=== bkerensa is now known as hypodermic
=== hypodermic is now known as bkerensa
[04:54] <mlinscott> •sanbar• heya!
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
[13:20] <KNRO> How do you make packages from outside universe available to packages being built by launchpad build system?
[14:00] <geser> you mean for use in your PPA?
[14:10] <KNRO> geser: yes?
[14:10] <KNRO> geser: yes for PPA
=== yofel_ is now known as yofel
[14:24] <geser> KNRO: hmm, good question. You might need to ask that in #launchpad.
[14:26] <geser> which package do you need for building from multiverse?
[14:27] <KNRO> geser: my own packages in my ppa that I just built
[14:28] <geser> ah, that should be automatically picked up
[14:28] <KNRO> I have libqsi-6.0.3 that is already built, and I have indi-qsi which depends on libqsi-dev which is provided by the previous package
[14:28] <geser> have you a link to the failed build at hand?
[14:28] <KNRO> geser: sure , hold on
[14:29] <KNRO> https://launchpadlibrarian.net/88715108/buildlog.txt.gz
[14:31] <KNRO>  pbuilder-satisfydepends-dummy: Depends: libqsi-dev which is a virtual package.
[14:39] <geser> KNRO: libqsi - 6.0.3-0~3~oneiric1 seems to be the first one which build successfully and got published 1 hour ago. So it might be bad timing that it wasn't published yet when the other build got tried (it was about something more than 1 hour ago)
[14:39] <KNRO> so you think it's a timing issue?
[14:39] <geser> yes, retry it
[14:39] <KNRO> how about Depends: libcfitsio3-dev but it is not going to be installed. ?
[14:40] <geser> let me check
[14:40] <KNRO> this is another package...
[14:41] <KNRO> however, this is an official package in MAIN... so I'm not sure what the problem is there
[14:43] <geser> me currently neither
[14:46] <geser> give it retry, perhaps a fallout from the previous error
[14:52] <KNRO> hmmmmm ok
[15:45] <jtaylor> anyone know how to get ports.ubuntu.com merged with the regular archive with apt-cacher-ng? just adding it to the mirror list does not work :(
=== Guest76320 is now known as JackyAlcine
=== JackyAlcine is now known as Guest39737
[17:16] <micahg> jtaylor: merged?  are you trying to get the arch all packages?
[17:17] <jtaylor> yes
[17:17] <micahg> add the same line as the regular archive, just put [arch=i386] after deb
=== Guest39737 is now known as JackyAlcine
=== JackyAlcine is now known as Guest48809
=== Guest48809 is now known as JackyAlcine
[17:20] <jtaylor> you mean in apt/sources.list?
[17:22] <jtaylor> I'm trying to get it done in the cacher itself
[17:28] <micahg> jtaylor: oh, no idea how to do that, but the regular archive is faster (mirrors and whatnot), so adding the second deb line should speed things up as well
[17:28] <micahg> it actually needs to be above the ports one though
[17:29] <jtaylor> my connection is so slow that the slowness of ports does not matter much, I'm trying to avoid downloading stuff multiple times
=== JackyAlcine is now known as jalcine
[19:50] <psusi> apport can be configured to collect .xsession-errors right?  how does one do that?  it would really make the apport reports for gparted much more usefull
[20:03] <Ampelbein> psusi: Yes, you need a package hook. See https://wiki.ubuntu.com/Apport/DeveloperHowTo
[20:08] <psusi> does debian also use apport?
[20:09] <psusi> i.e. would they care to include the hook, or will that need to be an ubuntu deviation?
=== el2ro_ is now known as el2ro
[20:14] <psusi> what's this do? report['SuspiciousXErrors'] = xsession_errors(re.compile('CRITICAL.*assertion.*failed'))
[20:14] <psusi> does that pull lines from .xsession-errors that mention assertion failure?
[20:15] <psusi> if so, that's exactly what I'm looking for
[20:46] <Alison_Chaiken> I'm sorry to bother you folks again, but I've been wrestling with reprepro and dh_install and debian/rules for days now.
[20:47] <Alison_Chaiken> The nub of the problem I'm having is "dh_install: <package>-dev missing files (usr/include/*), aborting"
[20:48] <Alison_Chaiken> I see that usr/include/* is called out in debian/<package>-dev,install.
[20:48] <Alison_Chaiken> oops, debian/<package>-dev.install
[20:49] <Alison_Chaiken> I'm not sure if dh_install can't find the source of the files or the destination.
[20:50] <Alison_Chaiken> I'm sure that I have to change the rules to tell dh_install where to look, and I'm sure it's under the binary-arch target, but I can't figure out how to get it to emit a more verbose error message so that I can figure out what it's missing.
[20:50] <jtaylor> which compat level?
[20:51] <jtaylor> you get more verbose output by exporting DH_VERBOSE=1
[20:52] <jtaylor> or running dh_install -v
[20:54] <Alison_Chaiken> I set DH_VERBOSE=1 in debian/control, but the "missing files" line above is all I get before dh_install dies.
[20:54] <Alison_Chaiken> I'm not sure what "compat level" is.
[20:54] <jtaylor> the value in debian/compat
[20:54] <Alison_Chaiken> I was using debuild at the CLI.
[20:54] <Alison_Chaiken> compat level 8
[20:55] <jtaylor> is anything in debian/tmp/usr/include after the build?
[20:55] <Alison_Chaiken> I should try to figure out how to invoke dh_install from the CLI, I guess.    Maybe that would give more info.
[20:55] <jtaylor> yes with -v
[20:55] <Alison_Chaiken> There is no debian/tmp/usr
[20:56] <jtaylor> thats most likely your problem
[20:56] <Alison_Chaiken> There's debian/<package> and debian/<package>-dev though.
[20:56] <ajmitch> what's DESTDIR set to?
[20:56] <Alison_Chaiken> So the tmp directory is not getting created, or is in a weird place?
[20:57] <Alison_Chaiken> I didn't set DESTDIR by hand.
[20:57] <Alison_Chaiken> INSTALL_ROOT is set in the Makefile though.
[20:57] <Alison_Chaiken> and a manual "make; make install" does what I expect.
[20:57] <jtaylor> what buildsystem?
[20:57] <Alison_Chaiken> cmake.
[20:58] <jtaylor> and debhelper tiny rules?
[20:58] <Alison_Chaiken> maybe the problem is that cmake's files are in build.
[20:58] <Alison_Chaiken> I know not the debhelper tiny rules.
[20:58] <jtaylor> how does your rules file look like?
[20:58] <psusi> #2  0x00007fcb279c1452 in ped_assert () from /tmp/tmp5MOi2b/lib/x86_64-linux-gnu/libparted.so.0
[20:58] <psusi> No symbol table info available.
[20:58] <psusi> taht means that the retracer couldn't find the dbg file right?
[21:00] <jtaylor> DESTDIR must be set, but usually that is done by the packaging helpers
[21:00] <jtaylor> if you aren't using them you have to set it yourself
[21:01] <Alison_Chaiken> My rules file just has DH_VERBOSE=1 and "include blah/debhelper.mk"
[21:02] <jtaylor> no cmake.mk?
[21:02] <Alison_Chaiken> I'll try setting DESTDIR and see what happens.    What goes in DESTDIR?
[21:03] <ajmitch> Alison_Chaiken: you mean "include /usr/share/cdbs/1/rules/debhelper.mk"?
[21:03] <Alison_Chaiken> Huh, looks like it gets a lot further if I include the cmake.mk besides debhelper.mk!
[21:03] <jtaylor> $(CURDIR)/debian/tmp
[21:04] <jtaylor> as expected when you use a cmake buidlsystem :)
[21:04] <Alison_Chaiken> DESTDIR=$(CURDIR)/debian/tmp then?
[21:04] <jtaylor> you don't need to set it yourself if you use cdbs cmake module
[21:05] <Alison_Chaiken> Adding cmake.mk means that I now have debian/tmp with files in it.
[21:05] <ajmitch> great
[21:05] <jtaylor> yes cmake.mk sets up cmake correctly for you
[21:05] <Alison_Chaiken> Still stops eventually, but much further along.
[21:05] <ajmitch> where does it stop now?
[21:06] <Alison_Chaiken> Just to be clear, what was wrong before?   dh_install couldn't find the sourcedir or couldn't find the destdir?
[21:06] <jtaylor> DESTDIR was not set
[21:06] <jtaylor> probably cmake was not executed either
[21:06] <jtaylor> and it just used a preshipped makefile
[21:06] <Alison_Chaiken> My biggest problem with the packaging system is that it is too terse.   If I knew which files were missing, I'd figure it all out.
[21:07] <ajmitch> so it wasn't installing the built files into the right place for dh_install to pick up
[21:07] <jtaylor> with export DH_VERBOSE=1 it prints every single command
[21:08] <Alison_Chaiken> The commands are printed all right, but the error messages are canned and don't contain enough info about the failure.
[21:08] <ajmitch> cdbs does cover up a bit too much of the magic for some people
[21:08] <jtaylor> it should ahve contained something like this: cp -a debian/tmp/... file not found
[21:08] <jtaylor> what more info do you need?
[21:09] <Alison_Chaiken> ajmitch, when you talk about dh_install picking up, I'm still not sure if it found the files and didn't know where to copy them, or if it didn't find them.
[21:10] <Alison_Chaiken> so the problem then jtaylor is that debian/tmp wasn't created (it was the missing DESTDIR I guess) and dh_install tried to look for files in debian/tmp/usr/include and there weren't any?
[21:10] <ajmitch> it didn't find them, there should have been other errors printer, like jtaylor said
[21:10] <ajmitch> right
[21:11] <Alison_Chaiken> So debuild compiles the source files using the package's native build system, it copies the binaries and sources to debian/tmp, and then it tars and compresses them?
[21:12] <jtaylor> no
[21:12] <jtaylor> it uses the native build system to install to debian/tmp using DESTDIR=debian/tmp and PREFIX=/usr
[21:12] <jtaylor> then it installs from there into debian/package-name and tars it
[21:12] <jtaylor> if you only have one package you can skip the redirection to debian/tmp
[21:13] <EvilResistance> is there a guide to setting up debhelper to run a setup.py python script for a package i'm building for a project/team?
[21:13] <Alison_Chaiken> I want to build the source and binary packages, so that's already two, I guess.
[21:13] <EvilResistance> general packaging, not necessarily PPA packaging ;P
[21:13] <jtaylor> no the source package is what you are building from, the binary packages are the result
[21:13] <jtaylor> but a libsomething + libsomething-dev are two binary packages
[21:14] <Alison_Chaiken> Ah, okay, package and package-dev are what I meant.
[21:14] <Alison_Chaiken> Thanks so much you guys for your help.
[21:15] <Alison_Chaiken> I will thank you on our project wiki . . . once I have my archive there!   (RSN, we hope, before deadline.)
[21:17] <ajmitch> EvilResistance: it can be as simple as http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/precise/python-configglue/precise/view/head:/debian/rules
[21:18] <EvilResistance> will it auto-detect the setup.py script?
[21:18] <jtaylor> if there is no makefile yes
[21:18] <EvilResistance> or do i have to tell the thing to run it, say, post-install
[21:18] <EvilResistance> ah, very good, so its easier than I thought :P
[21:18] <jtaylor> this one is a bit more complex: includes tests and python3 packaging http://anonscm.debian.org/viewvc/python-modules/packages/ipython/trunk/debian/rules?revision=19761&view=markup
[21:26] <EvilResistance> i think the initial one that ajmitch linked me will work, i'll test once i actually package the thing up :P
[21:26]  * EvilResistance does numerous test builds with pbuilder before releasing a package publicly :P
[21:32] <ScottK> Here's another example that's not as complex as ipython: http://anonscm.debian.org/viewvc/python-modules/packages/python-ipaddr/trunk/debian/rules?revision=19761&view=markup
[21:33] <ajmitch> automagic 2to3 at build time?
[21:33] <jtaylor> if upstream supports it yes
[21:34] <jtaylor> hm ipaddr does not use --force for install
[21:34] <jtaylor> that can cause issues with timestamps
[21:35] <jtaylor> but I forgot the situation where that bit me :/
[21:39] <jtaylor> btw thx ScottK for the qt py3 packages
[21:39] <ScottK> Here's another on of the magic 2to3 packages: http://anonscm.debian.org/viewvc/python-modules/packages/pyspf/trunk/debian/rules?revision=19761&view=markup
[21:39] <ScottK> jtaylor: Thanks to your bug, they even work now.
[21:40] <ScottK> BTW, POX gets some credit/blame for them too (although he declined to say so in debian/changelog)
=== EvilResistance is now known as Resistance