=== YDdraigGoch is now known as WelshDragon === chuck_ is now known as zul [05:50] hiş [06:09] if I want the installation of a package to remove another package if it exists, do I use Replaces, Conflicts, or both? [06:10] I'm finding http://www.debian.org/doc/maint-guide/ch-dreq.en.html a little ambiguous on what a packager manager should actually do with those [06:10] mrooney: from what I understand, Conflicts to express "Only one of the two should be installed at any given time", and Replaces+Conflicts to express "Users of previous package foo should be upgraded to bar" [06:14] jdong: okay, the example I'm playing around with is a metapackage to change the browsing experience, so I've got Depends: midori, and then I should Replaces+Conflicts firefox? [06:15] mrooney: I'd say just Conflicts firefox [06:15] if you want Firefox to be removed [06:16] ah okay, yes [07:04] hmm, are they are guides on debian/.install files? I can't figure out where it finds the files; I assume the lines are where it puts them [07:04] (I'm looking in https://wiki.ubuntu.com/PackagingGuide/Complete) [07:07] mrooney: man dh_install [07:08] jmarsden: awesome, thanks! perhaps it might be appropriate for me to add that to the wiki? [07:09] Well, in general all the debian/stuff files have corresponding dh_* commands and so man pages... but sure, if you think it will be useful to add it, go for it. [07:10] (all the debhelper bits --> debian/stuff) [07:10] Right... I wasn't clear... man debhelper to see more details... [07:14] oh hey dtchen, sorry for changing the importance on that bug; didn't realize it was part of your workflow [07:16] jmarsden: maybe I need to take a step back. I have an XML file I need to load in a postinst script, so I'm trying to ship it in /usr/share so I can do that, and was looking at the .install files as a means to do that [07:17] mr no worries [07:17] wow, uh, tab completion fail [07:17] dtchen: you use irssi? [07:18] lifeless: indeed [07:18] dtchen: if so, its a known 'feature' - irssi detects 'paste' by the rate at which characters are received, and if it thinks something is a paste [07:18] it doesn't interpret it [07:18] I think this is daft [07:19] but [07:19] (I filed a bug once) [07:19] mrooney: So, if you put the XML file under debian/ and then put a line in debian/install that says xmlfile usr/share/packagename doesn't that do what you want? [07:22] jmarsden: hm, could be :) this is I think my second-ever package [07:22] let me check that out [07:22] Sure, go for it :) [07:24] be aware that xmlfile needs to get into $(CURDIR)/debian/tmp/usr/share/packagename; some people do this in debian/rules, etc., etc. [07:25] dtchen: if you run irssi over ssh, a network glitch can cause what you had happen; otherwise just typing really really fast ;) [07:25] lifeless: yeah, client was extremely lagged [07:27] dtchen: oh I see so THAT'S how the .install files work? I was confused because I only saw what I perceived as a destination, but not a source file [07:28] but actually you sort of mirror the path in debian/tmp and each line is the source and destination in the .install file? [07:28] mrooney: there are different semantics based on the syntax [07:29] hm okay, know any simple packages that do this off-hand that I can look at the source? [07:30] I will just google dh_install, maybe it has more examples than the man page [07:33] libao does [07:33] see its libao2.install for an example of both, intermingled [07:35] oh awesome, I see, thanks! [07:37] so that's grabbing a .conf file from debian/ and shipping it to /etc, is it acceptable to just put arbitrary xml files in debian and then ship them that way? [07:38] I don't know about arbitrary -- I would think that they're at least Debian-specifi [07:38] c === Lutin__ is now known as Lutin [10:02] Hi people, what's the correct way to get a GConf schema file installed. [10:02] ? [10:03] Is this more of an upstream thing? Everything that I'm reading is telling me to put it in the post install but surely a `make install` should take care of this. [10:05] I realise that this might not be a MOTU question exactly but it might become a MOTU issue when I get it wrong ;) [10:07] lamothe: If you install the application "by hand" after building it from the upstream source tarball, how does the Gconf schema get where it needs to be? [10:08] schemadir = $(GCONF_SCHEMA_FILE_DIR) ... is that what you're asking? [10:09] No, I mean if you follow whatever install instructions upstream provides with their software, how does that file get installed? [10:10] What puts it into place? [10:10] jmarsden: Oh, I am upstream. [10:10] Currently I don't have a schema file. Would like to add one. [10:10] So in an unpackaged situation... how are you getting the file into place? get that working, *then* package your software :) [10:11] jmarsden: I'm not, so I did some reading and it seemed to be pointing me at putting it in the post install which didn't seem right to me. [10:12] Before I go down a path that a MOTU will tell me is wrong later, I was looking to try and get it close to correct first time :) [10:12] OK... I think you are correct; if the software doesn't work / install correctly before it is packaged, it should be fixed... [10:12] But I am not a MOTU :) [10:13] But I bet that I'm still doing it wrong :) [10:14] Ok, so I need to find a solution in the Makefile.am. Are the install-data-local sections frowned upon? [10:22] lamothe: dh_gconf does this [10:22] I realize that this is not a support channel, but my Brasero has stopped burning CDs in a fresh Karmic install, it only burns DVDs. [10:23] Laney: Thanks,I'm on dh7. [10:23] When I burn a CD, it acts as if the burn completed successfully, but when I reinsert it, it's shown as empty. [10:23] I hope that this works, although isn't most of this automatic now? [10:24] yes, it should be called in the install target [10:24] export DH_VERBOSE=1 [10:25] you have to arrange for the schemas to be installed in usr/share/gconf/schemas first [10:25] Laney: But surely I need to somehow instruct it ... right. [10:26] Won't $(GCONF_SCHEMA_FILE_DIR) do that? [10:26] isn't that stuff from cdbs? [10:26] just use dh_install [10:27] I have "schemadir = $(GCONF_SCHEMA_FILE_DIR)" in the Makefile.am. [10:27] no idea [10:27] see how some other software does it [10:28] Then ... "schema_DATA = me-tv.schemas" [10:29] Yeah, I was looking at gnome-terminal. I find it hard to find a simple GTK/GNOME application using DH7. [10:29] And a schema file. [10:29] Anjuta doesn't even have a scema file. [10:29] you're talking about the upstream autofoo [10:29] the debian buildsystem shouldn't matter [10:30] as long as "make install DESTDIR=debian/tmp" does it right [10:30] Laney: Yes, this is a upstream (my) problem. But I was hoping not to make it someone else's. [10:30] that's what I'm saying [10:31] so you don't have to care about if it's dh7 or not [10:31] so you *can* look at any old bit of software [10:31] once you get your autofoo doing the right thing then dh7 will also just work [10:32] Yes, but it's how it gets it in there. Also, I noticed that gnome-terminal installs its schema on a `make install`/ [10:36] most of them should [10:36] but it's beyond me how to wrangle the autotools into doing so [10:36] i'd go hassle someone who has done it before :) [10:42] Laney: Sorry, there's a nasty bug in karmic, when I plug in the power on my laptop it shuts down. === asac_ is now known as asac === yofel_ is now known as yofel [14:17] I am looking at a package that used to FTBFS because it was compiled with -Werror we patched it to fix the errors. Upstream has released a new version, they did not fix the warnings but they removed -Werror so it builds fine now [14:18] the error was: ignoring return value of 'write', declared with attribute warn_unused_result [14:18] or warning rather [14:19] so my question is if I should tear out the old patch since it isnt needed to fix the FTBFS or keep it since it is a better coding practice? === ikt_ is now known as ikt [14:57] if there are no other changes, just drop it and sync the package [15:00] but the patch could probably be sent upstream (if you care enough) === nixternal_ is now known as nixternal [16:11] where do makefile.in's get there variables from? [16:40] EzraR: configure [16:40] EzraR: or, rather, configure.ac [16:44] yeah i figured it out [16:46] this bug (LP: #327174) ends up steming from this bug [16:46] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=524176 [16:46] Debian bug 524176 in automake1.10 "AM_PATH_PYTHON should honor python's idea about the site directory" [Normal,Open] [16:46] i believe [16:46] anyone wanna make a second opinion? [16:47] i have the automake book sitting around...i should really get a round to reading it... [16:56] why nv cards crashs in x with ubuntu kernel? [17:04] Microsoft bribes Ubuntu development team to make bullshit kernel to not run 3d cards or commercial games on Linux. [17:05] Hi. Any MOTUs available to review my package on REVU? http://revu.ubuntuwire.com/p/gpxviewer - It's an application to view gps traces. Thanks :) [17:07] AndrewGee: I'll bite [17:08] AndrewGee: Ah, you're also upstream [17:08] Yup [17:08] Thanks for taking a look :) [17:20] hi all [17:21] !op [17:21] Help! Hobbsee, Riddell, sladen, fbond, mneptok, gnomefreak, Seveas, dholbach, elkbuntu, PriceChild, or jpatrick! [17:21] * sladen looks [17:22] Pici: ^^ [17:22] mok0: Thanks for the advocation. Do I upload a change to the standards version after some other MOTU has advocated then? [17:23] Andrew, Microsoft bribes Ubuntu development team to make bullshit kernel to not run 3d cards or commercial games on Linux. [17:23] sladen: see #u-d [17:23] AndrewGee: yes, the next advocate is likely to ask for some changes, then you can batch it in. If the second advocate does not have further issues he can make the modification before upload [17:23] mok0: Awesome! Thanks :) [17:26] thank you jpds :) [17:26] No problem. === neoXsys is now known as neoxsys [17:34] jpds: well, I guess that ends my chance of talking to the person [17:34] jpds: despite them not shouting any more abuse since I took it to privmsg [17:35] sladen: He attacked #ubutnu, #ubuntu-devel, #kubuntu, #kubuntu-devel and #ubuntu-ops. [17:35] jpds: and nobody (until I, apparently) bothered to ask the person specific *details* about what their issue was [17:36] jpds: admittedly it's not pleasant, but neither is being klined [17:36] sladen: http://paste.ubuntu.com/325505/ [17:41] jpds: *shrug*. I'd got a little further; but I didn't start of with telling them they were talking nonsense [17:42] jpds: although I do admit to being not entirely sure how to answer "is canonical a catholic enterprise?"---though the problem was sholved before I'd tried to [17:42] sladen: I did point him?her? at the bug tracker, but I suppose I should have been more precise. [17:43] catholic means universal, doesnt it? [17:44] wikipedia says yes: Catholic is an adjective derived from the Greek adjective καθολικός (katholikos), meaning "universal". [17:50] I wonder if it was meant as "closed shop"---I didn't have time to enquire [20:55] anyone in here familiar with the problem with automake and python.m4 setting wrong dirs? === menesis1 is now known as menesis === _stink__ is now known as _stink_ [21:56] lintian is outdated in REVU -> W: opensimulator-server source: newer-standards-version 3.8.3 (current is 3.8.0)