15:02:19 <pknirsch> #startmeeting Fedora Base Design Working Group (2014-07-18)
15:02:19 <zodbot> Meeting started Fri Jul 18 15:02:19 2014 UTC.  The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:19 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:27 <pknirsch> #meetingname  Fedora Base Design Working Group
15:02:27 <zodbot> The meeting name has been set to 'fedora_base_design_working_group'
15:02:31 <pknirsch> Now :)
15:02:37 <pknirsch> Hello and welcome everyone!
15:02:58 <pknirsch> On this gorgeous Friday afternoon (30'' C here in Stuttgart :)
15:03:19 <haraldh> <-
15:03:35 <haraldh> feels like 40° C inside here
15:03:39 <pingou> haraldh: +1
15:03:40 <pknirsch> #chair pknirsch dgilmore haraldh
15:03:40 <zodbot> Current chairs: dgilmore haraldh pknirsch
15:03:44 <pknirsch> hehe
15:03:47 <pknirsch> AC ftw!
15:03:54 * pknirsch loves the new office
15:03:58 <haraldh> :)
15:04:10 <haraldh> I remember the old one :)
15:04:10 <pknirsch> you should come visit us more often, haraldh !
15:04:13 <haraldh> true
15:04:15 <pknirsch> yea ;)
15:05:09 <pknirsch> And apologies for sending out the agenda so late, i was in marathon meetings the past 3 days at customers, so just didn't find time to send it our prior to today
15:05:10 <haraldh> msekletar is on PTO, IIRC
15:05:30 <dgilmore> hey
15:06:52 <haraldh> * vpavlin :No such nick/channel
15:06:52 <haraldh> * [vpavlin] End of WHOIS list.
15:08:21 <pknirsch> right
15:08:35 <pknirsch> ahm well, next week i'll be able to send it out earlier
15:08:36 <haraldh> seems to be in London, IIRC :)
15:08:40 <pknirsch> ah
15:08:48 <pknirsch> right, i think he mentioned that last week
15:09:00 <haraldh> together with msekletar and lnykryn
15:09:05 * pknirsch nods
15:09:29 <pknirsch> #chair jreznik
15:09:29 <zodbot> Current chairs: dgilmore haraldh jreznik pknirsch
15:10:01 <pknirsch> alright, nontheless we can quickly see if moben has anything new to report first :)
15:10:05 <pknirsch> #topic Status update builrequires cleanup
15:10:07 <pknirsch> bup
15:10:09 <jreznik> hey guys, sorry I'm late, I had medical check-up
15:10:23 <pknirsch> no worries jreznik, thanks for coming
15:10:35 <moben> Hi everyone!
15:10:48 <jreznik> pknirsch: actually two check-ups today :)
15:11:40 <jreznik> hi moben
15:12:04 <pknirsch> heya moben :)
15:12:06 <moben> I've mostly been working on some corner-cases where I have to do a content-aware comparison instead of relying on rpm's checksums to identify 'equal' builds
15:12:10 <pknirsch> jreznik: double doctor!
15:12:58 <moben> such as javadoc embedding timestamps in comments in html files, or typlelib files sometimes differing (not entirely sure why)
15:13:35 <haraldh> reminds me of my cpio patch for equal archives: http://lists.gnu.org/archive/html/bug-cpio/2014-07/threads.html
15:13:51 <moben> those obviously lead to cases where two builds even with the same buildrequires differ, so I got to handle them
15:15:04 <haraldh> I heard Suse did some heavy patching to ensure reproducible builds
15:15:24 <moben> haraldh: interesting, though so far I haven't run into any problems with inode numbers yet
15:15:37 <moben> haraldh: debian did some stuff there as well
15:16:02 <haraldh> moben, most likely rpmbuild does the archive and not cpio
15:16:26 <haraldh> would be nice to have a javadoc option to turn off timestamp generation
15:17:06 <haraldh> moben, what about compressors like xz?
15:17:49 <haraldh> at least gzip has the "-n" option and pgzip "-n -T"
15:17:57 <haraldh> to avoid timestamps
15:18:16 <tomprince> Another option would be to make the timestamp deterministic (for example, be the timestamp of the file from which the document was generated)
15:19:03 * pknirsch nods
15:19:05 <haraldh> tomprince, oh, then you will have to change the meaning of the header for compressors, e.g.
15:19:12 <moben> I'll need to look at those as well, but I guess I'll have to extract and compare anyway, because not every build is going to invoke them like that
15:19:31 <haraldh> moben, well, man pages are compressed
15:19:57 <haraldh> kernel modules
15:21:01 <haraldh> tomprince, but I agree in the javadoc case... take the newest timestamp of all source files
15:22:01 <moben> I think I'll try to get something rolling in the direction of reproducible builds
15:22:08 <haraldh> although, if only a comment was changed, the result is the same, so I would vote for removing the timestamp
15:22:29 <haraldh> moben, that would be absolutely awesome!!!
15:22:33 <moben> As I'm basically running into all the little problems on my way
15:22:58 <haraldh> as I said, you should probably talk to some suse guys
15:23:06 <haraldh> they did all that stuff already
15:23:36 <haraldh> upstream work would be nice, instead of workarounds
15:23:46 <haraldh> like suse did
15:23:46 <moben> haraldh: okay, I'll see if I can get in touch with someone there
15:23:59 <moben> always upstream first. I hate workarounds
15:24:08 <haraldh> yeah
15:24:25 <haraldh> in case of cpio I might have to wait half a year for a comment :-/
15:24:37 <haraldh> and another until its committed
15:25:11 <moben> That's probably the biggest complaint I personally have about the packaging guidelines: upstreaming patches is a SHOULD not a MUST :-/
15:25:11 * masta looks in
15:26:20 <moben> haraldh: also I'll see if I can get my script running on the list of buildrequire-heavy packages that you and pknirsch came up with, I guess I'll run into all kinds of those little quirks there, with that wide range of different projects :)
15:26:32 <haraldh> heh
15:28:07 <moben> on the bright side, I noticed that gcc's dep on libXtst is gone in rawhide. (seriously? gcc, libXtst? Sometimes I'm just amazed)
15:28:24 <haraldh> :-)
15:28:40 <moben> So, that
15:28:51 <moben> 's it from me for now, I guess
15:31:18 <haraldh> pknirsch, next topic?
15:31:30 <pknirsch> right, sorry
15:31:52 <pknirsch> #topic Talk with Vaclav Pavlin as candidate for WG
15:32:18 <pknirsch> Ok, so i totally forgot that vpavlin was away today, so unsure if it makes any sense for us to talk about him.
15:32:30 <haraldh> probably not
15:32:35 <pknirsch> mhm
15:32:54 <pknirsch> jreznik, do you know vpavlin or could tell us a bit about him from your perspective if you do?
15:33:05 <pknirsch> (or anyone else for that matter ;))
15:33:49 <haraldh> Well, he _was_ in my team...
15:33:56 <pknirsch> oh
15:34:11 <pknirsch> so what was your experience with him then haraldh ?
15:34:49 <haraldh> I think he took over most of notting's resort
15:35:18 <haraldh> so he would fit in of course from a distribution standpoint
15:35:28 <pknirsch> aye
15:35:40 <haraldh> he also has experiences with initscripts and systemd
15:35:40 <masta> cool
15:35:46 <pknirsch> #chair masta
15:35:46 <zodbot> Current chairs: dgilmore haraldh jreznik masta pknirsch
15:36:09 <haraldh> which comes handy to understand the inner workings
15:36:38 <haraldh> I don't know exactly though what he is working on right now...
15:37:06 <pknirsch> jup. and he's been very active with docker as well, which would be nice together with msekletar
15:37:08 <haraldh> so, this is why I would like him to speak for himself
15:37:14 <pknirsch> jup
15:37:21 <haraldh> ah docker
15:37:21 <haraldh> yes
15:37:52 <pknirsch> Ok, lets wait till next week to talk with him about it. But i'd love to get someone active into the WG in additionto msekletar and moben
15:38:06 <pknirsch> and my keyboard eats spaces again :/
15:38:22 <haraldh> the nice thing with docker is, that they want _small_ images
15:38:28 <pknirsch> aye
15:38:39 <haraldh> so it's in their best interest to have a minimal dependency base :)
15:38:45 <pknirsch> which aligns with your goal and moben's work
15:38:54 <haraldh> yep
15:39:04 <masta> yes!
15:39:16 <pknirsch> :)
15:39:16 <pknirsch> and masta's goal :)
15:39:18 <pknirsch> (and mine ;))
15:39:27 <haraldh> I guess nobody wants bloatware :)
15:39:33 <pknirsch> mhm
15:39:48 <pknirsch> ok, but lets talk with him next week when he's back
15:39:58 <haraldh> It's just those deps creeping in over time
15:40:09 <haraldh> unnoticed most of the time
15:40:14 <pknirsch> #info vpavlin in London today, will reschedule talk with him for next Friday
15:41:03 <pknirsch> Right, and thats why my goal is to get moben's work integrated into our release procedures somehow.
15:43:47 <pknirsch> ok, so lets open up the floor for other topics
15:43:51 <pknirsch> #topic Open Floor
15:44:01 <pknirsch> anyone got other topics to cover today?
15:48:56 <pknirsch> ok, i take that as a no :)
15:49:27 <pknirsch> with that, thanks everyone for joining today and have a wonderful weekend (with hopefully nice summer weather as we have here in Stuttgart) :)
15:49:32 <pknirsch> #endmeeting