*** douglaswinship has quit IRC18:17
*** paulsherwood has quit IRC18:17
*** benbrown has quit IRC18:17
*** benbrown has joined #buildstream18:19
*** flatmush has joined #buildstream18:20
*** paulsherwood has joined #buildstream18:20
*** douglaswinship has joined #buildstream18:21
*** juergbi has joined #buildstream18:24
*** phildawson_ has quit IRC18:35
*** slaf_ has joined #buildstream18:35
*** phildawson_ has joined #buildstream18:37
*** slaf has quit IRC18:37
*** slaf_ is now known as slaf18:37
*** phildawson_ has quit IRC18:40
*** traveltissues has joined #buildstream18:50
gitlab-br-botjuergbi merged MR !1827 (juerg/vdirectory->master: Improve `Directory` API) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/182719:00
*** narispo has joined #buildstream21:39
*** narispo has quit IRC21:41
*** narispo has joined #buildstream21:42
*** narispo has quit IRC22:25
*** narispo has joined #buildstream22:29
*** rdale has quit IRC23:01
*** narispo has quit IRC23:09
*** narispo has joined #buildstream23:10

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!