/srv/irclogs.ubuntu.com/2012/03/15/#ubuntu-arb.txt

mhall119wendar: ajmitch: stgraber: ping16:13
mhall119Someone on G+ is asking for some clarification about what his package should put in /opt/16:14
mhall119https://plus.google.com/108728447943307421833/posts16:14
mhall119can one of you give him an answer, of give me the answer and I'll give it to him16:15
mhall119http://paste.ubuntu.com/884992/ is you can't see that16:15
stgrabermhall119: in most cases, it's simply everything except the .desktop file16:16
stgrabermhall119: we also now allow /usr/share/doc/<packagename> but it's debhelper's default anyway so I don't think I ever saw someone actually put them in /opt (we always had to do it for them)16:17
mhall119stgraber: so the binary goes in /opt/ too?16:40
mhall119and the .desktop just points to it there?16:41
stgrabermhall119: yes16:43
mhall119stgraber: how about icons? Do they all go into /opt/ too, and if so does GTK know to find them there if you reference them by name instead of path?16:44
stgrabermhall119: "everything" :)16:44
mhall119ok16:44
mhall119thanks, I'll let him know16:44
stgrabermhall119: you need to put the path in the .desktop for it to work16:44
mhall119for Exec only, or Exec and Icon?16:45
stgraberboth16:46
stgrabergtk doesn't know to look in /opt and we don't really want it to either16:46
stgraberthe whole idea behind using /opt is to avoid having any of them automatically found/loaded by the OS unless you're calling them by their path16:47
mhall119ok16:47
wendarmhall119: there are more details at https://wiki.ubuntu.com/AppReviewBoard/Review/Guidelines19:42
wendarI forgot to ask about chairs for the next meeting at the end of last month's meeting19:43
wendar(I've added it back into the end of the agenda, so I'll remember in the future)19:43
wendarAnyone want to volunteer?19:44
* ajmitch checked next meeting date19:44
ajmitchwe're keeping it at the same UTC time?19:44
wendaryeah, same time19:44
ajmitchok, I can do it then19:44
wendarcool, thanks!19:44
ajmitchnext month I'll be back to UTC+1219:44
wendarlove those international daylight saving time shifts :)19:45
ajmitchyeah, it's just great :)19:46
wendarI recovered some lost action items19:46
ajmitchoh, what'd we miss?19:46
wendarapparently our "emergency" meeting in January wiped out several pending action items19:46
wendarI had one to ask about the deprecation cycle for python-support19:46
ajmitchok, I did follow up with dpitkin at least :)19:46
wendar(I've just emailed doko and barry to check in on that)19:47
wendarlfaraone had one about python-support in Quickly19:47
wendaractually, in python-distutils-extra, which Quickly uses19:48
wendarhttps://bugs.launchpad.net/python-distutils-extra/+bug/89458219:48
ajmitchI asked about #927588, #915902, #926032 to see if he'd prioritise getting those 3 fixed19:48
wendaryou had one about asking (someone, I'm not sure who) about ARB packages depending on backports19:48
wendarbut, I'm  pretty sure the answer to that one is "no"19:49
wendarat least, if it came to a vote, I'd be against ARB packages depending on backports19:49
ajmitchright, I asked broder & michag about the state of that, it's ok for runtime dependencies, but there's a LP bug that blocks build-dependencies19:49
wendarah, okay19:50
wendarI remember, it was a question of whether it was even possible for ARB packages to depend on backports19:50
wendarshould we keep a note of that somewhere?19:50
ajmitchright, but we mostly wanted it for the case of new libraries being added via backports, iirc19:50
wendaryup19:51
ajmitchhttps://bugs.launchpad.net/launchpad/+bug/888665 is the LP bug, it's due to it using an ancient forked sbuild19:51
wendarso, to say if it's safe, and non-disruptive, we could have an ARB depend on a backported library19:51
wendarif it's a new library, and not an updated version of an existing library in the old release19:52
ajmitchit'd just be hard to build against that library19:52
wendarso, at this point we'd be stuck with runtime python libraries19:52
wendarand no build-time C, etc libraries19:53
ajmitchactually a new library would likely work19:53
ajmitchsince the PPA can have backports dependencies turned on, and the backports version would be the only version to satisfy the requirement, it ought to work :)19:53
wendari.e. it's hard to give a one-size-fits all answer19:54
ajmitchyeah19:54
wendarHow about this addition to the ARB guidelines: right after "If your app depends on external libraries, please make sure that your app runs on the current versions shipped in Ubuntu."19:54
ajmitchwe do currently have "Applications must be able to be built with tools & libraries in the Ubuntu archive. Apps may bundle additional libraries they depend on, but may not include new versions of already packaged libraries."19:56
wendarwe add "(We're open to considering dependencies on backported libraries, on a case-by-case basis, only if the backport is a new library and not an updated version of an existing library.)19:56
ajmitchok19:56
wendarthat leaves it open for the devs to either bundle the library or backport it19:57
wendarwith the same effect, they can only depend on existing libraries, or a new library19:57
wendarbut not updated versions of existing libraries19:57
ajmitchthat's fair - no depending on a new version of Qt & boost :)19:58
* wendar shudders ;)19:58
ajmitchit'd be nice to get the LP bug fixed, but it's not a simple fix that just anyone can pick up & do, I suspect19:58
ajmitchwendar: so I didn't really get round to doing any ARB work last night (wasn't feeling great), but I've pushed the icon change I did yesterday to the staging PPA & my branch, if you have time to take a look20:11
ajmitchonce we vote on it I'll need to know the right buttons to push in myapps to get it published :)20:12
wendarajmitch: sure, I'll take a look20:13
ajmitchthanks20:13
wendarajmitch: in override_dh_gencontrol, you do need to keep that cp line20:17
wendarajmitch: it's obtuse, but what that does is copy the image *outside* the package directory20:17
wendarajmitch: so it can be added to the package metadata20:17
wendar(stgraber had to fix one of my earlier packages)20:18
ajmitchright, that is a bit special20:18
ajmitchbut this is why I'm checking with you first, thanks :)20:18
ajmitchit does mean I have to bump the ppa version again, which is a bit of a pain20:19
wendaryou can use $CURDIR20:19
wendarsee http://bazaar.launchpad.net/~allison/+junk/extras-framingham/revision/1320:20
ajmitchin which case I ought to put it in the debian/directory as well20:21
wendarI like the way you've used patches for all changes outside the debian/ directory20:26
wendarin the debian/changelog, merge his changes and yours into one changelog entry (since there's only one Extras release)20:27
wendaryou can note who did what with [ Author Name ]20:27
wendarand a quick scan of postinst and postrm, I wonder if they could be removed?20:28
wendarwe don't technically allow custom maintainer scripts, but this looks like standard desktop stuff20:29
wendarprobably automatically generated20:29
ajmitchthis is what I was checking with stgraber - they're boilerplate postinst/postrm20:29
wendarwith debhelper 8, a lot of stuff that used to be in generated maintainer scripts is now handled automatically behind the scenes20:31
ajmitchyeah, I was debating whether they'd be needed at all due to the non-standard directories20:32
wendarnot sure that's the case with these specific pieces, but seems worth checking20:32
wendarah, yeah, I noticed one of the commands was acting on /usr/share/icons/gnome20:32
wendarwhich we don't modify20:33
wendarso, that one could certainly go20:33
wendarthe mime database is modifying /usr/share/mime, which we don't modify20:33
ajmitchright, stripped out all but #DEBHELPER#20:35
wendarlet's see, update-desktop-database modifies the cache of MIME types handled by desktop files, so that can go too20:35
wendarcool, yeah, makes sense20:35
wendarajmitch: overall, it looks great. I'd +1 it.20:37
ajmitchgood, it's nearly all cielak's work, so thank him for a good submission :)20:37
wendarthanks, cielak!20:37
cielakno problem, did my best to get it as good as I could ;)20:38
ajmitchcielak: it does make our lives easier, I'll put this up for vote & we should be able to get it uploaded :)20:48
cielaknice! thanks for handling that :)20:48
* ajmitch could really do with some more RAM, running a couple of VMs doesn't help when firefox & chrome are both open20:51
wendarajmitch: yeah, both firefox and chrome are total memory hogs, very annoying21:14
ajmitchespecially the way I use them, where I often can't see the icons on the chrome tabs because I have so many open :)21:22
ajmitchwendar: just checking, you marked the next meeting as the 23rd rather than the 30th?22:08
wendarajmitch: well, that's what Ubuntu Fridge says22:09
wendarajmitch: but it could be wrong22:09
ajmitchthat seems a bit broken, it's meant to be the last friday of the month22:09
wendarajmitch: is it supposed to be... ah yea22:09
ajmitchwho sets up the fridge calendar entry?22:09
wendarFridge won't let me do anything more complicated than "3rd Friday"22:09
ajmitchgreat :)22:09
wendaranyone who has access can edit it22:09
ajmitchit doesn't allow 5th friday?22:10
wendargo ahead and update the wiki page22:10
wendarI suspect it would just drop off the calendar when there is no 5th friday22:10
wendarwhat we want is the -1th friday :)22:10
wendarthe 30th actually works better for me this month, I won't have to rush out after the meeting to drive to Seattle22:11
* ajmitch can see where to add an event, but it looks to only be editable by its creator, sorry :)22:14
wendarajmitch: I think I actually can't edit it anymore, I only had admin priviledges on my canonical google account, which is gone now22:39
wendarajmitch: but, I'll find someone else to edit it22:39
ajmitchnow that is a pain22:39
ajmitchyour membership application is coming up in a meeting today, isn't it?22:40
wendarah, yeah22:40
wendarif I get membership I'll see about getting  calendar editing access again22:40
wendaror, you or stgraber could get it22:41
* ajmitch shall try & be around at meeting time22:47
ajmitchstgraber: thanks for looking at harmonySEQ - the postinst/postrm aren't empty unless debhelper is not putting anything in them - I thought that I'd leave in the #DEBHELPER# tags for now23:45
wendarajmitch: it's probably safe to remove them entirely23:47
wendarajmitch: I've been erring on the side of "rip it out" for ARB apps, just on the general principle of "they're supposed to be dumb and simple"23:47
ajmitchfair enough23:48
ajmitchI don't think there are any dh calls in the package at this time which put anything in postinst23:48
* ajmitch shall see if he can get it to ~ppa99 before it even makes it into extras23:52
ajmitchI see you've got a couple of other testimonials from some well-known ubuntu contributors now23:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!