14:59:18 <rclobus> #startmeeting
14:59:18 <MeetBot> Meeting started Tue Jun 28 14:59:18 2022 UTC.  The chair is rclobus. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:18 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:03 <rclobus> #topic Welcome to the meeting. Who is online today? Please introduce yourself.
15:00:46 * rclobus is Roland Clobus, taking over the role of host for the meeting today. Otherwise, working on reproducible live ISO images.
15:01:26 * vagrantc is Vagrant Cascadian working on reproducible builds in debian and guix
15:02:24 <rclobus> Let's see who else says something, until 17:05, then the next topic will be activated.
15:04:16 <rclobus> Ok, it's 17:05
15:04:36 <rclobus> #topic Who wants to report something?
15:05:31 <rclobus> Ariadne: jelle: fepitre: h01ger: rclobus: obfusk: bmwiedemann: aparcar: mapreri: Do you have anything to report today?
15:05:52 <vagrantc> i mgiht do a little something about guix
15:06:10 <rclobus> vagrantc: Ok.
15:06:18 <fepitre> Hello, no at all.
15:06:29 <rclobus> I'll have a short overview of live images.
15:07:01 <rclobus> Let's be creative with the agenda for today :-)
15:07:07 <rclobus> #topic guix
15:07:15 <rclobus> vagrantc: The stage is yours
15:08:17 <vagrantc> so i did an alaysis of guix, some of the links are in the draft reproducible builds report for june ... it's at around 85% reproducible on x86_64 ... filed a few bugs and pushed a few patches
15:09:01 <vagrantc> i would've expected better numbers, given that guix normalizes the build environment in many ways, but they have a large number of ecl (embedded common lisp) packages for which there is an upstream issue and proof-of-concept patch
15:09:14 <vagrantc> roughly half of the known issues are ecl related
15:10:05 <vagrantc> it's been really fun working with something that has reproducible builds built into it's design ... e.g. guix challenge, guix build --rounds=N, etc.
15:10:26 <vagrantc> ...
15:11:21 * vagrantc offers the mic to the next speaker
15:11:32 <rclobus> #topic live images
15:12:05 <rclobus> I just managed to sqeeze out a summary of my activities for the last 2 months
15:12:14 <rclobus> #info https://lists.debian.org/debian-live/2022/06/msg00041.html
15:12:50 <rclobus> In short: all images are reproducible, only Cinnamon requires a patch.
15:13:24 <rclobus> Functionality testing is ongoing, automatic image publication is ongoing.
15:13:50 <rclobus> "All that's left, is just a matter of distribution of time" (which is hard)
15:14:42 <rclobus> That's the short summary from my side.
15:15:12 <rclobus> #topic Any Other Business
15:16:08 <rclobus> In #debian-reproducible, some days ago josch wrote about a CMake-related non-reproducible issue.
15:16:45 <rclobus> It is a known issue, but do we have an automated way to mark packages if the reproducibility fails with a specific pattern?
15:17:32 <vagrantc> there is a script that semi-automates stuff based on diffoscope output, but it is not automatically run
15:17:56 <vagrantc> reproducible-notes.git/bin/auto-classify
15:18:08 <rclobus> That could help. I'll take a look at it later.
15:18:17 <vagrantc> i've had mixed results with it
15:18:47 <vagrantc> i used to get it to work with reprotest ... just needed a few tweaks, and when i went to try and "upstream" those tweaks, it just stopped working for me entirely
15:19:27 <vagrantc> i think lamby wrote it
15:20:15 <rclobus> There is a specific pattern in the diffoscope output that will signal which packages could be affected. Thanks for the pointer.
15:20:28 <rclobus> Other AOB topics?
15:21:28 <rclobus> Ok. That's it for today. A very short meeting.
15:21:51 <rclobus> #info next meeting: 2022-07-26T15:00Z
15:21:56 <rclobus> #endmeeting