17:02:23 <mapreri> #startmeeting Reproducible Team Meeting - 2015-06-23
17:02:23 <MeetBot> Meeting started Tue Jun 23 17:02:23 2015 UTC.  The chair is mapreri. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:02:23 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:02:35 <mapreri> #chair deki
17:02:35 <MeetBot> Current chairs: deki mapreri
17:02:50 <mapreri> #link https://wiki.debian.org/ReproducibleBuilds/Meetings
17:03:02 <mapreri> #info agenda on https://wiki.debian.org/ReproducibleBuilds/Meetings just open that page
17:03:25 <mapreri> because -v it's funny here's the bullet points:
17:03:35 <mapreri> * discuss todays agenda
17:03:36 <mapreri> * go though last meetings summary and look for unactioned action items
17:03:36 <mapreri> * package/issue updates + r.a.d.o repo state
17:03:36 <mapreri> * rp.d.n updates+issues
17:03:36 <mapreri> * GSoC updates
17:03:37 <mapreri> * any other business
17:03:40 <mapreri> * announce next meeting
17:03:50 <mapreri> #topic discuss todays agenda
17:04:09 <deki> anyone has an additional topic?
17:04:11 <mapreri> Somethings we should change on the agenda? 2 mins
17:04:30 <mapreri> (/me is for very-flash and organized meetings ;))
17:04:39 <deki> +1
17:06:00 <mapreri> good.
17:06:05 <mapreri> #topic go though last meetings summary and look for unactioned action items
17:06:22 <mapreri> deki: did you do everything you had to do? ;)
17:06:28 <deki> mapreri: yes :)
17:06:31 <mapreri> s/ do?/?/
17:06:34 <mapreri> :*
17:07:00 <deki> one of the cloned bugs was even closed on the same day :)
17:07:13 <mapreri> cool
17:07:27 <mapreri> i don't see anything else on the previous minutes
17:07:38 <mapreri> was there anything left?
17:07:42 <deki> no
17:07:44 <mapreri> were* ?
17:07:59 <mapreri> fucking language
17:08:05 <mapreri> we all should just draw what we want
17:08:15 <mapreri> cool +1
17:08:28 <mapreri> #topic package/issue updates + r.a.d.o repo state
17:08:50 * mapreri is a bit unprepared, gimme a sec :)
17:09:11 <deki> debbindiff 23/24 was released yesterday/today with an awesome speedup by helmut. :)
17:09:20 <mapreri> in the meantime check if you see something to do/weird on the repo and/or patched packages
17:09:25 <mapreri> YAY
17:09:51 <mapreri> #action mapreri to reschedule package where dbd timeouts to see whether they are fixed by dbd 24
17:10:07 <mapreri> (tomorrow, one dbd 24 will be on jenkins)
17:10:34 <deki> repo state looks okay
17:11:07 <mapreri> i think we wanted to patch dh to export SOURCE_DATE_UTC or whatever pod2man started looking for?
17:11:24 <Lunar^> it's help2man, and that would be SOURCE_DATE_EPOCH
17:11:30 <mapreri> right
17:11:39 <mapreri> my memory → /dev/null
17:12:41 <Lunar^> I'm going to focus on debbindiff regarding coding. I hope someone beats me to this
17:13:23 <mapreri> i'm not sure where/how to do that, elso because afaik that would be the first env var exported by dh, so i can't just copy.. (and also i know nothing about perl)
17:13:39 <mapreri> volunteers?
17:13:43 <mapreri> oh, jumapico! o/
17:14:04 <mapreri> we just found a volunteer, right :P
17:14:07 <jumapico> sorry, volunteers for what?
17:14:12 <mapreri> eheh
17:14:13 <jumapico> :D
17:14:14 <Lunar^> only when building n?
17:14:16 <Lunar^> no
17:14:17 <Lunar^> ?
17:15:07 <mapreri> Lunar^: erm, can't understand
17:15:50 <nthykier> mapreri: for prototyping, you want set_buildflags in Dh_Lib
17:15:53 <nthykier> .pm
17:16:19 <mapreri> nthykier: cool
17:16:22 <mapreri> nthykier: thanks!
17:16:39 <deki> #info for prototyping SOURCE_DATE_EPOCH in debhelper, you want set_buildflags in Dh_Lib.pm
17:16:51 <nthykier> That sets flags from dpkg-buildflags under compat 9+
17:17:09 <mapreri> nthykier: is there a place in dh that knows the changelog date?
17:17:21 <nthykier> Probably not
17:17:35 <mapreri> sad
17:17:41 <helmut> Lunar^, deki: if I make the debbindiff timeouts fatal in rebootstrap, will you fix the fallout? ;-)
17:18:16 <nthykier> but you can make it know :D
17:18:21 <Lunar^> helmut: probably not now
17:18:23 <deki> helmut: did you have anything that still times out?
17:18:28 <mapreri> #info dh needs to learn the last changelog date somehow
17:18:29 <helmut> deki: cannot tell yet.
17:18:30 <AGWA> I can work on exporting this var - dh_strip_nondeterminism is already grabbing the changelog date so I should be able to adapt code
17:18:44 <mapreri> AGWA: maybe move some code to dh?
17:18:52 <AGWA> yes, that would be a good idea
17:18:55 <mapreri> alright
17:19:16 <mapreri> #action AGWA to investigate/work on getting dh export the SOURCE_DATE_EPOCH
17:19:19 <mapreri> cool!
17:19:28 <mapreri> anything else on this topic?
17:19:28 <nthykier> and TZ?
17:19:31 <h01ger> hi. sorry for being late & thanks for starting without me! (no irony! :)
17:19:36 <nthykier> might as well, while you are at it
17:19:43 <mapreri> h01ger: <3
17:19:44 <AGWA> will do
17:19:53 <deki> and also LC_ALL? :D
17:19:55 <h01ger> #save please
17:19:59 <mapreri> #save
17:20:23 <AGWA> any other vars? ;-)
17:20:24 <mapreri> here it is: http://meetbot.debian.net/debian-reproducible/2015/debian-reproducible.2015-06-23-17.02.html
17:20:47 <h01ger> SOURCE_DATE_ISO8xxx?
17:20:50 <mapreri> well, i don't think we really really want LC_ALL in dh. I'd rather discuss it a little more...
17:21:06 <jumapico> BUILD_DATE, BUILD_TIME and BUILD_TIMESTAMP? (with the format of gcc __DATE__, __TIME__, and __TIMESTAMP__)
17:21:14 <h01ger> TZ should also be set to utc by dh
17:22:11 <h01ger> "debbindiff 23/24 was released yesterday/today with an awesome speedup" - that wasnt clear to me from reading debian/changelog
17:22:12 <mapreri> AGWA: export whatever you think we can always drop stuff later :D
17:22:42 <jumapico> +1
17:22:47 <AGWA> mapreri: OK
17:22:50 <deki> h01ger: "* Speed up diff collection."  :)
17:22:57 <mapreri> even because if you use that function it'll work only for compat > 8
17:23:06 <mapreri> well, you can always patch it out
17:23:07 <mapreri> meh
17:23:10 <mapreri> just do something :)
17:23:18 <mapreri> let's move on
17:23:28 <mapreri> #topic rp.d.n updates+issues
17:23:31 <mapreri> status:
17:23:44 <h01ger> deki: ah, i totally missed the 23 entries
17:23:58 <mapreri> i did a bunch of stuff, but i want h01ger to be available to do a long deploy. full stop.
17:24:09 <deki> h01ger: yes, had to do 24 unfortunately because i missed something stupid...
17:24:15 <mapreri> so everything that not a typo will wait some weeks (1/2?)
17:24:42 <mapreri> + i think akira wanted to do something?
17:24:52 <h01ger> yeah, i will be back "regularily online" in a bit more than a week..
17:25:16 <mapreri> in one week i'll have physics+analysis, so meh
17:25:19 <mapreri> and then algebra
17:25:24 <mapreri> :|
17:25:28 <mapreri> whatever
17:25:29 <h01ger> (in the meantime please contact me via email if you need me. feel also free to ping me via irc to tell me i have urgent mail, but please use mail.)
17:25:59 <Lunar^> mapreri, AGWA: I think we should stick to one variable for SOURCE_DATE_*
17:26:04 <mapreri> for reference
17:26:06 <mapreri> [10:30:50] <akira> I was wondering if it would be useful to have a different symbol or maybe another color of the # to indicate pending bugs?
17:26:07 <akira> yes I wanted to create a symbol for pending bugs
17:26:11 <mapreri> :)
17:26:24 * h01ger has fun with a outside network with 50 wireless links and similar amount of APs etc pp :)
17:26:25 <mapreri> h01ger: could you put that in TODO?
17:26:34 <h01ger> mapreri: what?
17:26:40 <mapreri> ↑ 2 line up
17:27:06 <h01ger> "create a symbol for pending bug"?
17:27:13 <mapreri> or color
17:27:41 <mapreri> if you don't have something i'd move on
17:28:18 <mapreri> #info akira would like a symbol or a color to distinguish pending bugs in rb.d.n
17:28:23 <mapreri> #topic GSoC updates
17:28:33 <mapreri> let's go :)
17:28:44 <akira> atm I am working on two toolchain problems
17:29:14 <akira> one is with tex4ht which affects 5 packages
17:29:23 <akira> I have an idea where to make the fix
17:29:27 <mapreri> ( Dhole: ping since i've not read something by you yet)
17:29:42 <mapreri> akira: issue name?
17:29:49 <akira> and the other is doxygen latex or texline-bin
17:29:56 <h01ger> Dhole told me:
17:30:04 <h01ger> <Dhole> I will not be home this afternoon (we have a festivity inn the evening today in Spain)I
17:30:05 <h01ger> <Dhole> regarding the GSoC updates for the meeting, last week wass my last week of exams and presentations from uni (it was pretty
17:30:05 <h01ger> hectic) so I couldn't work much on the project
17:30:07 <akira> well there is no issue for tex4ht
17:30:15 <mapreri> nice doxygen latex i'm annoyed a bit by it :)
17:30:20 <mapreri> akira: create it?
17:30:33 <mapreri> h01ger: ok, thanks
17:30:38 <akira> yup I will create it too
17:30:46 <mapreri> thanks
17:31:12 <mapreri> akira: question/something you got for us?
17:31:15 <mapreri> (
17:31:29 <mapreri> (unwanted bracket, urg!)
17:31:41 <akira> yes, can the debbindiff output on rp.d.n be extended?
17:32:00 <akira> I have very long debbindiff and I cannot see some problems
17:32:09 <mapreri> ah, the limits
17:32:14 <mapreri> deki: ↑
17:32:16 <mapreri> Lunar^: ↑
17:32:17 <akira> without rebuilding it myself which takes ages
17:32:25 <mapreri> oh, the timeout!
17:32:31 <mapreri> (or the diff limit?)
17:32:39 <deki> hm, the reason it is cut off is not because of timeout
17:32:50 <h01ger> akira: provide patches? ;-)
17:32:57 <deki> yes, the size is just limited to not get huge html files
17:32:58 <mapreri> akira: btw if the problem is the rebuilding you can reschedule stuff with -a (--artifacts) and use that to diff
17:33:49 <mapreri> (we're out the 30 mins limit, note)
17:34:10 <akira> h01ger: I could extend the limit but I did not want to do that without discussing it first
17:34:35 <mapreri> deki: btw, we got hundred-MB rbuild, i guess a something a bit bigger .dbd.html could be acceptable?
17:34:50 <h01ger> akira: please do!
17:35:05 <deki> mapreri: hm yes. but the dbds are looked at by humans in the browser. the rbuild normally not
17:35:05 <akira> h01ger: okay
17:35:18 <Faux> Depends if you're diagnosing ftbfses or not.
17:35:20 <h01ger> (it only has an effect for very few packages, so i dont think that will be an issue)
17:35:21 <mapreri> akira: yes, just not something too exaggerated, but i think we can extend it a bit.
17:35:45 <mapreri> akira: + remember you don't need to build stuff at home if you don't have to :)
17:35:53 <mapreri> akira: anything else? ^^
17:36:14 <akira> would 4MB be too big?
17:36:29 <mapreri> nah, i think it ok. how it compares with the current one=
17:36:30 <mapreri> ?
17:36:37 <mapreri> it's*
17:36:44 <akira> I dont know what the current one is
17:36:54 <akira> because I dont know where to change it
17:37:27 <mapreri> difference.py, i guess
17:37:32 <mapreri> nah
17:37:59 <mapreri> deki: tell him/her (still to learn, sorry) ↑ :)
17:38:08 <deki> i'm looking
17:38:20 <h01ger> shall we move on with the meeting?
17:38:31 <akira> sure I dont have other questions
17:38:39 <mapreri> ok
17:38:40 <deki> i think it's presenters/html.py
17:38:45 <mapreri> #topic any other business
17:38:59 <mapreri> deki: the MAX_LINE_SIZE?
17:39:03 <deki> DEFAULT_MAX_PAGE_SIZE = 2000 * 2 ** 10
17:39:05 <mapreri> so, something else?
17:39:14 <josch> DEFAULT_MAX_PAGE_SIZE
17:39:19 <josch> ah damn one second too slow XD
17:39:24 <mapreri> :P
17:39:45 <josch> deki: thanks i can show it to akira now
17:39:52 <h01ger> are there any other topics?
17:39:53 <deki> josch: good :)
17:40:00 <mapreri> i think we figured there was nothing else outside the main track
17:40:07 <Faux> Can we reschedule all the things that have failed due to disc space?  I am happy to make the list.
17:40:12 <mapreri> Faux: sure
17:40:28 <Faux> Also, does anyone have an opinon on whether to usertag ftbfs bugs that are independently findable?
17:40:32 <mapreri> Faux: i though i collect most of them but if you find one just reschedule or make up a huge list for me
17:40:35 <Faux> i.e. don't just happen on jenkins?
17:41:37 <h01ger> Faux: i'd mark the others
17:41:48 <mapreri> Faux: i believe it's for everything that ftbfs (at least, that's how i behaved without getting too many complaints)
17:41:51 <Faux> mapreri: https://paste.debian.net/256924/ I think.
17:41:58 <mapreri> wow
17:42:11 <mapreri> Faux: all unstable?
17:42:14 <Faux> Yes.
17:42:34 <Faux> Well, I doubt other people care about our usertags on their bugs. :)
17:42:53 <h01ger> any other topic?
17:43:08 <Faux> I'm good.
17:43:20 <jumapico> a question (maybe offtopic) why remove __DATE__/__TIME__ macros instead replace by others that can be set in the command line?
17:43:41 <deki> because they make no sense
17:44:10 <jumapico> mmm, maybe for debian builds, but for a custom build?
17:44:36 <mapreri> still, i don't thing it usually matters
17:45:15 <mapreri> but i'm for putting them inside #ifdef clauses or something like that if that makes sense
17:45:28 <mapreri> Faux: ah, thanks for the list :)
17:45:30 <h01ger> lets close the metting?
17:45:40 <mapreri> #topic announce next meeting
17:45:51 <mapreri> so?
17:46:06 <mapreri> in two weeks?
17:46:11 <Faux> Sure.
17:46:12 <h01ger> yup
17:46:31 <mapreri> 2015-07-07?
17:46:46 <Faux> Ooh, that's actually excellent for me personally.
17:47:11 <h01ger> mapreri: yes
17:47:13 <mapreri> #info next meeting on 20150707T170000Z
17:47:20 <mapreri> go on, parse that!
17:47:21 <deki> :D
17:47:22 <mapreri> :P
17:47:26 <h01ger> :))
17:47:32 <h01ger> in epoch?
17:48:05 <mapreri> #info next meeting on 1436288400
17:48:09 <mapreri> :D
17:48:18 <mapreri> who'll send an email?
17:48:21 <h01ger> pffft^w*g*
17:49:04 <mapreri> ok, i'll do
17:49:13 <mapreri> #action mapreri to announce the next meeting on the ML
17:49:34 <mapreri> #topic final greetings
17:49:40 <mapreri> o/ everybody!
17:49:44 <deki> lol
17:49:47 <mapreri> nice meeting :)
17:50:03 <mapreri> #endmeeting