=== bluesabre1 is now known as bluesabre === balloons is now known as Guest87130 [09:47] cyphermox, when you come online. you might want to upload a signed-grub2 along with grub2 to trusty proposed. I am getting partial upgrade-only mode again [10:22] smb: I'll do that right now. [10:22] cyphermox: ^ [10:23] infinity, ah thanks [10:27] ^--- Self-reviewing that, it's just a grub version bump. [10:28] smb: Iz done. [10:29] infinity, ok cool. I will check a bit later for it (and what else is pending in proposed) [10:51] tumbleweed: You around? [10:53] tumbleweed: Who owns the reverse-depends(1) server-side bits? And, if it's you, could I ask you to add the debian-installer components to the Packages files you consider? [10:53] tumbleweed: Policy doesn't allow namespace overlap between debs and udebs, so it shouldn't cause any issues to just concat it all together. [10:56] tumbleweed: (Well, I suppose it could cause issues in your scripts, depending on how you process Provides, in some curious corner cases, but if that's a concern, you could have it not included in the default component set and require a --di switch or something) [11:08] infinity, nodejs block-proposed ping. is this still necessary? [11:09] doko: I think so. My "rebuild node-*" attempt didn't end well. [11:09] doko: I think we can maybe still get it in, but only if the rdeps get fixed up. I've not had time to look at the failures, only to see that lots of stuff fails. [11:10] doko: See node-* in https://launchpad.net/~adconrad/+archive/ubuntu/ppa/+packages if you're curious. [11:16] looks more like a task for the x-series ... [11:17] doko: Possibly, yes. Though, it's also a bit curious. See node-srs in wily-proposed. [11:17] doko: That version was specifically updated for node4 support, and built in Debian, but failed in Ubuntu. [11:19] doko: What's our toolchain delta these days? Just --as-needed, or also still some tighter security/sanity flags? Seems weird that these build logs should differ. [11:19] doko: g++-5 in both cases, but the debian CXX object builds cleanly, on Ubuntu, it's warning/error city. [11:19] yes, security when the options are not passed using dpkg-buildflags [11:21] * infinity wonders if maybe node's build system needs to relax some flags or some such. [11:21] But yes, possibly an x-series task, if no one has the round tuits to dig deeper. [11:24] chrisccoulson, infinity: looks like oxide-qt is missing some architecture specific bits, which are present in the other plethora of webkit versions. any reason for that? [11:24] I know nothing of oxide. [11:25] doko: Actually, I wonder if this node issue isn't node at all, but gcc5. I should retry my rebuild tests without proposed enabled, using the old nodejs. [11:26] doko: (not that I'm saying gcc5 is broken, just that new warnings/error might have blown up the node world) [11:27] Still curious that it's working in Debian, though. [11:27] doko: Do the gcc-5 binary packages include a handy "here's the difference from Debian" note, or do I just need to read the source package and sort it out? [11:30] infinity, no, best thing would be the diff of the debian dirs, and then maybe grepping for distrelease / derivative / distribution [11:30] doko: *nod* [11:32] doko, oxide is not webkit [11:33] chrisccoulson, but it ftbfs in webkit ;-P [11:33] It sort of is. [11:33] -- Running gyp [11:33] gyp: Undefined variable platform_heap_asm_files in /«PKGBUILDDIR»/third_party/chromium/src/third_party/WebKit/Source/platform/blink_platform.gyp [11:33] CMake Error at CMakeLists.txt:178 (message): [11:33] Running gyp failed [11:34] On those architectures, It's missing amongst plenty of other things, compilers for V8 [11:34] which is what nodejs 4.0 should provide ... [11:34] Which is now a solved problem in upstream V8, so oxide could use an update. [11:35] we use whatever chrome ships [11:35] Yay for everyone including static versions of everything. :/ [11:35] (But bundling libraries is NEVER a problem!) [11:36] chrisccoulson, firefox dropped my build fix [11:36] Trying to deliver frequent updates for something that has a dependency on a library with no stable ABI that's consumed by other applications in the archive would be significantly more of a problem [11:36] doko, what fix? [11:37] http://launchpadlibrarian.net/219814387/firefox_41.0%2Bbuild3-0ubuntu1_41.0%2Bbuild3-0ubuntu2.diff.gz [11:42] doko: Was the pkgconfig thing intentional in that diff, or just an oops? [11:42] doko: (But thanks for the -latomic fix) [11:43] I still consider it a gcc bug that -latomic isn't somehow autodected as needed on platforms where it's required. [11:43] oops [11:43] But I think I've seen more than one upstream bug where people disagree with me on that. [11:43] So I've never tried to argue it. [11:44] gcc-5 in wily now has a workaround, but as firefox is uploaded to released versions as well, it should carry that patch [11:44] doko: Wasn't the gcc-5 workaround specifically about -latomic for some C++ atomics, not for the entire atomic family? [11:45] doko: As in, powerpc (and a few arches we don't build for) is still a unique snowflake that needs -latomic for pretty much all atomics, even from C. [11:45] the former. but this was fixed upstream, but still showing on powerpc only [11:47] yes, because the builtins aren't resolved to some instructions, but are left "unimplemented". but some support is missing, like 8byte atomics. that's why I was looking into the insighttoolkit4 build [11:50] infinity, did you cancel the gdb build on armhf? [11:50] for the test rebuild [11:50] Nope. [11:51] I've only been babysitting the PPC builders, since they're "special", I haven't even looked at armhf. [11:52] I'm tempted to just update the PPC builders to utopic kernels, so gdb and gcc stop killing them, but that feels like a total cop-out when we should be able to debug and fix 3.13. But no one from either Canonical or IBM has found the time, so... :/ [11:56] pitti, one more debhelper merge? [12:29] doko: ENOVERB? I don't see any upload in unapproved? [12:30] pitti, no, was asking if you would do it? it's in unstable, not ready for upload =) [12:30] I think he's asking for you to put one there [12:30] as TIL [12:30] cheeky [12:30] doko: ah; https://tracker.debian.org/news/717055 looks rather intrusive, and has new depends in universe (dh-strip-nondeterminism) [12:31] that looks a little intrusive at this point; are you looking for a particular fix? [12:31] pitti, isn't 05 dropping that? [12:31] th dh_movefiles thingy [12:31] not according to https://tracker.debian.org/news/717170 [12:31] doko: ack, I'll cherry-pick that then [12:32] doko: https://anonscm.debian.org/cgit/debhelper/debhelper.git/commit/?id=c530fe28c07 ? [12:33] pitti, yep [12:34] doko: ack; test-building [12:39] doko: ^ [12:41] pitti: +1 for removing the errant .rej file too. :P [12:41] infinity: yeah, some dude forgot that on the last merge :) [12:41] careless people that === Odd_Blok1 is now known as Odd_Bloke === davmor2_ is now known as davmor2 [14:06] davmor2: I'm curious if it's the fact that it was missing a merge, or the fact that it was really really utterly broken, that makes my fixed usb-modeswitch work for you :) [14:08] cyphermox: sorry how do you mean? You asked for someone with a huawei to test it so I did [14:08] yes [14:08] the package was very very broken, not just for huawei [14:09] cyphermox: I guess people don't use dongles as much so no one noticed [14:09] cyphermox: the original package you mean not your fixed one right [14:09] yes [14:10] the original wouldn't trigger for a lot of dongles, I expect. I'm surprised, it doesn't seem like much change to make that not as popular; except maybe tablets. [14:10] cyphermox: I have a mifi (3g accesspoint) so don't need the dongle [14:31] infinity: yeah, I own them [14:32] infinity: https://code.launchpad.net/~stefanor/+junk/reverse-deps [14:42] cyphermox: in fact I have to switch the sim out of my mifi and put it in my dongle cause the sim in it is defunct [14:47] heh [14:47] I washed the best dongle I had once too many times, it no longer works [14:48] come to think of it, maybe it was the one good huawei dongle I had too [14:48] yup, oops. === Guest43458 is now known as med_ === med_ is now known as med === med is now known as med_