[06:53] <dholbach> good morning
[12:02] <mitya57> dholbach: "/usr/share/doc/ubuntu-packaging-guide-html/singlehtml/index.html does not have styling" — looks like a bug in dpkg,
[12:03] <dholbach> oh yeah?
[12:03] <mitya57> my changes make singlehtml/_images and _links be links to
[12:03] <mitya57> ../{_static,_images}
[12:04] <mitya57> the result is, when you install the new package, they are empty directories
[12:04] <mitya57> purging and re-installing the package fixes that
[12:04] <dholbach> hummmm
[12:04] <dholbach> ^ anyone ever experienced anything like this?
[12:05] <mitya57> the best thing I can do is revert all the packaging changes
[12:05] <mitya57> they weren't really needed for the search to work
[12:06] <dholbach> if that's easily possible and makes things work again, that might be useful
[12:06] <dholbach> thanks also for the typo and code-styling fixes
[12:06] <dholbach> it looks much better now than before
[12:06] <Zhenech> dpkg link handling needs special care
[12:06] <Zhenech> if you migrate from non-links to links
[12:07] <mitya57> yes, that was the case
[12:09] <mitya57> Zhenech: any documentation for what "special care" is needed?
[12:14] <Zhenech> mitya57, short google didnt yield what I wanted and at work currently
[12:14] <Zhenech> i bet you get good results in #debian-devel over in oftc :)
[12:16] <mitya57> OK, will now upload my branches and ask that
[12:21] <mitya57> dholbach: the new mp is at http://pad.lv/mps/ubuntu-packaging-guide
[12:21] <dholbach> mitya57, great
[12:24] <cjwatson> mitya57: dpkg will never replace a symlink with a directory or vice versa; if you want to do that then you must either remove the old symlink/directory in the preinst or fix up the state to what you want in the postinst
[12:24] <cjwatson> This is an intentional safety feature
[12:27] <mitya57> cjwatson: thanks, hm...
[12:29] <mitya57> dholbach: are localized versions supposed to be in separate packages?
[12:29] <dholbach> mitya57, yes
[12:30] <dholbach> otherwise they will very likely get too big
[12:30] <mitya57> dholbach: and will they depend on the non-localized one?
[12:30] <dholbach> no - the html/pdf text is generated during the build
[12:31] <mitya57> I mean that if they do depend, we can have one _static & _sources and link anything else there
[12:31] <dholbach> yes
[12:31] <dholbach> maybe the -html* packages should depend on a new -common package
[12:31] <dholbach> getting this right is not as easy as I thought
[12:31] <dholbach> but still I think it's worth the effort
[12:32] <mitya57> ok, so I'll wait until the localized packages are ready, and then try to do that using a way suggested by cjwatson
[12:33] <dholbach> can you explain what you mean by "until the localized packages are ready"?
[12:33] <mitya57> #LANGS = $(basename $(notdir $(wildcard $(PODIR)/*.po)))
[12:33] <dholbach> ah yes of course
[12:34] <dholbach> let me try something
[12:34] <mitya57> :)
[12:34] <mitya57> ah
[12:34] <mitya57> there's one more problem with dh_sphinxdoc
[12:35] <mitya57> it won't work well when there are two or more sphinx projects built
[12:35] <mitya57> so let's not use it at all :)
[12:36] <dholbach> ok, if we manually enable the languages in LANGS the -es packages look like they work
[12:36] <dholbach> impressive work by the translators
[12:37] <mitya57> though, we can mimic its behaviour by linking our {jquery,underscore,doctools,searchtools}.js files to the ones in libjs-sphinxdoc package
[12:37] <dholbach> do you think you can file a bug for it?
[12:37] <mitya57> I think I can do it :)
[12:38] <mitya57> s/do/implement/
[12:39] <mitya57> dholbach: btw, should the -epub package really depend on jquery & underscore?
[12:39] <dholbach> mitya57, to be honest, I don't know
[12:39] <mitya57> does asomething know?
[12:42] <mitya57> hm, when trying to "make epub", I get 5 warnings
[12:42] <mitya57> "WARNING: unknown mimetype for _static/*.js, ignoring"
[12:43] <dholbach> hm
[12:44] <mitya57> ah, it seems that epub is something like an archive with html/js/... inside
[13:00] <mitya57> dholbach: lp:~mitya57/ubuntu-packaging-guide/use-libjs-sphinxdoc -- something like this
[13:00] <mitya57> I don't have time to test that now, maybe will do that in the evening and propose the merge
[13:01] <dholbach> ok, I'll have a look
[15:26] <mitya57> dholbach: I'm back with a question: debian/rules uses $DEBIANDIR, but that's not defined for me
[15:26] <mitya57> can I do s/$$DEBIANDIR/debian/ there?
[15:27] <dholbach> mitya57, I don't know - I didn't write all of the code
[15:27] <dholbach> and I will have to run for a call in a bit
[15:28] <mitya57> I will do that — I think it's safe
[15:28] <dholbach> mitya57, but I had noticed that you were missing a 'done' in your debian/rules snippet
[15:28] <mitya57> dholbach: https://bazaar.launchpad.net/~mitya57/ubuntu-packaging-guide/use-libjs-sphinxdoc/revision/184
[15:28] <mitya57> "add a closing done"
[15:29] <dholbach> perfect :)
[15:29] <dholbach> as I said - I'll take a look at it later on
[15:59] <jtaylor> universe is still open or?
[15:59] <Laney> yeah
[15:59] <Laney> ffe vanishingly unlikely thoughj
[19:20] <micahg> ajmitch: congrats on being reARBd
[19:25] <ajmitch> micahg: lucky me?
[19:25] <micahg> ajmitch: gives you an excuse why you're not uploading to the main archive :P
[19:25] <iulian> ajmitch: Congratulations!
[19:34] <ajmitch> micahg: yeah I'll run with that excuse
[19:35] <ScottK> That and we aren't bumping boost versions, so nothing to do.
[19:35] <ajmitch> sounds good to me, I'll head back to sleep
[19:36] <micahg> ScottK: oh, are we staying with boost 1.49 for R?
[19:38] <ScottK> I would assume so.
[19:38] <ScottK> Generally getting ahead of Debian is a bad idea.
[19:38] <ScottK> We should ship the newest that's packaged, but not make it the one we use in packages.
[19:41] <ajmitch> & then generally run into co-installability problems if some library needs a newer version of boost
[19:41] <ajmitch> it ends up a little bit of a mess, but that's to be expected
[19:43] <ScottK> Much better than trying to fix stacks of FTBFS due to API/ABI changes in the newer boost.
[19:43] <ScottK> xnox: ^^^ make sense to you?
[19:43] <ajmitch> especially if it's updated after feature freeze :)
[19:44] <ScottK> Sid has 1.50, so we should get that, but leave it in Universe.
[19:47] <micahg> !info boost1.50 quantal
[19:48] <ajmitch> quick! get it synced! :)
[19:48] <micahg> !info libboost1.50-dev quantal
[19:48] <ajmitch> ok, so it just doesn't show the source package
[19:49] <ScottK>  boost1.50 |   1.50.0-1 | quantal/universe | source
[19:49] <ScottK> So no change, really. (there's a -2 in Sid)
[19:49] <micahg> looks like we'll get libav 9 in R
[19:50] <ajmitch> how much breakage is there with that one?
[19:50] <micahg> lots of internal APIs becoming private