[00:00] wondering if any reviewers have time to take a final look at my pyrocket package http://revu.ubuntuwire.com/details.py?upid=2857 [00:00] kostmo, sure, I'll look [00:01] thx [00:02] kostmo, loose the debian-binary file unless its needed for some reason [00:03] ok i'll delete it [00:03] Nice job on putting the manual page in the right section :-) [00:04] looks good to me [00:07] WHere do mime files go? [00:40] someone can make a revision in my package? http://revu.tauware.de/details.py?package=umit [00:42] * NCommander fires a massive cannonball size hack [00:42] ^Into Code::Blocks [01:10] Hello motus! can pbuilder be used to build source packages (as though i ran 'debuild -S'). I just managed to get pbuilder installed and running, and i'd rather not create a hardy chroot now as well (sorry to those who saw this in -bugs too) [01:15] kgoetz: Why would you do that? [01:20] wgrant: why would i do what? use pbuilder or a chroot? [01:22] * NCommander finally got codeblocks nearly lintian clean [01:22] take a look in /var/cache/pbuilder/result/ [01:22] or specify --buildresult [01:25] i might have to keep reading on pbuilder *grin*. seems i'm not as aware of whats its doing as i thought [01:26] is it just me, or is Ctrl-Alt-R not normally supposed to restart X? [01:27] Anyway, NCommander, I need more help [01:27] help on what? [01:27] with my watch file [01:27] or my uscan invocation [01:28] oh, I see [01:28] hang on [01:28] * NCommander is down to two lintian issues [01:30] if the openssl debacle taught us anything, it's that you should never try to make your packages lintian-clean [01:30] directhex, is that how that happened? [01:31] ok fixed that [01:31] that's half the story, yeah [01:31] any place where I can read the full story? [01:32] ok, so now I get one Lintian error, which is expected [01:32] (I don't clear a lintian error if its invalid; I always add overrides) [01:32] I do need to do something about it, I'm just not sure what [01:33] it's about the lack of a SONAME [01:33] what's the specific lintian error? [01:33] Oh [01:33] You should fix that [01:33] and I know I should add a SONAME, I just can't decide what it should be based around [01:33] because upstream has no SONAME, and AFAICT, the intention is to keep the current release (2.4) binary-compatible [01:34] so I think I can just add a SONAME and forget [01:34] NCommander, short version: patch is made for odd-looking code to fix lintian moan, patch is applied to originally planned location, plus an identical looking place further on in the source (don't you love variable names like "e"?). one of these places fixed lintian trouble, t'other caused massive reduction in entropy [01:34] *grumble* [01:34] I'd never change code over a lintian issue [01:35] WIth the exception that I had to change a hardcoded path to move plugins from /share to /lib [01:36] is -soname ignored when not making a shared object? [01:38] kgoetz: Why would you not build the source package on your normal machine? [01:39] what's the format of a soname again? (being stupid) [01:44] NCommander, I've got a few dpkg-source warnings I'd like to ask about? [01:44] Shoot [01:44] Version number suggests Ubuntu changes, but Maintainer: does not have Ubuntu address [01:45] Version number suggests Ubuntu changes, but there is no XSBC-Original-Maintainer field [01:48] coppro, run update-maintainer in the debian folders [01:48] (and then fix the changelog to reset your name) [01:48] since its a new package, you'll have to pass --section=universe [01:49] I [01:49] 'm also getting Use of uninitialized value in pattern match (m//) at /usr/bin/dpkg-source line429. [01:49] I've seen it before, but I forgot >.<; [01:52] ok thanks for the help [01:52] ok, now I've got warning, `debian/libmk4-dev/DEBIAN/control' contains user-defined field `Original-Maintainer' [01:57] coppro, that's normal [01:57] It's because Ubuntu uses a nonstandard field that Debian doesn't recongize [01:57] ok [01:57] I've considered submitting a patch to make that warning disappear, but the headache of getting a patch into dpkg and friends sounds like too much pain [02:00] NCommander: I re-uploaded the package without the debian-binary file: http://revu.ubuntuwire.com/details.py?upid=2891 [02:00] ... is there a second reviewer around, by chance? [02:01] kostmo, I'm not an MOTU, I can't sign your package [02:01] how do I get this stupid soname to work? [02:01] ah IC [02:01] the only thing I can think of is modifying the upstream Makefile.in [02:02] do you have any suggestions for the best communications channel to find a couple MOTUs for this? [02:03] y [02:03] disregard that [02:09] Heya gang [02:20] this is fail :( [02:20] NCommander, I got the SONAME in, but now I can't link /usr/lib/libmk4.so to /usr/lib/libmk4.so.4! [02:29] ok done [02:29] yy [02:29] *yay [02:33] ScottK, ping [02:34] NCommander: scottk is out [02:35] darn [02:35] emgent, care to review a package that's about to hit REVU? [02:36] NCommander: not now, i'm working. anyway feel free to link here :) [02:36] You do work? [02:38] NCommander: phpmyadmin security [02:38] how do I add a lintian override? [02:39] http://lintian.debian.org/manual/ch2.html - coppro [02:40] so I make a file then install it to usr/share/lintian/overrides, correct? [02:41] Alo! [02:47] wgrant: because i'm A. not running ubuntu, and B. dont want to fill my system with build-deps [02:58] Ok, question: what is the standard place to put your lintian-overrides file? Like, should it be .lintian-overrides? And what commands should be in rules? [03:07] kgoetz: If you're running something Debianish, you're fine. You generally don't generally need to install the build-deps. [03:12] and away codeblocks goes to revu [03:13] yay! [03:13] go NCommander [03:13] now can you answer my questions? [03:16] what was your question? [03:16] for lintian-overrides, I understand where they go in the package [03:16] but in my source folder, how should I put them, and is there a special dh_installblah command? [03:16] dh_lintian [03:17] But its only available in Intrepid/Sid, so if you want to backport, don't use it, and manually at it to your package [03:25] what package has all the development tools? [03:25] http://revu.tauware.de/details.py?package=codeblocks [03:25] someone care to review? [03:26] NCommander, upstream ships a debian/ directory? :( [03:27] cody-somerville, yeah [03:27] But its pretty broken [03:27] I had to redo at least 70% of the files in it [03:27] NCommander, you should see about getting upstream to remove it [03:27] I will [03:28] They release debs because no one had packaged Code::Blocks for Debian or Ubuntu [03:28] But whoever did it wasn't a DD [03:28] NCommander: that's the same situation I'm in with Bibus [03:28] NCommander: it's a mess upstream, but it's not in the repos yet, so hopefully once this mess is sorted tonight It will be soon :-) [03:28] tbielawa, yeah, once its in the archive, I'll use it as justification to removing it [03:28] ^upstream [03:30] umm... code::blocks provides .debs [03:30] @all, regarding lintian overrides. my upstream version ends with a dash then a final number, ex: 2.4.3-2 [03:31] that throws up the natice package with dashed version lintian warning. Is this justification enough to override the warning with an override? [03:31] I said it does [03:32] coppro, but their debs don't have proper dependencies in them, and violate the LSB [03:32] yeah, but that's why they have a debian/ [03:32] coppro, I think we're well aware of that. [03:32] No, I know [03:33] ok [03:34] I'm planning to submit this package to Debian after it gets in here [03:34] And ask them to remove the Debian folder [03:34] NCommander, Remove the URL from the package descriptions and put it in the Homepage field of the source package. [03:35] Ack, I knew I forgot to do something; and I also forgot to add the watch file >.<; [03:35] Sorry, its late [03:35] what package is debuild in? [03:38] NCommander, Also, codeblocks-contrib should be a suggest and not a recommend. [03:39] cody-somerville, *grumbles*, I can't find a watch file line that will work with sourceforge since they changed the layout [03:39] any suggestions? [03:39] NCommander: I'll post one online that works fo rme [03:40] thank you tbielawa [03:40] * NCommander hasn't done any real packaging in so long its good to get the rust off [03:40] http://sf.net/codeblocks/codeblocks-(.*)\.tar\.gz ? [03:40] Didn't work, but I'll retry it [03:40] http://revu.ubuntuwire.com/revu1-incoming/bibus-0807090710/bibus-1.4.3-2/debian/watch works for me [03:41] tbielawa, it mostly works, but I'm getting binary packages like [03:41] http://qa.debian.org/watch/sf.php/codeblocks/codeblocks-8.02-0suse103-x86_64.tar.gz [03:41] It seems suse > ubuntu to uscan ;-) [03:41] haha [03:41] er [03:42] html > 0, as I discovered last night [03:42] http://downloads.sourceforge.net/codeblocks/codeblocks-(.*)-src\.tar\.bz2 [03:42] good luck, i had enough problems getting that one to work [03:42] how should the suggest/recommend relations in a library package be set up [03:42] just the dev recommends doc, or anything else? [03:43] Not to be lazy, but how can I get uscan to download the package file for get-orig-source? [03:44] actually [03:44] Debian policy says I should download the same source file, so just a simple wget [03:47] cody-somerville, thanks for reviewing this ;-) [03:48] cody-somerville, upload in progress [03:51] wow, stupid question day! [03:51] wait, nvm [03:51] * cody-somerville raises an eyebrow. [03:51] I was about to ask a stupid question [03:51] but then I thought of an answer [03:51] wgrant: then you and Hobbsee just gave me different answers on th same subject [03:52] wait, is there any way for dh_install to change the name of the file it installs? [03:52] cody-somerville, upload complete [03:52] THank you for taking the time to review this [03:53] or should I just use /usr/bin/install [03:53] * NCommander is always amazed at how simply, yet effective the deb package format is [03:54] someone help me... I can make my next upload as soon as I resolve these stupid overrides [03:54] NCommander, The md5 of your source tarball mismatches with upstream's [03:54] coppro, yes [03:54] I have debian/libmk4.lintian-overrides [03:54] I want it to be install to /usr/share/lintian/overrides/libmk4 [03:54] what should I put in rules? [03:55] cody-somerville, mostly because I repacked it into a tar.gz as per the debian policy guide [03:55] NCommander, except that isn't debian policy [03:55] * NCommander rereads the policy guide [03:55] It used to be [03:55] NCommander, it is possible to make it a .tar.gz w/o repacking it. [03:56] help? [03:56] get-orig-source: [03:56] cd .. && wget http://internap.dl.sourceforge.net/sourceforge/codeblocks/codeblocks-8.02-src.tar.bz2 [03:56] bunzip2 codeblocks-8.02.tar.bz2 [03:56] gzip codeblocks-8.02.tar [03:56] Maybe something accidently ended up in the orig [03:56] NCommander, gzip -9 [03:56] coppro: if you just want to install a file you can use the standard pkg_name.rules file in your debian/ directory [03:56] Let me just remake it [03:57] what command in rules should I use though? [03:57] cody-somerville, nice catch [03:57] dh_install doesn't seem to support renaming, and /usr/bin/install won't create the directory structure [03:57] install -d will create the entire path [03:57] (and is bad anyway) [03:57] coppro, there is dh_installdirs [03:58] hmm... oh wait [03:58] i [03:58] but dh_install -i -p should be in your install: build target [03:58] am an [03:58] idiot [03:58] coppro, No, you're not. You're simply rushing. Slow down. [03:58] no im an idiot [03:59] Use this as an opportunity to redeem yourself then, coppro :) [03:59] cody-somerville, I'm uploading now. Sorry for these basic screwups >.<; [04:00] bah [04:00] Such a slow upload cap here [04:01] cody-somerville: upstream tarball is messy. bad folder name for the source, hidden directories like cvs and ide folders, and it's got a debian/. Does this reconstitute a reroll of the .orig.tar.gz ? [04:02] tbielawa: I think -D is the one you meant [04:02] coppro: I think you're right :) [04:02] tbielawa, did you export it from svn? [04:03] thanks for pointing out that the option existed though [04:03] cody-somerville: nope, it's from their sourceforge site. [04:03] tbielawa, then no [04:04] kgoetz: It will need some, but not all, build-dependencies for a source buil. [04:05] yay, time to upload again to REVU (still no docs package though) [04:05] cody-somerville: thanks for clarification. I just re-read the "Complete Ubuntu Packaging Guide" section on non-native packages again and it's like I never saw it before [04:05] coppro: :) [04:07] cody-somerville, updated package with properly done orig.tar.gz ;-) [04:07] what's the command to update the version and add a blank changelog entry? [04:07] dch [04:07] -i [04:08] * cody-somerville never uses that. Causes too many mistakes :P [04:08] and how do I test with pbuilder again? [04:09] coppro: it's involved. https://wiki.ubuntu.com/PackagingGuide/Complete#head-2979f8a826fc15379a92139c25bb513633d2023c [04:09] ty [04:09] it provides a link to the full guide [04:09] np [04:09] cody-somerville, http://revu.ubuntuwire.com/details.py?package=codeblocks [04:11] * NCommander hides in fear from the wrath of Cody [04:12] ok, pbuilder had success! time to actually upload! [04:13] coppro, first package? [04:13] yes [04:13] Nice, welcome to the club;-) [04:13] first actual package that is [04:14] I made one before but someone beat me to it on Debian, so I finished just for practice [04:14] That's what ITPs are for ;-) [04:14] I'm just happy to finally get codeblocks into the archive [04:14] yeah, but whoever it was didn't close the ubuntu needs-packaging bug [04:15] and I didn't realize that Debian had a package when I made it [04:18] sebner: New courier merge with your name on it. [04:18] wow, ScottK [04:18] he speaks [04:18] When I'm here I can almost never manage not to. [04:19] http://revu.ubuntuwire.com/details.py?package=libmk4 [04:19] yay! [04:20] ScottK, care to review codeblocks? [04:20] ScottK, http://revu.tauware.de/details.py?package=codeblocks [04:20] NCommander: Sorry, I've got another project I need to concentrate on with the little time I have tonight. [04:20] It's ok [04:20] * nxvl checks [04:22] wow, this is really counterintuitive [04:22] Blimey, sonicmctails@gmail.com overwrote my package [04:23] an RK4 simulation of an inverted pendulum runs SIGNIFICANTLY faster on C#/Mono than Java [04:23] even when just directly using ikvm to on-the-fly translate the bytecode from Java to Mono [04:23] LucidFox, whoops >.>; [04:24] LucidFox, I should have checked revu before I started packaging codeblocks [04:24] lol [04:24] owned [04:24] hrm or me? [04:24] * coppro ducks [04:24] both [04:25] him for getting his package blasted, you for putting all that hard work in [04:25] java IPModel > /dev/null 35.31s user 0.07s system 100% cpu 35.301 total [04:25] ikvm IPModel > /dev/null 24.76s user 0.06s system 99% cpu 24.824 total [04:25] He got his package blasted? [04:25] NCommander: are you using XSBC-Original-Maintainer: for some reason? [04:25] this is news to me... since when was CLR faster than Java HotSpot? [04:25] I didn't, it's still available [04:25] oh [04:25] well, it's still funny [04:25] NCommander> Please remove .debhelper.log files, at least [04:25] nxvl, upstream shipped a debian folder, I left the original mantainer [04:25] jdong: Did you get my ping about Feisty/Gutsy backports ... [04:26] NCommander> You should repackage the orig.tar.gz and remove it [04:26] Are any MOTU's around? [04:26] Scott, but he's busy apparently [04:26] argh, damn it. I thought dpkg-buildpackage -S cleaned before building [04:26] I'm a MOTU [04:26] LucidFox, I uploaded it again. [04:26] cody, can you REVU my package please? [04:27] ScottK: no, this shell is terrible at keeping history [04:27] LucidFox, so I don't understand what you mean? [04:27] howdy! any chance you'd be inclined to revu mine also? http://revu.ubuntuwire.com/details.py?upid=2891 [04:27] * LucidFox headdesks [04:27] jdong: Dapper/Hardy are caught up. There are a couple in Feisty/Gutsy you need to deal with. [04:28] * NCommander steals LucidFox's headdesk [04:28] But then, given how I submitted four packages for intrepid to REVU so far and they all have been ignored, maybe I was doing something wrong [04:28] cody-somerville: http://revu.ubuntuwire.com/details.py?package=libmk4 [04:28] ScottK: gotcha. I've got tomorrow off work; I'll jot that on my TODO list. Right under "get wisdom teeth pulled" [04:28] :( [04:29] Ow [04:29] yeah, not looking forward to it. [04:29] LucidFox, I'm a porter. I do new packages *really* rarely [04:29] jdong: I've done my massively crack addled backport for this month. Your turn. [04:29] But Code::Blocks is something I use quite a bit, and I wanted to prove I can package [04:29] Obviously I'm proving I can't ;-) [04:30] coppro, you should use use a patch system instead of editing files outside of debian/ directly. [04:31] huh? [04:31] say what? [04:31] cody-somerville, uploading again with all the afforemethoned suggestions [04:32] coppro, You modified unix/Makefile.in [04:32] wgrant: 'some' is still quite a lot if i want to work with varying packages [04:32] coppro, you didn't document it in your changelog either [04:32] NCommander: why is it recommended to remove debian-binary? Actually I can't remember why I put it in, but I'm pretty sure it was on a list of of files that should be in debian/ [04:32] ok, what do you mean by a patch system though? [04:32] coppro, you also modified unix/configure and didn't mention it in your changelog [04:32] coppro, ie. https://wiki.ubuntu.com/PackagingGuide/PatchSystems [04:33] NCommander: list I read was probably outdated [04:33] kostmo, its generated automatic for whatever version of dpkg file format is now used [04:33] ok ty [04:33] ok [04:34] coppro, You also don't need to put [Your Name] above ever item in your changelog. [04:34] that was left over from a script [04:34] coppro, I don't care. :) [04:34] lol [04:35] NCommander, My computer is still working on compiling the first version you uploaded :P [04:35] cody-somerville, ahahaha. need to upgrade a little bit? [04:36] Maybe [04:36] (then again, distcc, and -j4 are my friends) [04:37] NCommander: Aren't you a porter to $SLOW_ARCH? === asac_ is now known as asac [04:37] like I said, distcc and -j4 [04:38] emgent, try getting some of your fixes upstream :P [04:41] does " pdebuild --pbuildersatisfydepends" need to be called with any other options? i dont see any in the man page, but it doesnt install build-deps for the package i want to build. http://paste.ubuntu.com/28184/ [04:43] kgoetz, does your package depend on things from universe? [04:44] cody-somerville, could you revu my package, please? http://revu.ubuntuwire.com/details.py?upid=2891 [04:46] cody-somerville: sure, i use first see if in BTS bug is open, and than talk with maintainer on OFCT [04:46] anyway thanks for suggest :) [04:48] NCommander: i'm not sure, but they look like main packages (python-all-dev (>= 2.3.5-10) cdbs (>= 0.4.41) zlib1g-dev) [04:49] Hrm [04:49] Well, try adding the universe repo to your .pbuilderrc [04:49] ok, i'll try that and let you know how it goes. thanks. [04:52] * NCommander waits for cody-somerville's computer [04:55] cody-somerville, ping? [04:56] NCommander, yeees? :] [04:57] cody-somerville, how goes the compiling? [04:57] Maybe it would be faster to upload to your PPA and let launchpad do it [04:57] NCommander, it is almost done. [04:57] * NCommander waits patiently [04:57] NCommander lies. [04:57] He's not patient at all. [04:58] * NCommander clubs ScottK [04:58] Well, the security guys were scared with a 32 minute turn around to backport a patch to dapper, feisty, and confirming gutsy wasn't effected .... [04:59] so I dunn0o [04:59] Maybe I'm just used to fixing packages fast to make up for slow build times [04:59] (or I need some outlet for my excess energy; its been completely dead at work; damn fire company going to go out of business at this rate) [05:00] ScottK, so yeah, I now do security patches in addition to FTBFS .... I think I've got no life [05:01] NCommander, you should help contribute to Xubuntu. [05:01] What does Xubuntu need? [05:01] (I run it, so I'd be glad to help, but I've never seen a todo) [05:01] and I'll brb, dinner [05:04] NCommander, linda is dirty. [05:05] NCommander: i've rebuilt and i think it contains universe, but i still dont get the build-depends satisfied [05:06] kgoetz, pastebin your build log [05:08] cody-somerville: package build? half a mo [05:08] cody-somerville: http://paste.ubuntu.com/28187/ [05:10] kgoetz, that doesn't look like it is building in a chroot [05:12] NCommander, you didn't add a Homepage field [05:13] cody-somerville: so its not attempting to use pbuilder? :/ [05:13] I've never used pdebuild myself [05:14] mm ok. === nenolod is now known as yamarisu [05:15] Okay, I'll do one more review before I go to bed/take a nap. [05:17] Damn it, woohoos [05:17] *sigh* [05:17] I'm really out of it [05:17] I'm going to call it quits after I upload this last one, and stop torurting cody-somerville [05:20] cody-somerville, does it need Homepage: as in or loose the brackets [05:20] Lose the brackets [05:21] Ok [05:21] Uploading [05:21] I'm not going to ask you to stick around and wait for it to build again [05:21] I can find another sponsor, since your probably ready to stab my eyes out [05:23] NCommander, nah. I don't need to stab anyone's eyes out yet :) [05:24] well, if you do, go for my blind one [05:24] I don't need it [05:24] ;-) [05:24] or ... [05:24] .-) [05:30] cody-somerville, its there === yamarisu is now known as nenolod [05:41] \sh: moin [05:44] and I'm back [05:44] SOrry, duty called for the "Automatic Alarm in Building 47 at RIT" [05:44] cody-somerville, I assume your gone? [05:48] I just noticed that the rules file my upstream included actually called another Makefile they havein their Setup/ directory [05:50] It sounds like your having "fun" [05:50] I can't decide if that's better or worse then something say like checkinstall [05:50] this whole package is a messssss. [05:50] How long have you been working with packaging? [05:52] Sometimes its easier to -r debian, and start over [05:52] Unless your just referring to the packages build system [05:52] Debian/Ubuntu, or collectively? [05:52] Collectively [05:52] Hard to say for sure, but long enough to seen some real messes [05:53] :) [05:53] The worse are the package build systems themselves [05:53] Go try building RPM, then tell me this package is bad [05:53] This was the first real piece of software I've tried to package. It was debianized so I thought it would be easier [05:53] small hint [05:53] If its debianized, but not in the archive [05:54] It usually means a lot of pain and hurt [05:54] I can contest to that [05:54] code::blocks was ok [05:54] THe hardest part I have with this is deciding how to go about it. I've wanted to just rm the debian/ and some other stuff. But then I have questions like "how much not in debian/ am I supposed to edit without patching and etc?" [05:55] But I still had to redo at least half of their rules file, and apply a nasty hack to the source to get it to meet standards [05:55] eck [05:55] icky [05:55] tbielawa, you can modify anything in the package you like, but you need to add a patching system [05:55] * tbielawa uses dpatch [05:55] same [05:55] I've used quilt before [05:55] But I find I end up fighting the tool more then anything else [05:55] heh [05:56] My sole pet peeve with dpatch is every once in awhile, it looses its mind, and unpatch fails [05:56] Then again, so does quilt [05:56] 0_0 [05:56] It usually means moving the debian folder to safety, and rm -ring the source folder [05:57] this author made it so that it will build as a deb, but it will able make from this make file in their setup directory. THEN I find out just now when I'm about to publish my (hopefully) final REVU upload, that rules calles Setup/Makefile [05:58]