[01:32] tumbleweed: Not quite. Things were different in those days. [01:33] tumbleweed: Security updates ere done by handing a source to the security team, who uploaded it to a dak instance, built everything there, then uploaded source+binaries to LP. So most secuirity uploads from back then probably have oddly named changes files :/ [03:17] tumbleweed: requestbackport is looking awesome. i just ran through all the list of pending-but-approved backports and it came up with the right information for all of them [03:18] remaining issues: (a) it looks like it's including reverse-suggests. i don't expect people to test recommends or suggests, so those should be ignorable [03:19] (b) now that we have the checklist, the only other testing that needs to be documented is b/i/r for the package itself...which we could probably do by adding entries to the checklist itself [03:19] and then we can probably drop the "Testing performed" section entirely [03:25] tumbleweed: hmm, possibly (c) if the user specifies a non-existent source package, check to see if it's a binary package name before bailing (and adjust appropriately) [08:03] wgrant: suprised I haven't run into more of them, then. I'm scraping all upload history, for udd.debian.org [08:12] tumbleweed: Interesting. Perhaps they sometimes were _source.changes despite containing binaries. [08:14] oh, I just checked my logs. Yes, there were many [08:14] the practice seems to have started around dapper [08:15] Laney: um, that's a problem, I guess we need to survive without them [08:15] oh, right, we are. nm [08:29] broder: I didn't filter suggests, enhances, etc, because I thought they were worth mentioning. Should I mention them, but not as checklist items, or add a sentance to the instructions saying they can be ignored? [08:38] I can't see enough information in the LP API, to locate the source package that built a binary, but we can ask apt === bulldog98_ is now known as bulldog98 [09:23] broder: how's that? (r1223) [10:18] morning [11:12] Hi all... I'm trying to use requestsync but I've some problem :/ [11:12] Laney: what is it? [11:12] err l3on [11:12] Please edit the report and give an explanation. [11:12] Not saving the report file will abort the request. [11:12] Laney, broder: happy with requestbackport ? [11:12] what is a report exactly? [11:13] I mean.. which kind of info I've to report there ? And, is there an example somewhere? [11:13] l3on: requestsync is going to file a bug report. The report is the bug report that it'll file [11:13] l3on: https://wiki.ubuntu.com/SyncRequestProcess [11:14] I'm reading it.. [11:14] What I understand is: [11:14] 1. You can use LP web iterface or reqeustsync [11:14] if you want use requestsync use command in this way: [11:14] blabla blalba blala [11:14] So.. what's wrong? :) [11:15] are you saying you can't read the wiki page because it's too long? [11:15] tumbleweed, maybe [11:15] manage-credentials?! [11:15] Laney: that may still be necessary on lucid... [11:16] tumbleweed: haven't tried it, sorry [11:16] Laney: why do you think I'm prodding you :) [11:17] l3on: "Content of a sync request" answers the first question you asked [11:17] how about you tell us what you are trying to file a sync request for? [11:17] apt-cacher [11:18] and thanks for input :) [11:18] so, we currently have an ubuntu-delta for apt-cacher [11:18] what do you mean with ubuntu-delta ? [11:19] we've changed it in Ubuntu [11:19] the current version is 1.6.12ubuntu1. The "ubuntu" means it's been modified [11:19] I know the basic tumbleweed :) [11:19] The last upload was a merge, so the current state of the delta should be summarised in the changelog https://launchpad.net/ubuntu/+source/apt-cacher/1.6.12ubuntu1 [11:20] Changes are now in debian [11:20] fantastic, thesn please file a sync request and say so :) [11:20] requestsync --lp apt-cacher [11:20] but I'm not sure about a file "apt-cacher2" that's in 1.6.12-1ubuntu1 but not in new version 1.7.1 [11:27] looks like apt-cacher2 moved to apt-cacher, and doesn't include the changes we have to apt-cacher2 [11:33] tumbleweed, mmm [11:33] sure? [11:33] I see a NEWS somewhere... [11:34] line 89 [11:40] tumbleweed, in debian changelog I can read: [11:41] * Add changelog and NEWS.Debian to installer_files_regexp [11:44] ah, right, yes, I think that covers it [11:44] Ok I proceed :) [11:50] tumbleweed, maype there's a lack on debug symbol package [11:52] ah, yes. You'll need to do a merge. And please pass that patch to the debian maintainer [11:52] no [11:52] package_files_regexp includes ddeb [11:52] yes you're right [11:52] d|deb [11:53] (u|d)?deb [11:53] yep :) [12:00] tumbleweed: sorry, we hit my too-tired-to-be-useful point about 2 hours ago. i'll take another look after catching some shuteye [12:01] broder: err yes, why are you awake? [12:01] not really sure :) [12:01] tumbleweed, thanks for help... sync reported :) [12:02] l3on: seeing as I've already reviewed it, I'll sponsor it [12:02] thanks again :D [12:02] bug 889448 [12:02] Launchpad bug 889448 in apt-cacher (Ubuntu) "Sync apt-cacher 1.7.1 (universe) from Debian unstable (main)" [Undecided,New] https://launchpad.net/bugs/889448 [12:03] there are a fair number of open bugs for it in LP. Do you think any of them are fixed by the new version? [12:04] let me see [12:07] tumbleweed, this 83987 is fixed in debian [12:07] bug 83987 [12:07] Launchpad bug 83987 in apt-cacher (Ubuntu) "apt-cacher doesn't know about Translation-[lang].bz2 files" [Medium,Confirmed] https://launchpad.net/bugs/83987 [12:09] tumbleweed, this bug 219095 is fixed [12:09] Launchpad bug 219095 in apt-cacher (Ubuntu) "apt-cacher: keep getting 400 No Request Recieved" [Low,Confirmed] https://launchpad.net/bugs/219095 [12:10] neither of those were explicitly closed in the debian changelog (only the debian versions of the bugs) [12:10] Can you make a comment in the sync request bug, with the list of bugs to close, when it's synced? [12:10] of course! [12:11] great, then we can close them, then [12:12] tumbleweed, It's right: [12:12] According to debian bug system, these bug could be closed because fixed in debian: [12:12] * https://launchpad.net/bugs/83987 [12:12] * https://launchpad.net/bugs/219095 [12:12] Launchpad bug 83987 in apt-cacher (Ubuntu) "apt-cacher doesn't know about Translation-[lang].bz2 files" [Medium,Confirmed] [12:12] Launchpad bug 219095 in apt-cacher (Ubuntu) "apt-cacher: keep getting 400 No Request Recieved" [Low,Confirmed] [12:13] or more info is needed ? [12:14] sounds fine [12:16] comment filed :) [12:18] thanks. Hopefully I'll remember to close them when the sync is processed. [12:18] wait [12:18] :) [12:18] there's another bug :P [12:19] bug 366293 [12:19] Launchpad bug 366293 in apt-cacher (Ubuntu) "Non-existent i18n files lead to needless server failover" [Undecided,New] https://launchpad.net/bugs/366293 === yofel_ is now known as yofel === Quintasan_ is now known as Quintasan [16:29] Someone knows why "LDFLAGS += -Wl,--no-as-needed" should be important in debian/rules? :) [16:30] it shouldn't [16:30] you only need that in very rare cases [16:30] I'm trying merge gnome-phone-manager [16:30] and it does it [16:31] its better to fix the build to work with as-needed [16:31] It's the only change, others are now in Debian, and I don't know if I've to request a sync or not [16:31] does it build in ubuntu? [16:31] this is the build without any changes: [16:31] http://debomatic.debian.net/precise/pool/gnome-phone-manager_0.68-1ubuntu1/gnome-phone-manager_0.68-1ubuntu1.buildlog [16:31] of course! [16:31] and are all plugins fully linked? [16:32] aka does it run [16:32] Let me try :) [16:33] do a build without that flag and check dpkg-shlibdeps [16:33] The build I showed you is without that flag [16:34] seems fine then [16:34] to be sure check ldd -r of all shared libraries [16:34] if thats ok it can be synced [16:35] I can't install it damn it [16:36] because I'm on oneric right now [16:37] jtaylor, some suggestions ? :) [16:39] why can't you install it in oneiric? [16:39] depends on [16:39] gnome-phone-manager depends on libebook1.2-12 (>= 3.2.1); however: [16:39] Version of libebook1.2-12 on system is 3.2.0-0ubuntu1. [16:39] gnome-phone-manager depends on libedataserver1.2-15 (>= 3.2.1); however: [16:39] Version of libedataserver1.2-15 on system is 3.2.0-0ubuntu1. [16:39] did youbuild it for oneiric? [16:40] no precise [16:40] if it isn't installable in precise - no sync [16:40] to test you can build in oneiric [16:40] so far I know there are no toolchain differences that are relevant [16:41] but ifyou are going to be doing any reasonable amounts of work on Ubuntu, you probably want to set up local chroots / pbuilder, so that you can test things like that [16:41] tumbleweed, I know ... I'll do it :) [17:21] jtaylor, bug 889563 [17:21] Launchpad bug 889563 in gnome-phone-manager (Ubuntu) "Sync gnome-phone-manager 0.68-1 (universe) from Debian unstable (main)" [Undecided,New] https://launchpad.net/bugs/889563 === Guest7778 is now known as JackyAlcine [21:43] Hi all, I'm receving this error during merge cdebootstrap [21:44] error: ignoring return value of 'fgets', declared with attribute warn_unused_result [-Werror=unused-result] [21:44] And in debian/rules I have: [21:44] CFLAGS_DEB = -Wall -W -Werror -ggdb [21:44] What's the best way to fix it? [21:44] remove -Werror or control fgets return value ? [21:46] well, it's a bug. It should be fixed upstream by checking the return value, and appropriately handling error situations [21:47] tumbleweed, and for now what you suggest to do ? [21:47] if it doesn't look particularly dangerous, then -Wno-error=unused-result may be the easiest solution for Ubuntu [21:47] ok, i'll try :) [21:48] tumbleweed, take a look at this: [21:48] CFLAGS_DEB = -Wall -W -Werror -ggdb [21:48] CFLAGS_STATIC = -Wall -W -Werror [21:48] CFLAGS_UDEB = -Wall -W -Werror [21:48] I've just to edit CFLAGS_DEB ? [21:48] I assume those apply to the three binary packages it builds [21:48] so probably all of them [21:49] ok, thanks.. I'll try :) [21:50] you probably want to add support for precise while you're there [21:51] did you notice bug 884185? [21:51] Launchpad bug 884185 in cdebootstrap (Ubuntu) "Please merge cdebootstrap 0.5.8 (universe) from debian unstable" [Undecided,New] https://launchpad.net/bugs/884185 [21:51] I knwo and I did :) [21:51] oh wow! But it's not reported in m.u.com/universe ! [21:51] well :) [21:52] that was two weeks ago [21:52] dupondje: still working on it? [21:53] the previous uploader is considered responsible for merging into the next release, and it's a good idea to ask before starting a merge. But if you don't, the worst you can do is waste your time [21:53] *before starting someone else's merge [21:55] btw, if you are looking for packages that are most in need of merging: http://qa.ubuntuwire.org/oldmerges/ and http://qa.ubuntuwire.org/bugs/rcbugs/precise/ [21:55] tumbleweed, thanks! :) [21:57] tumbleweed, well... learn me to use that websites :) [21:58] ops.. s/learn/teach/ :) [21:58] which one? [21:59] The oldmerges are merges older than 90 days, it's clear :) [21:59] so I can take one of these without carry about last uploader [21:59] but.. second one? [21:59] right, of course some of the onse at the top of oldmerges are rather hard / not actually mergeable [22:00] but many others are just neglected [22:00] The rcbugs page lists packages that have had RC (release critical) bugs fixed in Debian, but we haven't got those fixed versions yet [22:01] and I can take it without contact old uploader ? [22:01] the bug fixed may not affect Ubuntu, but they're a good way to find easy improvements for ubuntu [22:02] well cdeboostrab now builds :) [22:02] contacting the old uploader is never a bad idea, but nothing requires you to [22:02] you can also look a tthe upload history, to get afeeling for how much someone cares about a package [22:03] hi ! [22:04] started with cdebootstap, but did some bugreporting in debian first [22:04] to get it fixed there [22:06] l3on: its indeed build with -Wno-error atm [22:06] thats quite fine, but they should fix it upstream also, anyway :) [22:07] also add Precise to dists [22:07] and your fine [22:07] feel free to upload the merge [22:07] Ok :) [22:07] I edited how tumbleweed suggested [22:07] and it builds [22:08] well its currently also build with additional options for those build errors [22:08] I've to contact upstream ? [22:10] you could, I already asked them to add oneiric & precise [22:10] not the build errors yes [22:10] Also you need to check if NO_PKG_MANGLE is still needed [22:10] l3on: we like to push everything upstream whenever possible. Maintaining deltas (differences) in Ubuntu is painful (it requires merging, and we can't keep up with all the merges) [22:11] dupondje, how can I do it ? [22:11] tumbleweed, thanks :) [22:12] l3on: well there is where I got stuck ... :D [22:12] lol [22:12] here the buildlog: http://debomatic.debian.net/oneiric/pool/cdebootstrap_0.5.8ubuntu1/cdebootstrap_0.5.8ubuntu1.buildlog [22:12] l3on: you should be building for precise, not oneiric [22:12] also, that's the failed build [22:12] gosh.. sorry, wrong url [22:13] http://debomatic.debian.net/precise/pool/cdebootstrap_0.5.8ubuntu1/cdebootstrap_0.5.8ubuntu1.buildlog [22:13] :) [22:15] so, you want to know if NO_PKG_MANGLE is still needed [22:15] if you read the ubuntu changelog, you can see why it was added [22:15] I would know what is NO_PKG_MANGLE and I'm googling it :) [22:16] hrm, don't know if that's well documented anywhere [22:16] we have a package called pkgbinarymangler, that we install in our build chroots (although debomatic doesn't appear to have it. DktrKranz?) [22:17] it allows us to mangle the packages a little during the build, for the things that we do differently to debian [22:17] its for translations I think [22:17] but not totally sure [22:18] e.g. it sets the maintainer to "Ubuntu Developers", removes translations, yes, optimises PNGs, removes changelogs https://launchpad.net/pkgbinarymangler [22:19] and how you know if we need it ? [22:19] oh, sorry debomatic does have it. I missed that [22:19] well, cdebootstrap is a bit odd. You'll notice the debs it builds contain debs [22:19] DktrKranz: unping, nm [22:21] yep, it's right I've debs inside deb [22:21] http://debomatic.debian.net/precise/pool/cdebootstrap_0.5.8ubuntu1/cdebootstrap_0.5.8ubuntu1.contents [22:21] the changelog entry that added NO_PKG_MANGLE says: [22:22] helper/Makefile.am, helper/Makefile.in: Set NO_PKG_MANGLE while building nested packages so that pkgstriptranslations doesn't kick in. [22:22] I don't think anything would have changed there, that's still a potential problem [22:23] you can, of course, forward that patch to Debian too, it will do nothing in Debian, and make Ubuntu's life easier [22:23] that has been forwared to debian [22:23] and rejected btw [22:23] so :) [22:24] lol [22:25] right, debian bug 486899 [22:25] Debian bug 486899 in cdebootstrap "cdebootstrap: Please set NO_PKG_MANGLE while building nested package" [Wishlist,Open] http://bugs.debian.org/486899 [22:34] well tumbleweed, if I understood.. NO_PKG_MANGLE should allow us to not have nested packeges [22:35] but... using it I've still debs inside debs [22:35] That's new contents: [22:35] http://debomatic.debian.net/precise/pool/cdebootstrap_0.5.8ubuntu1/cdebootstrap_0.5.8ubuntu1.contents [22:36] l3on: the problem is that debian package building is modified by the presense of pkgbinarymangler. But this can be disabled by exporting NO_PKG_MANGLE [22:36] we don't want those modifications on the nested packages [22:36] this is the patch : [22:36] http://paste.ubuntu.com/736721/ [22:36] that looks right [22:36] ok :) [22:36] i'll debdiff and upload the bug :) [22:37] thanks (again!) :) [22:42] ok, done bug 884185 [22:42] Launchpad bug 884185 in cdebootstrap (Ubuntu) "Please merge cdebootstrap 0.5.8 (universe) from debian unstable" [Undecided,New] https://launchpad.net/bugs/884185 [22:42] thanks tumbleweed dupondje :) [22:48] + - Patch 002-Fix-netsted-packages.patch setting NO_PKG_MANGLE while building [22:48] typo [22:54] dupondje, mmm... is it important ? :/ [22:54] I need to redebuild ?! [22:56] and I see the series seems adjusted in the code [22:56] and there is also a patch included ... [22:56] double type, patch name heading with 0002_... [22:58] s/type/typo/ [22:58] ok.. maybe it's time to go bed :) [22:59] dupondje, suggestions ? :) [23:15] well I've to go.. bye bye and thanks for all :) [23:15] see you :) [23:50] tumbleweed: ok, actually taking a look at the requestbackport changes. i would like the installability checkbox to be per-binary package, to make it explicit that people test all of them [23:50] (i screwed that up once in the past)