17:58:27 #startmeeting 17:58:27 Meeting started Mon Oct 12 17:58:27 2020 UTC. The chair is h01ger. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:58:27 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:58:47 #topic say hi / indicate your presence and review the agenda at https://pad.sfconservancy.org/p/reproducible-builds-meeting-agenda 17:59:09 Good evening all. (Chris Lamb) 17:59:11 * h01ger says^wtypes 'hello' while eating the last bits of his early dinner 17:59:15 hello! 17:59:27 * sangy is happy to see this reboot 17:59:35 Hello 17:59:40 Hello :) (Morten Linderud) 17:59:44 ohai! 18:00:11 * david-a-wheeler waves 18:00:37 * h01ger is happy to see many faces^wnicks again 18:01:01 Ah, david is even here! 18:01:08 * h01ger is about to remove 'add your topic here' from the agenda (there's still 'any other business') 18:01:10 hello (Paul Spooren) 18:01:29 Funny, I'm usually considered odd :-). 18:01:29 hello (Santiago Torres-Arias) 18:02:04 I hope everyone is doing well & healthy. 18:02:25 You too, David. 18:02:51 * h01ger hopes so too! 18:03:21 so i take it we have an agenda for todays meeting which should not last longer than 60min (but certainly may) 18:03:25 I'm doing well. I'm now at the Linux Foundation focusing on security of OSS. 18:03:40 nice! 18:04:10 so we've prediscussed that there are generally 2 kinds of meetings: 18:04:18 a.) meta meetings, which are 18:04:52 a.) meta meetings, which are non going into details of topics and rather collect summaries, stati and delegate work to a different date+time 18:05:26 b.) on-topic meetings, which might be boring for half the attendes if their topics are not discussed. 18:06:00 these are not the best summaries of those two kinds, but i hope you get the idea 18:06:22 in the prediscussion we concluced to want type a.) meetings here 18:06:51 #save 18:07:10 meta: gunnar: meetbot.debian.net has the log if you lost bits due to reconnects 18:07:14 and possibly plan on scheduling some topic-specific b. style meetings? 18:07:20 definitly 18:07:25 Meta-meetings make sense to me. However, I would urge where possible to try to do things *aychronously* (e.g., via email lists, issues, etc.). 18:07:42 david-a-wheeler: agreed 18:08:06 Hi 18:08:07 It's hard to get everyone to a meeting at the same time (synchronously), especially since people are around the world who have an interest in this topic. 18:08:09 Hello: Bernhard M. Wiedemann here. (was fighting with new internet home setup) 18:08:46 for those late or with lost connections: backlog at http://meetbot.debian.net/reproducible-builds/2020/ 18:08:47 There's probably an ideal middle ground somewhere where we get the benefit of seeing faces/nicknames and the positives of async conversations. 18:08:48 #save 18:09:31 these irc meetings are not ment to replace the lists or bug tracker :) 18:09:50 lamby: Agree. If nothing else, occasional synchronous meetings force people to finish up something, as well as help community interaction. 18:09:57 so it is a complementary channel 18:10:26 i sort of see irc meetings as helpful to kickstart a stalled conversation, or bring up new topics to frame them for a larger conversation through other channels 18:10:29 But I don't think sync meetings need to happen *often* for that to work. Also - the more warning time the better (so we can clear our calendars) 18:10:38 * h01ger mostly made this topic so you know what to expect. i'm not sure there is much to discuss now. i'd say: lets try and if this (the meta meetings) doesnt work, we can discuss again and change plans as needed 18:10:56 h01lger: Makes sense to me! 18:11:15 (meeting frequency is the next topic) 18:12:13 It also felt sensible to me to mention the type of meeting up front (even though there is little to discuss right now) so we can point back to this conversation if we get too detailed on some particular technical topic. 18:12:34 These meetings might also be a nice introducion for some people to start contributing. Shouldn't ignore that factor either :) 18:12:43 Foxboron: very much so! 18:13:01 also just to show the bots we have something to say as well 18:13:30 :) alright then... next topic 18:13:40 #topic 3. meeting frequency 18:14:12 my idea/suggestion is to try to go with bi-weekly, very 2nd monday, at 18 utc. aiming for an hour. 18:14:25 s#very 2nd monday#every 2nd monday# 18:14:28 wfm 18:14:37 sounds good 18:14:38 was monthly too far appart? 18:14:47 sangy: i think so, yes 18:15:04 Fair enough. I'm mostly concerned about having enough content and continuity 18:15:09 if we end up scheduling some "b." style meetings, that might get to be a bit much 18:15:14 * h01ger also knows any time will be bad for someone and is happy to do a new doodle poll for a different time say every 3 months 18:15:21 I think the reason we lost some of the momentum was that the last meetings were lacking stuff to talk about 18:15:26 * david-a-wheeler Googles, finds that's 2pm US Eastern Time right now. 18:15:29 sangy: yup 18:15:52 * h01ger googles "US eastern time" 18:16:10 18 UTC might be a bit inconvenient for the extreme ends of the timezones. Like Australia and the... east coast of the US? 18:16:29 But time can maybe be decided upon on a case-by-case? 18:16:32 * h01ger notes he drew in the next topic (time) into this (frequency) 18:16:35 Foxboron: (West coast, I think you mean?) 18:16:45 Foxboron: i'd say so yes. 18:16:47 (Yes, west coast :)) 18:16:52 I wonder if we can do bi-weekly and also alternate TZ's? 18:16:54 lamby: 18UTC isn't bad for me west coast ... 18:16:55 some projects do alternating times to better cover the globe 18:17:23 US West coast is -3 hours US East coast, so 11am US West Coast. That doesn't seem bad. Any time is going to be horrible for some TZ. 18:17:28 18utc work for me here 18:17:29 sangy: i'd avoid alternating TZ. its hard to remember a bi-weekly meeting already, if you add different times to it, more people will forget 18:17:32 * bmwiedemann wonders how many Asian/Australians do we not have here today 18:17:36 biweekly meetings sounds great 18:17:59 most people who are here now can make 18 utc work for them. else they wouldnt be here. hence my happiness to do dudle polls for times every 3 months 18:18:13 i would think the people who showed up for this meeting bias the results of 18UTC ... but this meeting time was selected after making a poll 18:18:14 18 UTC works for me, I'm just forwarding the concern as I recently had to consider it a bit :) 18:18:17 ack 18:18:18 Biweekly 18 UTC works for me right now. I can't promise to be at every meeting :-(. 18:18:24 #agreed we do type a.) meta meetings 18:18:36 #agreed biweekly is a good frequency 18:18:54 To clarify: Biweekly = alternate weeks 18:19:00 fortnightly 18:19:00 every two weeks 18:19:06 david-a-wheeler: noone expects you to, though many people will be happy when you show up 18:19:21 #agreed every 14 days is a good frequency 18:19:36 #topic 4. meeting time 18:19:45 h01ger: Very kind, thanks! 18:20:01 "meeting time" was mostly discussed in the previous topic 18:20:06 Oh. sorry :x 18:20:53 seems like the proposal was to go with 18UTC, though there were some ideas about having each meeting switching the time? 18:20:59 maybe we can also have this rule: if two or more people want a dudle poll for a new time, we do it 18:21:24 vagrantc: there was the idea of switching time each time but i really dont think it will work well. 18:21:30 yeah, I just threw the proposal out there, but I"m not very strongly opposed to single TZ 18:21:51 the TZ is UTC :) 18:21:51 I've just seen the format exist in other groups to allow for people who would not otherwise be able to join (existing and future members) 18:21:58 btw: in 2 weeks is DST change in EU and then, 19 UTC would be easier for me 18:21:59 sorry, s/tz/meeting time/ 18:22:18 /o\ DST! 18:22:19 bmwiedemann: same here :) 18:22:24 h01ger: Many groups have alternating times due to TZ, but they usually only do that once the group gets bigger. 18:22:36 david-a-wheeler: ic 18:23:01 that makes sense 18:23:08 i'm not strictly against alternating times.. 18:23:11 e.g., I think it's 2AM right now in China/Japan 18:23:17 i just know that regular times help me make a meeting 18:23:22 +1 18:24:08 can i say: _agreed we use 18 UTC for now and if two or more people want a dudle poll for a new time, we do it 18:24:09 ? 18:24:18 s#say#record# 18:24:21 h01ger: sounds reasonable 18:24:31 I think it's reasonable as well 18:24:36 sounds good 18:24:41 For now. I think I'll have a conflict once DST starts, but I guess I'll see. 18:24:49 * vagrantc watches jelle haggle 18:25:11 wfm 18:25:30 #agreed we use 18 UTC for now and if two or more people want a dudle poll for a new time, we do it 18:25:39 Hmm. Confirmed, I'll have a conflict once US DST starts. I may ask for 19 UTC then, but that's up to the group to decide if they want to change. 18:25:43 * bmwiedemann votes for having a dudle for after DST 18:25:55 * david-a-wheeler seconds bmwiedemann 18:25:56 DST changes at different places on a different date 18:26:05 yeah, that's what makes it fun 18:26:14 that said: "if two or more people want a dudle poll for a new time, we do it" 18:26:14 2020-10-25 in EU, 2020-11-01 in US 18:26:29 h01ger - I know, that's why I said "US". It *used* to be coordinated :-(. 18:27:11 We're here now. Can we just use 19 UTC starting November? 18:27:16 bmwiedemann: i heard the world is bigger and i wouldnt be surprised if there are different places in europe :) 18:27:17 well, it's within anyone's power to find another person and propose the doodle poll :) 18:27:19 anyhow 18:27:21 next topic 18:27:40 #topic 5. quick brainstorm of potential topic-specific irc sessions 18:27:54 i think topic 6 is/might be one ;) 18:28:01 and topic 8 too 18:28:32 so, the idea i had here was to have a no-discussion brainstorm of ideas for topic-specific sessions ... kind of like the in-person breakout sessions we've had at the summits 18:28:34 besides that i have "debian: distributing .buildinfo files for real" 18:28:49 any clarification needeD? 18:29:02 vagrantc: I think we had set some goals too on the summit which we where going to revise the next summit 18:29:29 and "debian: a new .deb format including .buildinfo files" (which is what i suggested in my dc20 talk and for which i still need to present a complete written down rationale, which i plan to do in the next 2 weeks) 18:29:52 Look at/discuss current published research being done in repro builds might be interesting. It's also relevant as we intend to collect some of it on the webpage. 18:29:59 ok, brainstorm on! 18:30:29 applying reproducible builds to achive diverse-double-compiling in the real world 18:30:38 brainstorm idea: at each meta meeting we nominate one topic for a type b.) meeting 7 days after that meta meeting 18:30:50 vagrantc: I like that one :-) 18:31:48 I want to use dettrace more in my r-b tools - especially the autoclassify. 18:31:51 brainstorm: strategize a vision for reproducible builds 1/3/5 years in the future 18:32:06 Idea: "Default Debian (or Ubuntu/Fedora), as delivered, is fully reproducible". Or maybe that should be broken down further? 18:32:33 david-a-wheeler: i want a pony. IOW: this needs to be broken down more 18:32:38 vagrantc: https://github.com/bmwiedemann/ddcpoc 18:32:38 #save 18:33:11 * vagrantc taunts people to come up with more ideas :) 18:33:35 I've lost the bubble on how close things are to that vision. Can someone break that vision down into the steps needed? 18:34:36 maybe discuss our set goals on the last summit https://reproducible-builds.org/files/ReproducibleSummit5EventDocumentation.html#__RefHeading___Toc14828_2303288670 18:34:37 david-a-wheeler: happy to do after the meeting :) 18:34:39 brainstorm: open office hours Q&A session to catch people up on missing pieces 18:35:02 david-a-wheeler: in openSUSE, there are roughly 95% of packages that can build reproducibly, but some of the workflows fail when we normalize mtimes. 18:36:09 brainstorm: rebuilder progress / pain points 18:36:16 bmwiedemann: david-a-wheeler: in debian there are also 95% of packages that can be build reproducibly, but that wasnt davis's question. he asked for "as delivered, is fully reproducible" (and i'm not refering to the 95 vs 100% but rather to "can vs is") 18:36:18 maybe we should close this topic for now, and revisit it? 18:36:18 bwiedemann: I imagine that they can be prioritized. I would prioritize the "default package set" first. 18:36:30 vagrantc: yes 18:36:47 david-a-wheeler: bmwiedemann: lets stay meta for here and now please. 18:37:07 vagrantc: do you have your brainstorm topic resolved? did you take notes or will you? 18:37:27 h01ger: i'll put together a summary from the meeting logs 18:37:30 brainstorm: I'd like to see a Debian and other OS versions of https://github.com/bmwiedemann/reproducibleopensuse/blob/devel/howtodebug 18:37:30 IOW: how do we get from this brainstorm collection to the next on topic meeting? 18:37:42 vagrantc: coolio 18:37:44 #commit vagrantc to summarize brainstorm session 18:37:47 okay, meta-level is to identify *specifically* what is needed to get the "minimum/default set of packages" reproducible on a common distro (or more than one) 18:37:54 #action: vagrant will provide a summary of this discussion 18:38:13 next topic then 18:38:15 h01ger: thanks :) 18:38:28 #topic https://ismypackagereproducibleyet.org/ wants feedback, and possible additional data sources 18:38:31 bmwiedemann: ^ 18:38:59 bmwiedemann: i love it, is there a way to do a regexp or glob search, since different distributions may use different package names? 18:39:12 does this compare the CI tests? 18:39:26 the links are not really readable for me (blue on black background) 18:39:49 * h01ger registers ismypackagereproducibleinpracticeyet.com and puts a big NO on it :/ 18:39:51 yes, CSS is improvable. it takes the reproducible.json published from tests.r-b.o 18:40:02 I really like ismypackagereproducibleyet.org. Maybe we can nag Fedora/RH friends to provide equivalent data? 18:40:06 maybe worth to point that out? 18:40:18 david-a-wheeler: I haven't seen a lot of movement from them on this for a while honestly. 18:40:29 david-a-wheeler: noone from fedora/RH is working on anything reproducible atm ttbomk 18:40:55 h01ger: That's too bad :-(. 18:40:57 h01ger: I have talked with someone who worked on it for Fedora but forgot his name 18:40:57 h01ger: Neal Gompa disagreed with me last time I said that :p there is *some* movement *somewhere* but nothing visible. 18:41:01 bmwiedemann: style is a bit hard to read 🙂 18:41:02 vagrantc: possible, but I tried to keep it simple for now, so only exact matches atm. 18:41:07 jelle: Neal? 18:41:10 Are those reproducible.json files all the same over multiple distros? 18:41:15 bmwiedemann: is there any specific data source you are missing (and which you know exists)? 18:41:15 I know Neal :-) 18:41:17 Foxboron: whoever was in #archlinux-projects 18:41:21 then it is Neal 18:41:21 That was Neal :) 18:41:49 we have 3 more topics and 15min left 18:41:54 h01ger: I was wondering what else is on tests.r-b.o ? FreeBSD? alpine? does openwrt have packages of e.g. bash? 18:41:57 I do have a request for https://ismypackagereproducibleyet.org/ - can it also include hyperlinks to the specific results? It must get the data from somewhere...! 18:42:07 good point! 18:42:11 bmwiedemann: i think https://repology.org/ tracks differing names ... now sure how hard it would be to integrate 18:42:19 bmwiedemann: ok, happy to discuss this directly after the meeting if that works for you 18:42:36 OK, let's close this topic. thanks for the great feedback. 18:43:06 where is the source? 18:43:29 or i can give the answer here: freebsd ports are not being tested. alpine tests are broken. openwrt packages are build, but only as package set (there is no data to be exported on t.r-b.o) 18:43:30 oh nevermind 18:43:34 aparcar[m]: it's the second or first link on the page 18:43:35 happy to go into more detail later 18:43:38 maybe we should link to ismypackage... from tests.r-b.org 18:43:39 next topic.. 18:43:51 #topic 7. updates on projects status 18:43:59 jele, what do you mean with this? 18:44:04 I can do an Arch update 18:44:05 jelle, sorry 18:44:22 I can do an OpenWrt update at the next meeting 18:44:39 * h01ger thinks distro/project updates are good, but maybe even better send to the list? 18:44:44 aparcar[m]: also cool 18:44:49 hmm we could 18:45:12 h01ger: I was bringing up the topic as this was the first time we had a meeting since summit :-) 18:45:17 jelle: if you have prepared something, do paste it now? 18:45:40 jelle: sure. this is our first irc meeting, so i'm not surprised we're still a bit bumpy 18:45:45 ah, I forgot to prepare something :-) 18:46:11 heh, ok then 18:46:30 (speaking of bumpy, I need to leave in ~10min :/) 18:46:32 #agreed we very much appreciate project status updates send to the list (or shared via irc, blogs or cnn) 18:46:39 ok :) 18:46:40 next topic then? 18:46:57 #topic 8. rb-format 18:46:58 I think it's best if status is posted to mailing list, then all can see. 18:47:07 aparcar[m]: i suppose this was you? :) 18:47:13 (rb-format) 18:47:30 yea the mail thread was active for a bit but then went to sleep again 18:47:38 hint: it helps to tag topics with a name 18:47:43 I'm wondering if there happened something in the meantime 18:47:44 * vagrantc tried to stir it up a bit 18:47:49 bmwiedemann: the hint is on the agenda now 18:48:02 but mostly heard crickets 18:48:16 aparcar[m]: not much, it got stalled there. someone should write a summary of the replies and move the topic forward 18:48:28 vagrantc: what does 'hearing crickets' mean? 18:48:30 yeah I wonder if ti just needs a bump... 18:48:30 I remember working on a database layout with sangy but we never got anything in-toto related integrated 18:48:57 aparcar[m]: I did some sketch and I believe I submitted a WIP PR. It's been so long, so sorry if I was blocking on my side 18:49:04 #info this topic got stalled on the list and needs someone to drive further 18:49:05 h01ger: basically no response ... you listen, and all you hear is the noise of various insects in the distance 18:49:19 vagrantc: ah, i see 18:49:33 (= Grillenzirpen) 18:49:53 I love aglutinating languages heh 18:50:07 aparcar[m]: anyway. I wonder if we can bump the thread and I can shake the dust off of our discussion and take it from there? 18:50:12 okay if people are still interested in this I'll try to write a bump 18:50:25 or maybe sangy because you'd be the leader in in-toto related things? 18:50:30 * h01ger would like to stay on topic and use simple languange everyone understands. i love discussing language IRL 18:51:04 #action sangy or aparcar[m] or both will try to revive the topic 18:51:09 aparcar[m]: it might require summarizing the goals again, too 18:51:14 aparcar[m]: yeah. I can definitely take a look at where things left off and also send a bump. I don't think both of us reviving the thread would hurt :) 18:51:16 next topic for now then.. 18:51:18 vagrantc: ack 18:51:28 sangy: ack 18:51:31 #topic 9. any other business 18:51:58 i have one, a question: do you feel you spent the last hour well? 18:51:59 can someone send a rb sticker to hawaii? my laptop is a bit bald 18:52:02 it is always a good idea to be clear on the goals of what you do 18:52:16 aparcar[m]: mail me your shippign address and i'll happily do 18:52:21 h01ger: yes 18:52:33 aparcar[m]: i could too 18:52:37 aparcar[m]: nice 18:52:41 dunno if international shipping is any different... 18:52:50 h01ger: yes I think the time is well spent as it show that multiple people are interested on working on the various topics :) 18:53:07 I lost some reproducible packages and want them back 18:53:17 I did a "state of repro builds" during the weekends Arch Conf. The presentation might not be super interesting if you are in the weeds, but the Q&A is fairly informative I think. 18:53:20 h01ger: I think, the SNR was good, though interweaved discussions can be hard to read in IRC. 18:53:23 i think we've stirred up some old topics and look forward to renewed progress :) 18:53:39 bmwiedemann: +1 18:53:41 * h01ger is a bit exhausted from juggling topics (so i'm a bit empty right now) but is happy we are having these meetings again, hoping to include new and old contributors with it 18:53:46 Foxboron: link? 18:53:56 If this speeds up progress, the meeting was worth it! 18:54:10 I liked this meeting and looking forward to the next one. 18:54:22 happy to see nicks i hadn't seen in a while (or ever) 18:54:29 aparcar[m]: https://streaming.media.ccc.de/archconf2020/relive/6308 :) 18:54:31 while it's hot 18:54:34 +1 on what vagrantc just said 18:54:47 thank you 18:55:28 Clearly many people are interested in this important topic, and that's a good thing. 18:55:31 alright then, i think we can preparing closing this meeting in 4 mins. so i'll say "thank you all very much for attending and contributing to this meeting!" now and will say endmeeting in 3min 18:55:46 Thanks so much! 18:55:50 * h01ger waves happily 18:55:54 Thank you very much h01ger :) 18:56:02 Thanks h01ger ! 18:56:03 * david-a-wheeler waves back! 18:56:11 Very happy to could attend for the first time :) 18:56:11 Indeed, thanks h01ger. 18:56:11 david-a-wheeler: had hoped at some point to get your comments on: https://reproducible-builds.org/news/2019/12/21/reproducible-bootstrap-of-mes-c-compiler/ 18:56:36 Thank you for organsing this h01ger. 18:56:40 david-a-wheeler: yes, you might be very interested in that blogpost :) 18:57:09 vagrantc: I mentioned that as part of the Q&A since someone asked if repro builds solved anything if dependencies was bad :p 18:57:16 dependencies bad/compromised 18:57:26 #topic next meeting: monday, october, 26th, 18 UTC on #reproducible-builds on irc.oftc.net 18:57:28 I'm off, bye 18:57:34 #agreed next meeting: monday, october, 26th, 18 UTC on #reproducible-builds on irc.oftc.net 18:57:36 Thanks all o/ 18:57:38 thank you! 18:57:45 thank you all! 18:57:51 Thanks! 18:58:10 And thanks for pointing me to the MeS post, I miseed that somehow! 18:58:18 #endmeeting