:: #mythtv

Daily chat history

Current users (74):

aloril, amessina, andreaz, Anssi, bhaak, blahdodo, brfransen, Captain_Murdoch, ChanServ, Chutt, clever, coling_, CyberJacob, dblain_, dekarl, eharris, ElmerFudd, enyc, gary_buhrmaster, ghoti, gigem, gregbert, gregL, GreyFoxx, Guest41682, Hydr0p0nX, ikevin, J-e-f-f-A, jab416171, jams, jmcentee, jnylen, joki, jpharvey, jst_, jya, kc, knightr, KungFuJesus, lautriv_, lautriv__, libsci, markspieth, MythLogBot, natanojl, nephyrin`, peper03, poptix, pppingme, raven42, rich0, Roklobotomy, saaki, seld, ShapeShifter499, sphery, SpootDev, sraue_, stefan__1, stuartm, suffice, superm1, tgm4883, Tobbe5178, tonsofpcs, tris, unforgiven512, vincent43, wagnerrp, Warped, XDS2010, xris, _charly_, _iwc
Saturday, February 6th, 2016, 00:00 UTC
[00:00:43] knightr: dekarl, that lock was not necessary, he would not have posted back and only wanted to provide valuable input...
[00:01:15] knightr: I have dealt with Antonio for quite a while now so I know this was definitely overkill for him...
[00:04:56] dekarl: knightr: I did not mean harm. Or anything personal. (In fact I did not realize its one of our contributors) Its just that I get really confused when tickets are brought back from the dead. (Its been closed for three years)
[00:05:07] dekarl: I'd prefer an auto-lock after being closed 6 months actually
[00:25:13] MythBuild: build #37 of master-ubuntu-lts-armhf is complete: Failure [4failed compile core] Build details are at . . . hf/builds/37 blamelist: Karl Dietz < >
[00:36:16] arescorpio (arescorpio! has joined #mythtv
[00:56:24] gregbert: that doesnt look good
[01:07:04] blahdodo (blahdodo!~blahdodo@ has quit (Quit: Bye bye)
[01:09:18] blahdodo (blahdodo!~blahdodo@ has joined #mythtv
[02:29:23] andreaz (andreaz! has quit (Read error: Connection reset by peer)
[03:18:44] gary_buhrmaster: dekarl: Re Qt5 patches. Well, as you probably noticed, that whole series was "near and dear" to my heart (or at least keyboard). I was actually embarrassed to find the danielk had done some of the same work
[03:19:02] gary_buhrmaster: dekarl: (and I completely missed it and duplicated some of it). My bad.
[03:32:06] jmcentee (jmcentee! has joined #mythtv
[03:34:15] jmcentee_ (jmcentee_! has quit (Ping timeout: 240 seconds)
[04:12:37] dmfrey (dmfrey! has joined #mythtv
[04:13:22] dekarl: gary_buhrmaster: we need to get faster with patches... having them sit on trac for years stinks
[04:14:10] dekarl: and by stinks i mean "adds to technical debt in an uncontrolled chaotic way"
[04:24:27] gregL (gregL! has quit (Remote host closed the connection)
[04:26:27] gary_buhrmaster: Patches should likely not languish. Either accept, request revision (or a test case), or reject quickly. But, of course, that means someone has to do that work too. There is no free lunch here.
[04:26:50] dekarl: gregbert: no worries, we run the raspberry pi builders with 2 or 3 parallel compilers. sometimes that is just to much for the little memory they have
[04:29:12] dekarl: gary_buhrmaster: better tooling may help, but not do all work :/ (e.g. running patches against the buildbots, allowing some kind of code review and test)
[04:29:17] clever: dekarl: ive been playing with qemu-user lately, when combined with binfmt_misc, it allows you to run arm binaries on an x86 machine
[04:29:36] clever: so you can just grab an entire arm root fs, chroot into it, and run the native arm compiler
[04:29:52] clever: it will have full use of all of the x86 cores and ram, but will be a bit slower due to the emulation
[04:30:34] clever: 8 cores at 4ghz can still beat 4 1ghz cores though, even with that handicap
[04:30:42] dekarl: clever, I like the idea, but I'm playing with the idea of unit tests that run GPU code (and had the Pi collecting dust)
[04:31:01] clever: yeah, the gpu stuff cant be emulated
[04:31:11] clever: but if you can split the job up, compile on x86, scp over, then unit test
[04:31:39] dekarl: As long as the build job runs mostly <1h thats not worth it for me :)
[04:31:54] gary_buhrmaster: re: test patches. gerrit (which is likely a bigger project than you want) would run patches against the buildbots, but the issue is the time to review whether they make sense, not the building (at least I think that is true).
[04:31:55] clever: in my case, i have to wait 2 days 5 hours for qt4 to build
[04:32:00] clever: and another 2 days for llvm
[04:32:27] dekarl: well, if one jobs runs a week, then it'd be worth it, for me too :D
[04:32:31] clever: qt4 is at 1d 22h, so there is another ~7 hours left on it
[04:32:56] gary_buhrmaster: Re: qemu arm. It depends on your targets and hosts, but I have found that emulating arm takes about as long as running on "bare metal".
[04:33:12] clever: gary_buhrmaster: the key factor is qemu USER arm
[04:33:17] clever: it doesnt emulate the kernel
[04:33:35] clever: its more like wine, it sits between the arm binary and the x86 kernel, and translates the syscalls over
[04:33:46] clever: so you can use all the cores (normal qemu arm is single core)
[04:34:08] clever: the extra cores lets you beat the handicap caused by emulation
[04:35:51] clever:
[04:37:19] clever: ive only run into 2 blocking issues with qemu-user-arm, lld only works if you use a static qemu, and xargs fails with large command lines
[04:37:39] clever: and if your using a chroot, it has to be static, or its dyn libs must exist within the chroot
[04:40:38] ** clever heads off to bed **
[04:40:52] dekarl: sounds like a job for solaris zones or freebsd jails combined with foreign (e.g. fedora linux) userland
[04:41:09] dekarl: no idea what its called in linux land
[04:48:23] gary_buhrmaster: I wonder if the eventual release of the HuskeyBoard will obsolete many of the current challenges.
[04:52:10] dmfrey (dmfrey! has quit (Ping timeout: 250 seconds)
[05:03:45] dekarl: something like the big sister of DragonBoard 410c?
[05:04:26] dekarl: haha "it is not going to initially be $200. But it won't be $1000 either."
[05:07:30] joki (joki! has quit (Ping timeout: 256 seconds)
[05:13:53] joki (joki! has joined #mythtv
[05:15:59] gary_buhrmaster: The target was supposed to be $200–400, but AFAIK actual pricing has not been announced (just ranges being suggested).
[05:16:20] gary_buhrmaster: And while $400 is not pocket change (at least to me), I can see it possibly being made available as builders for some projects.
[05:38:41] dekarl: I wonder if there'll be an introductery 50% offer on that (and in EUR)
[05:44:24] dekarl: lets see if it comes with the 16GiB recommended minimum or 1Gib mandatory minimum memory (or a SO-DIMM socket)
[06:01:25] tgm4883: Thoughts on
[06:01:36] tgm4883: need to know if we need to add this downstream before the 18t
[06:01:40] tgm4883: 18th*
[06:15:08] dekarl: jya ^- thats the ffmpeg backports to fixes/0.27 to make it build with the upcoming Ubuntu LTS release
[06:44:42] jya: dekarl: what ffmpeg backport?
[06:49:31] MythBuild (MythBuild! has quit (Remote host closed the connection)
[07:47:29] blah_ (blah_! has joined #mythtv
[07:50:26] Roklobotomy (Roklobotomy! has quit (Ping timeout: 240 seconds)
[08:30:51] dekarl: jya, Ubuntu packagers noticed that they need three commits from ffmpeg, that we have in master, to compile fixes/0.27 on the upcoming release. #12634 I thought it makes sense to pass it by you, as you've been taking care of ffmpeg lately.
[08:30:51] ** MythLogBot **
[08:32:01] jya: I see.. So we're not doing 0.28 for 16.04 anymore?
[08:32:08] jya: I think we should now
[08:32:43] dekarl: but now that I looked at them again, they are not in ffmpeg/libav yet and partly untested...
[08:33:04] arescorpio (arescorpio! has quit (Quit: Leaving.)
[08:33:12] dekarl: but they are in the ffmpeg ubuntu packages
[08:35:49] jya: I compiled on 15.10; works fine
[08:36:09] jya: They will likely be back ported to the various branches.
[08:36:28] jya: Ffmpeg dev told me they were going to release 2.9 in time for 16.04
[08:36:52] jya: After that they will more on YYYYMMDD versioning.
[08:43:20] SteveGoodey (SteveGoodey! has joined #mythtv
[10:03:51] dekarl1 (dekarl1!~dekarl@mythtv/developer/dekarl) has joined #mythtv
[10:07:26] dekarl (dekarl!~dekarl@mythtv/developer/dekarl) has quit (Ping timeout: 276 seconds)
[10:10:17] dekarl1 is now known as dekarl
[10:36:52] blah_ is now known as Rokmonsta
[10:39:49] dekarl: still some stuff to go through!cl . . . port+-+Crash <- 11 crash reports... I get a 12th while trying to reproduce one ;)
[10:47:04] natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has joined #mythtv
[10:48:19] Rokmonsta (Rokmonsta! has quit (Remote host closed the connection)
[12:05:30] lautriv_ (lautriv_!~lautriv@funtoo/user/lautriv) has quit (Ping timeout: 250 seconds)
[12:17:58] lautriv_ (lautriv_!~lautriv@funtoo/user/lautriv) has joined #mythtv
[12:28:46] dmfrey (dmfrey! has joined #mythtv
[12:41:13] markspieth: clever: gary_buhrmaster: Ive done a qemu user build env for arm on x86 before. I replaced make and other tools with x86 versions, gcc/g++ as cross compilers x86 host/arm target. Build env then works as fast as normal as they are all host/x86 targets whilst unit tests run as arm
[12:42:02] markspieth: the qemu user shim looks at the elf target and executes the emu target as required.
[12:42:36] markspieth: this was all inside the arm chroot.
[12:46:08] dekarl: !seen MythBuild
[12:46:08] MythLogBot: MythBuild was last seen 5 hours 56 minutes 37 seconds ago
[12:46:19] dekarl: stuarta -^ fyi
[13:50:39] clever: markspieth: i have thought of that, but thats more work to mix arches in the same root
[13:50:57] clever: and my current OS is heavily against mixing stuff like that
[13:56:33] clever: it hashes all of the build inputs, and enforces them being the same arch and the latest version
[13:57:06] jpharvey (jpharvey!~jpharvey@ has quit (Quit: Ex-Chat)
[14:10:16] jpharvey (jpharvey!~jpharvey@ has joined #mythtv
[15:01:22] dmfrey (dmfrey! has quit (Ping timeout: 256 seconds)
[15:57:33] gregL (gregL! has joined #mythtv
[16:52:21] Steve-Goodey (Steve-Goodey! has joined #mythtv
[16:58:28] andreaz (andreaz! has joined #mythtv
[18:08:17] Steve-Goodey (Steve-Goodey! has quit (Quit: Konversation terminated!)
[18:55:08] totesmuhgoats (totesmuhgoats! has joined #mythtv
[18:55:19] totesmuhgoats: hi guys
[18:55:30] totesmuhgoats (totesmuhgoats! has left #mythtv ()
[19:03:32] jheizer_ (jheizer_!~jheizer@2601:246:8200:887f:50d3:fcb1:f76f:1a67) has quit (Ping timeout: 252 seconds)
[19:54:37] Roklobotomy (Roklobotomy! has joined #mythtv
[21:21:24] dekarl: knightr: this string is missing newlines or spaces between the text snippets. . . . ngs.cpp#L115
[21:23:26] dekarl: That would make a nice check for cppcheck / clang / some-tool. I'm not sure how to correctly fix it (I never saw that screen, its in DVB-S setup) and if its a bit late in this cycle for the fix.
[22:16:38] SteveGoodey (SteveGoodey! has quit (Quit: Konversation terminated!)
[22:24:33] Tobbe5178 (Tobbe5178!~asdf@2001:2002:4e46:4de8:1464:20de:34e9:1a83) has quit (Read error: Connection reset by peer)
[23:39:16] knightr (knightr! has quit (Ping timeout: 248 seconds)
[23:44:24] Tobbe5178 (Tobbe5178!~asdf@2001:2002:4e46:4de8:68b8:c43:e0c1:49c9) has joined #mythtv
[23:50:12] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv

IRC Logs collected by BeirdoBot.
Please use the above link to report any bugs.