micahg | cyphermox: that freecad upload could've been another no change rebuild to fix the changelog entry | 04:19 |
---|---|---|
micahg | cyphermox: I meant build2 vs ubuntu1 :) | 04:20 |
ScottK | micahg: He knows. It was an accident. | 04:21 |
micahg | ScottK: was looking for scrollback and didn't see any | 04:21 |
ScottK | It was a PM. | 04:21 |
micahg | ok | 04:22 |
=== almaisan-away is now known as al-maisan | ||
=== al-maisan is now known as almaisan-away | ||
dholbach | good morning | 07:06 |
mitya57 | good morning! :) | 07:07 |
iulian | Morning dholbach. | 07:32 |
dholbach | hey iulian | 07:37 |
jokerdino | hey MOTU folks, can someone take a jab at this question - http://askubuntu.com/questions/144052/how-to-file-a-bug-report-for-a-specific-ubuntu-release | 09:43 |
jokerdino | i reckon you guys are the perfect ones to answer this question ^^ | 09:43 |
=== mitya57_ is now known as mitya57 | ||
=== ara is now known as Guest64238 | ||
=== almaisan-away is now known as al-maisan | ||
=== al-maisan is now known as almaisan-away | ||
=== almaisan-away is now known as al-maisan | ||
=== al-maisan is now known as almaisan-away | ||
muelli | hey folks. I have a subversion repository with a debian/ folder. It's libnfc from http://libnfc.googlecode.com/svn/trunk. And all I want it a source package that I can upload. Bonus points if I can alter the arguments to dh_auto_configure. git-buildpackage would work reasonably well, but svn-buildpackage doesn't work as expected :-\ I.e. I need to generate a .orig tarball myself... Is that really true? | 12:58 |
mitya57 | muelli: .orig tarball should be downloaded from upstream, you shouldn't change or regenerate it | 13:18 |
mitya57 | if svn-buildpackage doesn't work, you can unpack upstream and your source into one directory | 13:19 |
mitya57 | cd there and run debuild | 13:19 |
muelli | mitya57: I *have* already downloaded the SVN repository. So all the sources are there. | 13:24 |
mitya57 | muelli: if the upstream svn contains debian directory, the package should be native instead | 13:26 |
mitya57 | so, you need to change debian/source/format to "3.0 (native)" and | 13:26 |
muelli | well, that shows 3.0 (quilt) | 13:26 |
mitya57 | drop "-0" prefix from the first (from top) entry in debian/changelog | 13:26 |
mitya57 | muelli: change "quilt" to "native" | 13:27 |
mitya57 | this is the most simpliest solution, | 13:27 |
mitya57 | you won't have to deal with separate .orig and .debian tarballs | 13:28 |
mitya57 | *the most simple | 13:28 |
muelli | mitya57: alright. Sounds good. thank you very much. | 13:28 |
mitya57 | you are welcome | 13:29 |
muelli | Still a bit annoying. I presume, I can convince dpkg-source to just assume the format to be native. | 13:29 |
muelli | and quilt format would indicate there to be patches/ directory and a series file, is that correct? | 13:30 |
mitya57 | dpkg-source is for generating .dsc files | 13:30 |
mitya57 | quilt format is the most common format, it supports debian/patches but doesn't require that | 13:31 |
muelli | how would I best change the "original" from SVN then. I could obviously modify in place. But should I rather export the the patch (i.e. svn diff), put it somewhere and revert the change? | 13:35 |
mitya57 | just modify in place | 13:36 |
mitya57 | are you going to just build that package for yourself, or upload it somewhere? | 13:36 |
muelli | mitya57: well, I want to scratch my own itch first, but I want to do it as sustainably as possible. So it'll at least end up on my ppa | 13:39 |
mitya57 | in any case you'll have to modify that manually (or manually build the tarball, if you leave the format unmodified) | 13:40 |
mitya57 | every time the upstream changes | 13:40 |
muelli | thanks for your help so far. Highly appreciated. | 13:41 |
muelli | another dput related question. So far I have to do "dput -f ppa: ..." and I haven't had any success trying to configure a default. I want to make it "dput *.changes" only, i.e. without the -f. Is that possible? | 13:42 |
muelli | ah, my ~/.dput.cf has method sftp, while the dput -f does ftp *thinking* | 13:43 |
mitya57 | my .dput.cf has ftp | 13:44 |
mitya57 | it's anyway secure because it checks your pgp key on server side | 13:44 |
mitya57 | and I do just "dput my-ppa PKGNAME_VERSION_source.changes" | 13:45 |
muelli | yeah, well. sometime I can only go out of the network with port 22, 80 and 443. But right now it works, so I'll give it a shot | 13:45 |
mitya57 | I have to go out, sorry | 13:46 |
=== jbicha is now known as Guest70781 | ||
=== Guest70781 is now known as jbicha_ | ||
PaoloRotolo | Hi all! | 15:11 |
gnomefreak | is gir1.2-gjsdbus-1.0 still needed in 12.10? | 15:18 |
AmberJ | Hello | 16:09 |
AmberJ | I'm trying to create package(s) for OpenCog using instructions on ubuntu wiki.. | 16:10 |
AmberJ | When I run 'debuild -uc -us', all goes fine until I get this: http://pastebin.com/VqwrPZim | 16:11 |
AmberJ | Line 19 is where some errors start to pop up... | 16:11 |
AmberJ | .deb packages for opencog don't exist. I'm starting from scratch | 16:14 |
AmberJ | Line 19 says: "cp: cannot stat `../../opencog-0.1/opencog/embodiment/Control/MessagingSystem/router': No such file or directory" ... But this file exists in the specified location | 16:19 |
AmberJ | Also, opencog uses cmake and I didn't made any changes to debian/rules (autogenerated by 'bzr dh-make'). | 16:30 |
AmberJ | Though I read about some cmake edits that can be made to debian/rules, I didn't made any. And the build process seems to go fine without any edit to debain/rules | 16:31 |
=== jacob_ is now known as jacob | ||
=== Pici is now known as Guest55530 | ||
=== Pici` is now known as Pici | ||
jtaylor | bug 700036 | 18:02 |
ubottu | Launchpad bug 700036 in blcr (Ubuntu Natty) "package blcr-dkms 0.8.2-15ubuntu1 failed to install/upgrade: blcr kernel module failed to build - error: ‘struct signal_struct’ has no member named ‘count’" [High,Triaged] https://launchpad.net/bugs/700036 | 18:02 |
jtaylor | just wanted the title, lp times out in the browser | 18:02 |
tumbleweed | reloading often works when bits of lp timeout. If the relevant bits of table are in RAM, it may not timeout | 18:05 |
jtaylor | does not help :/ those blcr bugs almost always time out, probably because of the hundreds of dupes | 18:07 |
tumbleweed | opened for me, first time | 18:08 |
tumbleweed | oh, also for bugs, .../+text | 18:08 |
jtaylor | strange | 18:09 |
jtaylor | a +text works | 18:09 |
jtaylor | thx | 18:09 |
frceego | Hello | 18:26 |
frceego | How common (and difficult) is it to create multiple packages from a single version controlled repo of a project that uses cmake? | 18:29 |
frceego | The project has many component programs many of which can serve as independent program/libraries. | 18:30 |
frceego | But they are all tied under a single cmake project using top-level CMakeLists.txt file. | 18:30 |
jtaylor | you build multiple source packages? | 18:31 |
bcurtiswx | SpamapS, re Bug #1003465, do I still need all that information with a microrelease SRU ? | 18:31 |
ubottu | Launchpad bug 1003465 in empathy (Ubuntu) "SRU request for bug fix release 3.4.2.1" [Wishlist,New] https://launchpad.net/bugs/1003465 | 18:31 |
jtaylor | a source package can have multiple binary packages | 18:32 |
frceego | The code for the complete project is in a single upstream source package. | 18:32 |
SpamapS | bcurtiswx: I don't see a micro release exception for empathy here: https://wiki.ubuntu.com/StableReleaseUpdates/MicroReleaseExceptions | 18:33 |
SpamapS | bcurtiswx: so yes. At least you need to discuss the risk of regression created by introducing all of the changes. | 18:33 |
frceego | The code for the complete project is in a single upstream source package."cmake". I could not find much info about cmake+multipleBinaries on Ubuntu wiki... So, I'm reading debian's new maintainer's guide. | 18:34 |
frceego | oops, typo. I meant: I'm interested in this kinda project using "cmake". The upstream source is tied under a single cmake project() | 18:35 |
jtaylor | you're using cmake to build the package? | 18:35 |
bcurtiswx | SpamapS, hmm I thought there was some exception with GNOME packages and microreleases | 18:36 |
frceego | yes jtaylor | 18:36 |
jtaylor | I'm not familiar with that | 18:36 |
frceego | ok | 18:36 |
jtaylor | multiple binary packages from one source is possible (and normal), but no idea if cmake supports it | 18:36 |
SpamapS | bcurtiswx: none that I see. | 18:37 |
SpamapS | bcurtiswx: I think in practice, a lot of them have been waved through | 18:37 |
geser | jtaylor, frceego: the splitting into seperate binary packages is done after the "make install" during the package build by specifing in debian/$package.install which files go to which binary package | 18:39 |
bcurtiswx | SpamapS, hmm. OK. Thanks | 18:39 |
jtaylor | yes but how to teach that to cmake? | 18:39 |
jtaylor | or cpack in this case | 18:40 |
geser | you don't need to tell cmake it, cmake install into $DESTDIR and debhelper (more precisely dh_install) moves then the files around | 18:41 |
jtaylor | that requires modifying debian/ | 18:41 |
jtaylor | I guess what is wanted is automated debian packaging from cmake | 18:41 |
frceego | jtaylor: sorry, if I was not clear. I'm NOT using cpack. All I do is use cmake to build project. | 18:42 |
geser | to build a proper package you need files in debian/ anyways | 18:42 |
jtaylor | geser: I'm not sure about that cpack can output to debs and rpms | 18:43 |
jtaylor | I don't thnk you need a debian dir for that | 18:43 |
geser | I don't know about cpack, so can't comment on it | 18:44 |
jtaylor | frceego: in that case do what geser said | 18:44 |
frceego | yes. noted. | 18:44 |
geser | if you want something which can be uploaded to the archive (or a PPA) you need to be able to build the binary package with the usual Debian/Ubuntu tools (dpkg-buildpackage) | 18:45 |
frceego | yes, someone told that to me a few days ago. So, I quit the idea of using cpack and now I'm learning about debian/ubuntu tools .. | 18:46 |
AmberJ | Any ideas why my 'debuild -uc -us' fails? all goes fine until I get this: http://pastebin.com/VqwrPZim | 18:47 |
jtaylor | something is broken with the build, see line 19 | 18:48 |
AmberJ | jtaylor, Line 19 says: "cp: cannot stat `../../opencog-0.1/opencog/embodiment/Control/MessagingSystem/router': No such file or directory" ... But this file exists in the specified location | 18:51 |
geser | if computed the relative path correctly based on this log then it tries to find a file /home/amberj/opencog-0.1/opencog/embodiment/Control/MessagingSystem/router | 18:51 |
AmberJ | err..wait | 18:52 |
AmberJ | I was wrong. The file exists instead at /home/amberj/dustbin/opencog-0.1/opencog/embodiment/Control/MessagingSystem | 18:57 |
AmberJ | This has to do either with project's cmake or me messing with cmake's cache for the project. | 18:57 |
AmberJ | Let me retry by doing a fresh checkout.. Once I'm done, I'll report here. | 18:58 |
AmberJ | Thanks geser jtaylor | 19:04 |
AmberJ | Is there a (de facto) standard way to create upstream tarballs? | 19:17 |
tumbleweed | depends on the upstream | 19:18 |
AmberJ | The project I'm working with does not provides upstream tar files atm. So, I typically checkout/branch the bazaar repo and use 'find' to recursively delete .bzr/ | 19:18 |
AmberJ | And then, tar the directory | 19:19 |
tumbleweed | that works | 19:19 |
tumbleweed | bzr has an export command btw | 19:19 |
geser | can bzr export as a tar.gz too? | 19:20 |
tumbleweed | AmberJ: there are a bunch of recipes in the old packaging guide: https://wiki.ubuntu.com/PackagingGuide | 19:20 |
tumbleweed | geser: yes | 19:20 |
AmberJ | tumbleweed, thanks, bzr's export was what I was looking for. | 19:21 |
=== yofel_ is now known as yofel | ||
AmberJ | geser, jtaylor Just done with deb package... project's cmake config was fine; it seems I messed with cmake's cache last time | 21:05 |
=== highvolt1ge is now known as highvoltage | ||
dupondje | Somebody around with some GTK coding knowledge ? | 22:21 |
chrisccoulson | dupondje, sometimes it's better just to ask your question ;) | 22:41 |
chrisccoulson | (and you might get a better response in #ubuntu-desktop) | 22:41 |
=== jbicha_ is now known as jbicha |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!