=== timrc-afk is now known as timrc === freeflying_away is now known as freeflying === freeflying is now known as freeflying_away === mwhudson is now known as zz_mwhudson [03:26] thanks for the syncs Logan_. There should be a fair few more there [03:26] And another SRU [03:27] Noskcaj: lol, you should be happy I'm doing this now. first day of the new semester tomorrow :O [03:28] :) [03:28] I've got another week still [03:28] unfortunately there aren't any Ubuntu packaging classes I could TA [03:28] it's a shame [03:28] Are there any ubuntu packaging classes ever now? [03:29] if there were one, that would be awesome [03:29] make it happen! [03:30] i'll try. First i need all the xubuntu 14.01 stuff sponsored and to finish adopting xsane === freeflying_away is now known as freeflying === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying === freeflying is now known as freeflying_away [05:47] Good morning === zz_mwhudson is now known as mwhudson === freeflying_away is now known as freeflying [07:58] pitti, provisioning of VMs for adt is working with kernel -5 [07:58] jibel: indeed, I verified this morning [08:07] good morning [08:11] erk, since when do we switch consoles using Alt+Left/Right? that's very confusing, and I use these for other things already [08:12] is that a kernel thing, or X.org? [08:12] apw, RAOF ^ [08:13] morning [08:32] stgraber: if I may bother you again: I use apt-cacher-ng on my workstation; is it possible to tell lxc to set an apt proxy for creation of a container? or is that something you'd set up manually for each container after its created? [08:47] pitti, Courtesy of unity default keybindings I would guess. The kernel does not care about that [08:48] smb: I can switch between VTs with Alt+left/right as well [08:49] (nothing related to unity keybindings) [08:49] gah, it's only now that I realize how often I use Alt+left/right [08:49] switching between channels in IRC, going back and forth in firefox [08:49] I'm fairly sure this is new from today [08:50] or new with kernel -5, I was running the saucy kernel yesterday afternoon to test something else [08:50] I'll bisect [08:50] pitti: I didn't update yet and alt + left/right works here to switch IRC channels [08:50] pitti: let me see what I didn't update since yesterday [08:51] pitti, I woul d be pretty sure that the changes between those kernels did not contain something to change that [08:51] pitti: http://paste.ubuntu.com/6796175/ [08:51] quite a lot [08:51] want me to try upgrading to latest kernel first? [08:52] didrocks: yeah, fortunately most packages are unrelated to that [08:52] didrocks, that at reast would rule out the that or not [08:52] didrocks: if you want; I'll try downgrading the kernel first, and then some other updates [08:52] pitti, using -5 kernel here and Alt+Left/Right is working as expected [08:52] pitti, Thats still installed, no? [08:52] pitti: as you wish, I can apt-get install linux-image [08:52] smb: "still installed"> I'm running that, if you mean that [08:53] pitti, I mean downgrading to -4 [08:53] Both should be installed if you upgraded [08:53] via grub [08:53] smb: apt-get autoremove already cleaned it up, but I'll get it from LP [08:53] sec [08:54] apw, Isn't autoremove supposed to leave at least the previous one? [08:55] smb, didrocks: ah, a mere reboot fixed it apparently (without changing packages) [08:55] the last three, plus the current kernel if it isn't one of those [08:55] so it was either some funny thing that qemu or LXC did this morning, or some transient thing from dist-upgrading without rebooting [08:55] smb, apw: I had -5 installed and running, plus the saucy kernel, so -4 got autocleaned [08:56] which I was quite happy about, given how leaky -4 was [08:56] perhaps booting precise in LXC does some funny console setup or so? [08:56] pitti, It surely was not the best one. Though I would not expect autoremove to be that smart ;) [08:56] smb: ah, it doesn't have a "remove dangerous kernels" mode? :-) [08:57] no indeed :) [08:57] reminds me of http://www.youtube.com/watch?v=ytVdBLMmRno [08:57] (watch it till the end) [08:57] perhaps you were running the saucy kernel when you autoremoved [08:57] pitti, Maybe if we flag them as broken when we send them out. Like all those we do especially to check whether qa is paying attention. ;-P [08:59] apw: no, that was this morning when I ran -5 [08:59] pitti: oh :) [08:59] apw: anyway, I had the current, running, and a previous kernel [09:01] pitti, re proxy for LXC, set HTTP_PROXY to your apt proxy or to 'apt' [09:01] pitti, ie sudo HTTP_PROXY=apt lxc-create .... [09:01] jibel: ah, nice [09:01] or or sudo HTTP_PROXY="http://.../" lxc-create.... [09:03] hm, doesn't seem to work [09:03] I can stop apt-cacher-ng and it still works; and there's nothing in the container's apt.conf.d/ to set it? [09:03] jibel: ^ [09:04] jibel: or perhaps this doesn't work as it's using a cached base system already [09:04] pitti, yes, you may need to flush the cache with option -F of the template === mwhudson is now known as zz_mwhudson [09:05] jibel: trying [09:06] pitti, "sudo HTTP_PROXY=apt lxc-create -n test_proxy -t ubuntu -- -F -d" works here and it creates /var/lib/lxc/test_proxy/rootfs/etc/apt/apt.conf.d/70proxy with the right proxy === doko_ is now known as doko [09:07] jibel: ah, so I suppose it doesn't create that file when using a cached base fs [09:08] jibel: ah, now it's using apt-cacher-ng even during creation [09:08] jibel: merci [09:08] stgraber: ^ [09:10] jibel: ah, HTTP_PROXY=apt would use 127.0.0.1, which doesn't work in a container; but it proves that it did configure a proxy :) [09:11] * pitti updates it to 10.0.3.1 [09:12] pitti, right, =apt gets the configuration from apt, alternatively you can specify the IP or hostname of your host instead of apt [09:12] if it is on the same machine [09:13] ah, hostname sounds like a nice idea, avoids hardcoding the IP [09:13] so 70proxy is *not* in /var/cache/lxc/precise/, so lxc-create does create it after copying; but apparently not if it's using a cache [09:14] * pitti files a bug, that doesn't sound hard to fix [09:14] hello. it seems cloud-init or something is bricking my machines if I set up my own ephemeral... [09:20] stgraber, jibel: FYI, filed bug 1271455 [09:20] bug 1271455 in lxc (Ubuntu) "lxc-create does not honor $HTTP_PROXY when using a cached base image" [Undecided,New] https://launchpad.net/bugs/1271455 [09:23] yay, now my host ubuntu, schroots, autopkgtest, and now lxc are all using the same apt-cacher-ng; this thing rocks :-) === dosaboy_ is now known as dosaboy [09:51] pitti, any chance you could take a look at https://code.launchpad.net/~darkxst/ubuntu/trusty/gdm/nokill/+merge/202393 [09:52] darkxst: sure [09:52] note debian does not kill gdm in postinst, except in the case where they transition from gdm -> gdm3 [09:52] thanks [09:53] darkxst: hm, not in the recent big upload though, it seems? (http://launchpadlibrarian.net/162697567/gdm_3.8.4-0ubuntu2_3.10.0.1-0ubuntu1.diff.gz) [09:56] darkxst: uploaded [10:01] pitti, thanks, that code has been there for a while, but only just started triggering a kill === j_f-f_ is now known as j_f-f === _salem is now known as salem_ === timrc is now known as timrc-afk === salem_ is now known as _salem === timrc-afk is now known as timrc === _salem is now known as salem_ [11:29] I'm trying to build the suPHP package updated to 0.7.2 with uupdate [11:29] But it seems the libtool patch is causing the build package command to fail [11:30] Can someone look into this? 0.7.1 Has an exploit [11:36] TuxBrother: try filing a bug :) [11:36] Is that an Ubuntu-specific patch? [11:36] I don't know [11:37] What patch then? [11:38] http://www.suphp.org/Home.html < FYI The patch: 02_libtool.dpatch in debian/rules/patches [11:38] mitya57: Do I need to post the contents on pastebin? [11:40] TuxBrother: Err, what package is that patch in? Is suPHP packaged in Ubuntu? [11:40] I thought it was not. [11:41] Ah, it is, let me look then. [12:21] mitya57: Any luck so far? [12:26] TuxBrother: https://launchpad.net/ubuntu/+source/suphp/0.7.2-0ubuntu1 [12:27] Great1 [12:28] mitya57: Any chance to see this backported to precise? [12:29] TuxBrother: please file a bug first. [12:29] Then I'll attach a debdiff for security team review. [12:33] TuxBrother: http://paste.ubuntu.com/6797026/ — this is the relevant part of the diff, right? [12:36] mitya57: I'm not a developer, only a sysop [12:38] I will file a bug report later today [12:39] OK, feel free to subscribe me (~mitya57) there [12:53] zyga: Any reason why you plainbox autopkgtest is redirecting output to /dev/null? It makes it harder to figure out what happens. [12:53] Also, maybe you are going to fix the failure yourself? :) [13:14] mitya57: my ignorance while writing the initial tests, it's on my TODO list, just swamped with other tasks [13:15] mitya57: it prints to stderr and I just need to add that flag to make that safe [13:15] mitya57: if you remember how to do it could you please fix it in debian SVN? I'll make my life easier with the next release [13:17] it'll [13:18] zyga: OK, actually I've already found the failure and am looking at that [13:18] mitya57: what is the failure? [13:19] FAIL: test_perfect_match (impl.exporter.test_html.HTMLExporterTests) [13:19] ---------------------------------------------------------------------- [13:19] Traceback (most recent call last): [13:19] File "/usr/lib/python3/dist-packages/plainbox/impl/exporter/test_html.py", line 142, in test_perfect_match [13:19] self.assertEqual(self.actual_result, expected_result) [13:19] mitya57: hmm, interesting! [13:19] (LOTS of HTML output below) [13:19] mitya57: can you pastebin that somewhere? alternatively, can you open a bug on lp:checkbox with that output? [13:20] Will do [13:20] infinity, something for your "amusement" (as you seem to be touching eglibc quite often): bug 1271534 [13:20] bug 1271534 in eglibc (Ubuntu) "libc6-xen:i386 installation can cause panics on boot" [High,Confirmed] https://launchpad.net/bugs/1271534 [13:20] Let me try assertMultiLineEqual first [13:20] mitya57: many thanks! [13:20] mitya57: perhaps self.maxDiff = None will help [13:22] Ah, that is bytes, not str, so assertMultiLineEqual doesn't much help [13:37] zyga: filed bug 1271542 and attached output there [13:37] bug 1271542 in checkbox "Plainbox self-test failing on test_perfect_match" [Undecided,New] https://launchpad.net/bugs/1271542 [13:40] mitya57: thanks, we'll take it from there [13:40] mitya57: I love how launchpad renders graphics in that log :) [13:43] Yes, my Firefox also thinks it's HTML file ;) [13:44] mitya57: yeah, that test is checking our our HTML inliner injects various resources [13:44] this might be a security issue, we probably can inject js into launchpad.net domain and do XSS [13:51] zyga: here's the first non-matching lines: https://launchpadlibrarian.net/163098330/plainbox.diff [13:52] mitya57: ohhh [13:52] mitya57: I just did the same :) [13:52] mitya57: thanks :) [13:53] mitya57: yeah, I wonder why that happens, I cannot reproduce that failure [13:53] I also can't reproduce it locally, but can in trusty chroot [13:53] I meant in trunk, I'll check out that debian release though, this is curious [13:54] or better, I'll let roadmr do it, I'm busy wity something else :/ [13:54] mitya57: ideally we'd re-render those images and see what they really look like [13:54] mitya57: do you know how to do that? [13:54] They look the same [13:54] mitya57: hmm, maybe base64 is somehow not deterministic? (though that would be odd) [13:54] mitya57: like some junk or seed? [13:54] Just copy-paste the data:... URI into the adress bar [13:55] oh! [13:55] good idea [13:55] No, base64 is deterministic :) [13:56] mitya57: maybe different encodings are possible? python has b64 and standard_b64? [13:57] mitya57: theory, the images are the same, the encoding has junk in both versions, that junk is not used by the the png decoder, the junk is a reused buffer of some kind? === Sweetsha1k is now known as Sweetshark [13:58] * mitya57 looks [13:59] * zyga looks at the source [13:59] zyga: I think when you encode PNG, it is in bytes, not in unicode, so encoding shouldn't matter here [13:59] mitya57: encoding as in "this is how png bitstream looks like" [13:59] mitya57: maybe there's fixed-lenght buffer and the stuff "afte the end" of the real data is irrelevant [14:00] * zyga imagines interesting data injection capabilities in "identical" PNGs [14:01] the part that converts a file to data URI looks sane to me [14:01] zyga: No, the images are not equal. I can spot "Created with GIMP" comment in the second one but not in the first one [14:01] ohhh [14:01] wow, cool [14:02] so somehow, the images got changed [14:02] Ah, pngbinarymangler [14:02] I wonder when that happened, everything that lands to trunk runs the full test suite [14:02] !!! [14:02] *pkgbinarymangler [14:02] cool :D [14:02] why does it do that? [14:02] I mean, why does it strip the comment [14:02] In Ubuntu all PNGs are stripped [14:03] hmmmm [14:03] I see [14:03] I think there is some env variable to disable that [14:03] interesting, thanks for solving this bug then [14:03] mitya57: I'll add this to the bug report [14:04] zyga: export NO_PNG_PKG_MANGLE=1 in debian/rules should fix that [14:04] mitya57: interestingly, we also run the test suite at build time, but obviously adt runs it after the build, that's really tricky thing to find [14:05] Should I commit that to SVN? [14:05] mitya57: do you want to patch it in svn? [14:05] mitya57: yes please! [14:06] mitya57: if you can, do a changelog and I'll ask our sponsort to release it [14:06] changelog bump [14:11] zyga: feel free to ask p1otr now :) [14:12] mitya57: thanks, just sending email [14:14] mitya57: thanks a lot, you fixed that instantly :) [14:16] zyga: You are welcome :) [14:25] ev: bdmurray: cjwatson: looks like 12.04 installer still has the "creepy" webcam step, which was removed in raring and up. Should the webcam step be removed / disabled for 12.04.4, or let it be. [14:26] stgraber: ^ [14:26] xnox: I don't like it, but I would let it be at this stage [14:27] Maybe a month or two ago ... [14:27] * zyga liked the webcam step [14:27] why did we think it's creepy? [14:29] zyga, because photographs steal your soul [14:29] brendand: that's okay, we sold ours to get the mortage ;) [14:30] zyga, and ubuntu installer photographs assign copyright of your soul to canonical [14:31] brendand: don't spread rumours. [14:31] even better, after sending my photo to facebook, google and amazon I really need canonical to own a copy ;) [14:32] xnox: creepy? :( [14:32] * zyga mutters something about CLA, photos and forking [14:32] and gets back to wokr [14:32] zyga: creepy, cause you get to see your own face whilst installing ubuntu =) (and apw used that term) but in fact it's mostly useless as the pics didn't actually propagate anywhere. [14:32] (and places it did propagate to, got removed / changed since webcam introduction) [14:33] xnox: yeah but that's not the reason not to use it, my only complaint was that the "live feed" was much brighter, more naturally exposed, than the picture it took, but I guess that's related to video/still pic transition [14:33] xnox: I'd love if it could be my u1 user photo and if u1 had more features around users later === freeflying is now known as freeflying_away [14:37] and SSO for login! [14:40] brendand: yeah [14:45] jodh: regarding unusable kvm for you; kvm really works splendidly on both my amd and intel based trusty laptops; can you bring the laptop you're having trouble on next week so I can take a look? [14:45] ogra_: hey Oliver, how are you? do you have a sec for an arm kernel install question? [14:45] pitti, sure, shoot (and i'm fine) [14:46] ogra_: so, I dist-upgraded a raring arm calxeda node in the lab from raring to saucy [14:46] ogra_: it didn't come back when booting, so I suspect it might not like the 3.11.0-15 kernel [14:46] ogra_: I did the same upgrade to another node, but didn't reboot yet [14:46] ogra_: I'd like to tell the boot loader to boot with the raring kernel [14:47] ogra_: I suppose I'll change the symlinks in /boot (vmlinuz, initrd.img) to point back to the 3.8 raring kernel [14:47] pitti, ah, i havent dont much with server recently ... usually flash-kernel should do the right thing though ... [14:47] ogra_: is there something I need to runa fter that, like flash-kernel? [14:47] pitti, try dannf or rbasak [14:47] hallyn: are you referring to bug 1268906 or bug 1257352? The former kernel issue occurs on 2 systems I have. I'll be bringing one with me. [14:47] pitti, the package postinst should have run flash-kernel [14:47] ogra_: ack, talking to rbasak [14:47] bug 1268906 in linux (Ubuntu) "cpu soft lockup running kvm" [Medium,Confirmed] https://launchpad.net/bugs/1268906 [14:47] bug 1257352 in qemu (Ubuntu) "kvm hangs occasionally when switching out of the qemu console" [Medium,Confirmed] https://launchpad.net/bugs/1257352 [14:47] ogra_: right, but I suppose that flashed 3.11 (saucy kernel) on dist-upgrade [14:47] yeah [14:49] pitti, i remember a bug where flash-kernel forcefully always flashes the latest kernel ... newer f-k versions use linux-version for that iirc [14:49] jodh: the former [14:49] hallyn: right. It's entirely repeatable with every image I throw at kvm. [14:50] pitti, you might need to remove the 3.11 kernel completely and then run f-k manually [14:51] ogra_: ah, might be safer [14:52] (make sure to clean up /boot manually, iirc there is stuff left behind) [14:53] ogra_: ah, it also seems that it wasn't due to the kernel but due to a net interface problem [14:53] oh [14:53] phew, i thought ppisati tested those [14:58] rbasak, ogra_: FTR, merely changing the symlinks is sufficient; boot.scr has the symlinks and the boot loader reads/resolves those [14:59] good [14:59] Thanks [14:59] * ogra_ wonders why highbank uses boot.scr though ... dont we use uEnv.txt ? [15:00] flash-kernel writes out boot.scr, doesn't it? [15:00] That happened when we synced to Debian. Some time around the 12.10 timeframe IIRC. [15:00] well, for anda we switched to uEnv.txt long ago [15:00] ah, k [15:01] so its lool's fault then [15:01] :P [15:19] pitti: I've never tried that part of the Ubuntu template myself because I use a transparent proxy, but try setting HTTP_PROXY=apt before running lxc-create [15:19] stgraber: bonjour [15:19] pitti: looking at the code, that should query apt-config for the current proxy value and use it for the container [15:19] stgraber: right, that's what I did at last; works except for said bug [15:19] stgraber: not quite as my proxy is 127.0.0.1, but HTTP_PROXY=http://`hostname`:3142 worked [15:20] pitti: ah yeah, sounds like the proxy stuff isn't done at the right stage of lxc-ubuntu... [15:24] stgraber, the proxy is set when sources.list is written ie during initial download _or_ if the arch of the host and the container are diffrent [15:27] stgraber: do you have an idea about http://paste.ubuntu.com/6797756/? [15:27] stgraber: in particular "confused by argument: armhf"? [15:27] stgraber: that's from an arm machine [15:28] which is runnning saucy [15:28] pitti: you're running an old version of the cloudimg query tool which doesn't know about armhf [15:28] pitti: the same on trusty should work (at least it did last week) [15:29] stgraber: ok; I'll try upgrading this node to trusty [15:29] stgraber: or perhaps the cloud query tool first; which package is that? [15:30] cloud-image-utils [15:30] arch: all, how nice [15:30] stgraber: merci [15:56] doko: which package needed porting to bs4 again? [15:58] zul, http://people.canonical.com/~ubuntu-archive/component-mismatches-proposed.svg [15:58] doko: cool thanks [17:03] hallyn: can we wait until the next upstart release (next week) for a fix for bug 1263738 to land in trusty? [17:03] bug 1263738 in lxc (Ubuntu Trusty) "login console 0 in user namespace container is not configured right" [High,Triaged] https://launchpad.net/bugs/1263738 [17:04] jodh: yeah that's fine [17:04] jodh: i wanted it fixed by feature freeze if possible. containers are useful without the fix [17:04] jodh: thanks! [17:05] hallyn: ok, thanks. shouldn't be a problem as we've fixed it upstream now. === jodh` is now known as jodh === zyga_ is now known as zyga === frobware is now known as zz_frobware === zz_frobware is now known as frobware [19:06] infinity, xfce4-weather-plugin 0.7.4-5 is now in s-p-u, is that enough to sync to precise? === fire is now known as Guest63129 === zz_mwhudson is now known as mwhudson === baba is now known as fire_ [19:36] Is there anything i can do do make my motu application happen faster? It's three weeks since i sent the email [19:44] Noskcaj: the DMB isn't particularly good at e-mail applications. I'll poke people again [19:45] tumbleweed, thanks. [19:45] There's a heap of stuff xubuntu needs uploaded, and not many sponsors. === bfiller is now known as bfiller_afk === jackson is now known as Guest4739 === Ursinha_ is now known as Ursinha [20:07] Noskcaj: just saw you want to merge xfwm4 [20:07] is that a stable release? [20:07] its not mentioned on their homepage [20:07] they might be using old gnome versioning, odd unstable, even stable [20:07] so we probably don't want .11 in trusty [20:13] jtaylor, We try ro use the old gnome versioning, but everything is very stable anyway and contains stuff we need for xubuntu [20:13] 4.11 is mostly the same as 4.12, just 4.12 has to have everything released at once, but panel isn't fully ready [20:14] when is it expected to be ready? [20:15] jtaylor, No one really knows. Probably before trusty's final release, probably not before feature freeze [20:20] smb: Ew. Stop flipping your dom0 from Xen to bare metal. [20:20] smb: (But yes, good catch. I wonder why no one else has noticed this for years...) [20:23] robert_ancell, Any chance you could do the packaging for lightdm (and greeter) in debian? We've got a pointless duplication of work currently [20:23] Noskcaj, no, I want to release fast into Ubuntu. You can copy changes into Debian [20:24] ok. debian is pretty fast to upload though [20:27] smb: So, that's probably something I should fix in Debian but, also, worth noting that you don't need libc6-xen on Ubuntu at all, our default i386 flavour is also built with -mno-tls-direct-seg-refs === bfiller_afk is now known as bfiller [20:28] smb: (In other words, we shouldn't ship libc6-xen in Ubuntu at all, and I'm trying to decide if it's worth coming up with a migration plan to make that more obvious) [21:04] Meh, since I was updating the packaging for myself, it was trivial to generate the merge too: https://unit193.net/dropbear_2012.55-1.4ubuntu1_dropbear_2013.60-1ubuntu1.debdiff https://unit193.net/dropbear_2013.60-1_dropbear_2013.60-1ubuntu1.debdiff [21:59] ok. so in live-build, i have a package-lists/ubuntu-cloud that works for 12.04-13.10. [21:59] but in trusty, i want to *not* include the 'Task' of 'server^'. [21:59] anyone know how to put logic like that in there: [22:00] jtaylor, Are you able to sponsor xfwm4? [22:06] ok. [22:14] * seb128 wonders wth with dch in trusty [22:14] $ dch -i "Rebuild for the new poppler soname" [22:14] $ dch -r [22:14] -> gives "the" as a release [22:14] $ dch -i "Rebuild" [22:14] -> works fine [22:15] does anyone see anything wrong in there? (or feel like debugging dch? ;-) [22:19] https://ubuntuone.com/3vkRPzVaPpF8p6xferyRnT <- wtf. why i am on a vt with a mouse cursor, and unable to get back into my running xorg session, which btw, is still responding to presses of the multimedia keys?! [22:19] bahaha [22:20] happened during dist-ugprade, i think during configuration of the new kernel. my screeens just went dead but X was still running, and i could still control rhythmbox with my keyboard [22:20] seb128: check /usr/bin/dch line 1375 [22:20] Laney, shrug, is that a feature?! (and how did you debug that? ;-) [22:21] sk1llz [22:22] * seb128 is impressed by Laney p0wers [22:22] yeah I can trace perl code, that's why they pay me the big bucks [22:22] hehe [22:23] * seb128 notes to recommend that Laney gets more bucks in the next review [22:23] :> [22:23] I wonder if anyone even knows that feature exists any more [22:23] on the other hand, you now know how to work around it ... [22:24] desire to patch... slipping... === seb128_ is now known as seb128 [22:30] Riddell, could you get calligra rebuild for the new poppler soname change? (there is one small api change, http://cgit.freedesktop.org/poppler/poppler/commit/?h=poppler-0.24&id=ebe49d597a62aa94601c2e4595dbad1895ea7ef0 ... it's not likely it requires change but it would be good to test build still) === kyleN is now known as kyleN-afk [22:43] slangasek, it would be nice if you would fwd you upgrade fixes to Debian ;-) [22:44] seb128: if you mean mutter, the new upstream version was a -0ubuntu1 change... [22:45] slangasek, no, it just made me think of https://launchpad.net/ubuntu/+source/harfbuzz/0.9.24-2ubuntu2 which is our only diff on that source (crossed that this week because we were a version behind and it's our only diff and it's not forwarded) [22:47] seb128: ah, ok. yes, that should be forwarded; I'm just getting a little tired of having to garden these kinds of issues in libraries in Ubuntu, because they break update-manager in the devel series, which means either nobody's using update-manager or nobody minds being prompted about partial upgrades [22:48] so my motivation to forward these to Debian, who aren't affected by this update-manager code, is low [22:48] can we fix update-manager if the issue is limited to it? (or would it affect and "apt-get update" upgrade as well?) [22:51] slangasek, well, I would be fine fixing some of those/forwarding them if I understood the issue enough to explain the impact in a bug fwd to Debian (or is that just a "it's how it should be done")? === salem_ is now known as _salem [22:53] slangasek, (and yeah, I've to admit I'm one of those who don't pay attention to the partial upgrade dialog during unstable serie because we have enough transitions happening that's not un-usual, even if we didn't have bugs) [22:54] seb128: yeah, the issue is update-manager knows what Conflicts/Replaces means and will accept removing a package for that reason, but Breaks/Replaces (or Conflicts without Replaces) means something different according to policy so won't silently allow it [22:56] slangasek, I see [22:56] so yeah, long explanation for a patch that doesn't make any practical difference in Debian [22:57] having to maintain the diff/rebase is annoying as well [22:57] not sure that's worth the diff in fact, it only impact users who follow unstable and updated before that version [22:57] those changing series now are going to dist-upgrade anyway [22:57] so it could be dropped? [22:58] well, it doesn't hurt to keep that one until the LTS... [22:59] on that note, calling in a day [22:59] good night everyone === timrc is now known as timrc-afk [23:54] I usually commit those if I see them and can [23:54] as I did do for harfbuzz, IIRC