=== gusnan is now known as Guest97256 [12:21] I spent some time looking into the node-sha.js autopkgtest failures on ppc64el, which (along with other test failures in various node-* packages) prevent the migration of nodejs [12:21] it turns out it's a nasty bug in V8 that incorrectly optimizes a function on ppc64el only (https://github.com/nodejs/node/issues/34666) [12:22] this requires further investigation upstream, and I'm not sure what the best approach would be to get this moving forward in Ubuntu [12:22] (nodejs in turns is blocking firefox and thunderbird) [12:22] s/turns/turn/ [12:23] suggestions welcome [12:27] oSoMoN, would building with a lower -O workaround the optimization problem? [12:32] seb128, I don't know, I'm not familiar with how V8 optimizes code, but I can certainly try [12:33] oSoMoN, I'm just random suggesting, let's see if maybe doko or xnox can help there [12:36] oSoMoN: feel free to ignore me but I'm kind of a fan of the minimal `eval('')` patch you mention in the bug report [12:37] icey, that would just hide one symptom of the problem, but V8 would potentially incorrectly optimize other functions in other packages [12:37] fair enough [13:54] sil2100, hey, I just commented on bug #1889106, if you're still around and have some time for it, would you mind accepting chromium-browser 84.0.4147.105-0ubuntu0.20.04.1 into focal-proposed? [13:54] bug 1889106 in chromium-browser (Ubuntu Focal) "[SRU] Upgrading from 18.04 to 20.04 will keep the 18.04 chromium-browser due to a higher version than the transitional deb in 20.04" [High,Fix committed] https://launchpad.net/bugs/1889106 [13:55] that would fix the deb->snap upgrade path for users upgrading from bionic to focal [13:59] seb128: I have a merge ready for gnunet, but the my_bool change is still needed [13:59] I opened a bug upstream this time [13:59] that's our only delta [14:05] ahasenack, right, thank you! [14:05] seb128: it doesn't have dep8 tests, and I really don't know much about the pkg [14:06] seb128: was it blocking/breaking something else, something I could test with this new version? [14:06] otherwise, I'll jut get a quick review and upload [14:07] ahasenack, something was depwaiting on it on the proposed report but I forgot what ... the delta is minimal and that could be an autosynced update, I would trust Debian and not spent too much effort trying to test it [14:08] ok [14:08] I just got a build failure, taking a quick look [14:08] might be gcc10 fun [14:09] seb128: ah, gnunet-fuse perhaps [14:09] and all the other gnunet-* src packages looks like [14:09] which are syncs [14:10] ahasenack, right [14:10] ok, I ll handle it [14:27] ok, ftbfs resolved [16:18] Are the pipelines that build images for vagrant, etc public? As I understand, people in ubuntu build them. [17:08] vorlon: hi, good morning/afternoon. While reviewing the base-files changes regarding motd-news-config, cpaelzer came across one case which we are wondering if it's worth addressing [17:08] vorlon: basically the scenario is you have the current base-files and ubuntu-server installed, and the user removes /etc/default/motd-news (instead of editing it) [17:09] in that scenario, when you do the upgrade, /etc/default/motd-news is reinstalled when the new motd-news-config package is installed [17:09] you start with base-files/ubuntu-server, upgrade, and end up with base-files, ubuntu-server, and motd-news-config, as expected. But with the config file back, as it's now part of motd-news-config, and before, when the user removed it, it was part of base-files [17:09] is this worth addressing? If yes, how? [17:10] his comment, for reference: https://code.launchpad.net/~ahasenack/ubuntu/+source/base-files/+git/base-files/+merge/388835/comments/1022645 === ijohnson is now known as ijohnson|lunch === ijohnson|lunch is now known as ijohnson [20:04] Anyone? [21:54] ahasenack: I think it's impossible to migrate conffile between packages. [21:54] I'm doing it for the case of a modified conf file, when there is a backup [21:54] but a removed one... that's another story [21:54] ahasenack: at most you can check if the file doesn't exist in preinst and re-remove it. [21:55] ahasenack: backups and modified ones, also hard.