[00:23] very close to a working package - dpkg-buildpackage: warning: Failed to sign .dsc and .changes file what does this mean? [00:24] do i need to have a gpg key? [00:24] you don't have to [00:25] AdamDH: You really should when you actually publish it, but for now you can pass -us -uc to not sign them [00:25] hey Laney [00:25] hi james_w [00:25] good holidays? [00:25] (lenny GR results out for the interested) [00:25] well for the time I just want to test my rules etc I will sign them when I put them up [00:25] ooh [00:25] Laney: it was ok thanks, yours? [00:26] yeah quite good [00:26] watching my new Spooks DVD atm [00:27] cool [00:27] well it finished dpkg-genchanges: including full source code in upload [00:27] dpkg-buildpackage: source only upload (original source is included), where did it put the .deb? [00:27] Laney: when's the global bug jam again? [00:28] or did I miss something [00:28] james_w: February I think [00:29] an 20-22 [00:29] AdamDH: What command did you run [00:29] ? [00:30] Laney: I'd be well up for getting something going in the uk for that. [00:30] dpkg-buildpackage -S -rfakeroot [00:31] james_w: Definitely. Location is always the problem though [00:31] always [00:31] Do you think popey's house would work? [00:31] what area are you based? [00:31] I'm in Nottingham === `Chris_ is now known as `Chris [00:32] I could maybe get a room at the uni, but would people come? [00:32] oh cool, I'm in Nottingham currently [00:32] oh! [00:32] I'll be back there on Monday [00:33] the centre of mass does seem to be in the South though [00:33] quite [00:33] ah, damn, just going to ask if you were around tomorrow :-) [00:33] gone back to cambs for christmas :( [00:34] we could always have a couple of locations and an online presence for those that can't make one of the venues [00:36] I think nailing one down in London would be the best way to go [00:37] u-uk hasn't really run anything like this before afaik [00:37] yeah [00:37] I think a few people met in a pub last time [00:37] I was away I believe [00:38] I think picking a venue and then looking at alternative arrangements for those that can't make it is the best thing to do, as we will never find a location to suit everyone [00:38] I'll post to the list about it. [00:39] seem to have some issues about: make: *** No rule to make target `dh_testdir', needed by `configure-time-stamp'. Stop. [00:39] dpkg-buildpackage: failure: debian/rules build gave error exit status 2 dh_testdir? why that's an inbuilt function? [00:41] AdamDH: do you have a tab character before "dh_testdir" under "configure-time-stamp:"? [00:41] it needs to be a real tab character, and not some spaces [00:43] configure-time-stamp: dh_testdir [00:43] its a tab [00:43] but not on a new line [00:44] it needs to be a new line [00:44] thanks [00:44] if it's on the same line then it is not a command to run but a "target" that must be built before the current one [00:44] i am getting a feel for how this all works now [00:45] took so reading and playing about [00:45] *some [00:46] cool :-) [00:47] really should be doing uni work instead of packaging a cross compiler! [01:08] looks like I got most of it working just getting compile errors now on the source === vorian is now known as awesome === awesome is now known as vorian [01:24] what does it mean when make: build: Command not found? [01:24] after running sudo pbuilder build ../*.dsc [01:25] AdamDH: I suspect another syntax error in your debian/rules [01:25] AdamDH: could you pastebin it and I will have a look? [01:27] yer sure give me a mo [01:35] james_w: http://www.pastebin.ca/1294753 [01:36] ok [01:36] you have an "install" target, where you want to build the stuff before-hand [01:36] so you have "build" afterwards [01:37] as you have put it on the next line, starting with a tab make will interpret it as a command [01:37] to make it a dependency as you require then put it on the same line [01:37] "install: build" [01:38] same with "binary-arch: install" [01:38] and "configure: configure-time-stamp" [01:39] also, at the end of "build-time-stamp:" you do "touch configure-time-stamp" instead of "touch build-time-stamp", which would cause a problem in some cases [01:40] cheers I will add those changes [01:41] thanks for the help [01:42] no problem [01:46] does DIF mean that sync requests require a freeze exception before they are sponsored? [01:47] they arent automagic anymore bobbo, you can still do sync requests [01:49] vorian: ah ok, the Wiki page jsut semed like a lot of words, not making much sense :P [01:49] haha [01:49] i like that [01:50] "a lot of words" :) [01:56] wow even on my macbook pro compiling still takes some time, cant want till I get my dual quad core amd box [01:56] *wait [02:07] Hi [02:15] hey === Jazzva_ is now known as Jazzva [03:58] i am running sudo pbuilder build ../*.dsc and I get ~/msp430/msp430-binutils-ubuntu-package [03:58] any ideas? [04:19] If I am building a custom kernel, and building the debs for it, is there anyway to add extra version information such that I can install the same version of the kernel without my custom modifications? For example, I have linux-image-2.6.24-22-server installed and I would like to patch it and rebuild it. When I install the custom version, how can I make is so that I can have both the Ubuntu package installed and my custom package [04:25] maybe how I can create a custom flavour maybe? [04:25] something like server-custom ? === bluesmoke is now known as Amaranth [06:47] any revu admins around? [11:03] it says: dpkg-deb: building package `msp430-binutils' in `../msp430-binutils_2.18.msp430.cvs.0.0.20081228-1_amd64.deb'. [11:03] dpkg-genchanges >../msp430-binutils_2.18.msp430.cvs.0.0.20081228-1_amd64.changes [11:03] but i cant find msp430-binutils_2.18.msp430.cvs.0.0.20081228-1_amd64.changes [11:04] i mean .deb [11:06] check /var/cache/pbuilder/result [11:06] or wherever else you've set it [11:08] thankyou Hobbse! I just made my first debian package [11:09] it was in there [11:10] :) [11:11] something went slightly wrong as when I check with the gui package installer there is no included files === asac_ is now known as asac === DrKranz is now known as DktrKranz === LucidFox is now known as LucidFox_away [12:20] Question, I've created a Python distutils setup script which in combination with pysupport, debhelper.mk, python-distutils.mk is used to build our package. Lintian still throws a few warnings I need to fix, but I'm currently wondering how I should properly go about adding a postinst to do a notification (update-notifier). See package apt-file for an example. [12:20] You can find the control files here btw: http://code.google.com/p/nautilussvn/source/browse/#svn/branches/v0.12/packages/ubuntu/debian [12:22] I'm assuming there might be some debhelper available to do just this (I could obviously manually created a postinst file, but I'm looking for "The One True Way" (TM) [12:22] you could look at the apt-file source package...? [12:23] directhex: I did [12:24] See my reference to "The One True Way" (TM), also somebody else build a Debian package for us before and he was told "Please remove postinst and postrm.". http://revu.ubuntuwire.com/details.py?package=nautilussvn [12:26] directhex: oh wait, you might mean.... look closer [12:26] * Brucevdk looks closer [12:30] no luck, but I'll just manually created the postinst file for the time being. Somebody will just have to look at it whenever I might submit it for inclusion. [12:30] create* [12:31] Why do you think following the way apt-file's postinst does it is not right? [12:33] maxb: I don't think anything. But it's a package, it's not documentation (guideline, packaging guide, tutorial). I'm not familiar with all the intricacies of Debian packages, and obviously the source package doesn't explain "why" things are done the way they are [12:33] maxb: also it's not using the same methods to build the package (we use pycentral) [12:33] maxb: combined with the fact that some reviewer said to remove the postinst file, well that makes me wonder what "The Right Way" (TM) is [12:36] For example, what does the #DEBHELPER# in the postinst file mean/do? [12:36] The reviewer was objecting to the inappropriate content of the file, not its existence [12:37] #DEBHELPER# identifieds the point in the file at which debhelper will insert automatically addend snippets [12:37] *added [12:38] maxb: now we're talking, this is all I was asking for (a little help/pointers) ;-) Thanks. [12:39] maxb: can you point me in the right direction where I should look what arguments are passed to postinst? apt-file also does a check for [ "$1" != configure ]; then [12:40] The reviewer was calling attention to two specific problems - (1) It's unequivocally wrong to be including debhelper snippets in the source package's postinst, since debhelper is going to add them during the package build, and (2) There's no obvious reason why that symlink wasn't simply shipped in the package [12:41] The authoratative manual for core packaging details is the Debian Policy Manual - including a (rather complex) description of the arguments a maintainer script can expect to be passed [12:43] and yes, the many ways maintainer scripts can be invoked *is* hideously complex [12:43] there is a nice graph in the debian women wiki [12:45] maxb: just browsing the manual is complex ;-P [12:46] maxb: thanks for the help though, everything is just a little clearer than it was before [12:49] Most of the complexity comes from the fact that an installation process can fail, and be rolled back [12:50] e.g. if a prerm was run, but something went wrong and the package then isn't going to be removed after all, postinst gets used to roll back the result of running prerm [12:52] maxb: are you allowed to include comments in postinst etc. to explain this to readers? [12:55] Of course. It's just a shell script like any other [12:57] Well, actually, it's possible to write the maintainer scripts in other languages too, but it's seldom done [12:59] Oh, from now on I will write all maintainer scripts in LOLCODE [12:59] kthx [13:00] :-) [13:00] * StevenK kills Brucevdk [13:00] DO NOT WANT [13:01] sure? === LucidFox_away is now known as LucidFox === _JKHKBeJ is now known as LjL === LjL-Temp is now known as LjL [14:49] * hyperair grumbles at how long revu is taking to publish an uploaded package [14:51] http://revu.ubuntuwire.com/details.py?package=codelite <-- revu, anyone? [15:55] * nhandler is having a fun morning re-reviewing packages on REVU [15:57] * RainCT hugs nhandler :) [15:58] RainCT: If I don't, vorian will pass me up on the Hall of Fame. We can't have that, can we? ;) [15:59] of course not! [15:59] * RainCT whispers to vorian: go, go, go! :P [16:00] lol === fta_ is now known as fta [16:25] jpds: in case that you can manage to come, we'll meet at 20:30 in Plaça Catalunya (BCN) [17:12] could anyone tell me how long sync requests usually take to get done, after they've been approved [17:26] marnold: It depends on how busy the archive admins are. [17:27] thought so just wanted to ask [17:34] how do i set files to be excluded from a package.install file? [17:39] ..is there nobody in this room? [17:39] everyone is probably busy [17:40] hyperair: man dh_install suggests that this is only possible from the dh_install invocation command line, not in the controlling file [17:41] maxb: yeah, i've read the man page over and over. i'm quite sure i've seen a package that uses a reverse of dh_install though [17:41] somewhere [17:46] RainCT: Unmöglich. [18:00] Btw, are two acks for new packages required or not? https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages says no ("but greatly encouraged"), but I think I saw some other wiki page which said it's mandatory [18:05] RainCT: I don't think the REVU process is actually required by policy [18:06] Laney: I'm not talking about REVU (which isn't policy, but non-MOTUs getting acks from 2 MOTUs is), but about MOTUs requiring an ack from someone else [18:06] mm [18:06] like one year ago it wasn't, but I don't know if this has changed or not === geser_ is now known as geser [18:14] RainCT: it isn't a must have but a good advice [18:15] sebner: alright, so it hasn't changed [19:55] when an update is requested by apt/aptitude/synaptic/whatever, does ubuntu/debian have any kind of a mechanism to only download changed files or does it automatically download a complete package? [20:00] terli: in Ubuntu, currently the latter [20:02] crimsun : get me a cattle prod. [20:02] I'm going into shuttleworth's house. [20:02] is there a sane implementation of that other behaviour available somewhere? [20:03] I don't think so. [20:03] what? [20:03] it's a nontrivial thing, really [20:15] df -m shows 1181 free. aptitude install python-gtk2 ... "Need to get 3281kB of archives. After unpacking 12.0MB will be used." after that, df -m shows 1199. difference of 18. I can understand the 3.281 got saved somewhere, but that only accounts for 15.2. [20:15] where did the other 3mb go? [20:17] laga: I read a description of 'that' - it would require the server extract the diffs real time. which would load the cpu/ram and max out the server faster than the current way maxes out bandwidth. [20:18] at least thats what I got from it [20:18] if that's true or not is someone elses problem :) === jokaro is now known as jokaro_ === jokaro_ is now known as jokaro [22:58] would someone look over my merge please https://bugs.edge.launchpad.net/ubuntu/+source/azureus/+bug/311858 [22:58] Ubuntu bug 311858 in azureus "New Debian version needs merge" [Undecided,Confirmed] [22:59] Can I suppress the update-maintainer warning when building a source package? [23:04] marnold: I've only looked quickly, but I see some strange changes in your debdiff from Debian, such as dropping azureus-gcj. Wasn't that done in the Debian upload? [23:05] hmm maybe i did the debdiff wrong [23:05] which diff [23:05] debdiff debian.dsc your_merge_candidate.dsc [23:06] the second one [23:06] ok [23:06] It's a good idea to run lsdiff on your debdiffs just before uploading to see if they look right [23:06] i'll do it again and reupload [23:07] i did but i was working to try to combat sleep problems so i might not have done my best [23:18] Laney, corrected diff attached [23:20] marnold: Does azureus not use a patch system? [23:21] marnold: The changes to debian/rules aren't in the changelog [23:21] no [23:22] marnold: And the %f in the desktop file isn't explained afaics [23:22] thats because i took debian/rules from Debian [23:22] marnold: But the changes are showing up in your diff [23:24] previous uploader did the changes directly to the source directly without dpatch [23:25] so that is why the diff with debian looks the way it does [23:27] debian/rules was intended as a copy of Debian's as ours is cluttered with gcj changes [23:27] hang on [23:31] Laney, while I'm doing these corrections should i convert the multiuser patch to dpatch [23:32] or leave alone for now [23:33] imo we should get the thing building before we worry about that [23:35] i can always do a cleanup upload later [23:41] marnold: Right, it seems to have other changes directly on the source, so leave that there for now [23:41] marnold: But you need to explain the other things (rules, desktop) in the changelog or get them out of the diff [23:42] Note that these are just general comments - I know nothing about azureus specifically. ubuntu-java (I think) deals with java stuffs [23:42] rules adds a gconf scheme that we want to keep [23:42] almost forgot about that [23:43] marnold: There's the ant heap space change, and dh_gconf/link that were added [23:43] and %f in the desktop [23:44] azureus.template changes - are these covered by something you already mentioned? [23:46] * azureus.desktup: %U -> %f (otherwise file associations don't work) [23:46] Thanks to Francisco Reverbel for pointing this out [23:47] marnold: You need to mention all the changes you keep in your new changelog entry [23:47] User interface auto selection patch [23:47] Laney: what update-maintainer warning were you talking about? did you mean the dpkg-source error rather? [23:48] Adri2000: Yes, but I bodged around it [23:48] I coudn't see anything in the manpage of dpkg-source [23:48] It'd be nice if submittodebian would filter that change out [23:49] are you using debuild or raw dpkg-buildpackage [23:49] debuild [23:49] -S [23:49] as i recall dpkg-bp only warns [23:49] hm, interesting [23:50] debuilld actually enforces it [23:51] dpkg-source errors if $DEBEMAIL contains @ubuntu.com, otherwise it just warns [23:51] so a workaround is to use DEBEMAIL='' debuild/whatever [23:52] UTSL http://patches.ubuntu.com/d/dpkg/dpkg_1.14.23ubuntu1.patch :) [23:52] \o/ [23:53] thanks [23:56] Adri2000: works fine for me? [23:56] oh, never mind [23:56] Hobbsee: ? [23:56] :) [23:57] i thought you were meaning in general