:: #mythtv

Daily chat history

Current users (61):

aloril, Anssi, blahdodo, brfransen, bumblebeebat12, ChanServ, clever, davic, dym, eharris, ElmerFudd, enyc, fjear, frobnic, gary_buhrmaster, ghoti, gigem, gregl, GreyFoxx, hampton, Hydr0p0nx, ijc, jheizer, jpabq, jpharvey_, jya, kalamaja, knowledgejunkie, kwizart, libsci, mad_enz, Maliuta_, markspieth, markspieth2, MitchCapper, mkbloke, MythBuild, MythLogBot, MythNotifyBot_, nephyrin, ooshlablu, Panic, peper03, peterbennett, poptix, pppingme, ramshadow, rhpot1991, RokLobsta, sdrik, sphery, stuarta, stuartm, taylorr, Tobbe5178, tonsofpcs, tris, Warped, xris, zbot`, _charly_
Tuesday, January 8th, 2019, 00:10 UTC
[00:10:45] MythBuild (MythBuild! has quit (Remote host closed the connection)
[00:10:56] MythBuild (MythBuild! has joined #mythtv
[00:12:03] MythBuild (MythBuild! has quit (Remote host closed the connection)
[00:13:13] MythNotifyBot_: Service: HTTP Buildbot is Down – HTTP 502 – Bad Gateway
[00:13:17] MythBuild (MythBuild! has joined #mythtv
[00:13:48] MythBuild (MythBuild! has quit (Remote host closed the connection)
[00:14:58] MythBuild (MythBuild! has joined #mythtv
[00:15:06] MythBuild (MythBuild! has quit (Remote host closed the connection)
[00:15:25] MythBuild (MythBuild! has joined #mythtv
[00:16:12] MythNotifyBot_: Service: HTTP Buildbot is Up – HTTP 200 – OK
[07:09:24] peper03: mark____: I think in this case, the precision of the float is always going to be sufficient. I'd be interested to know whether the floating point code/hardware is optimized for 2^N values. I understand the principle, of course. I just have no idea whether the implementation matches that.
[07:09:32] peper03: I assume the '(2*
[07:10:13] peper03: Sorry. The '(2 * col_width)' is only determined at run-time?
[07:18:47] mark____: peper03: compile time for glsl  – which is c++ run time on video setup:) so it is pre-calculated. though to be honest, we don't stress even the oldest of gpus when it comes to calculations. texture sampling is by far the biggest slow down
[07:21:51] peper03: mark____: OK. Good to know!
[07:25:07] Tobbe5178 (Tobbe5178!~asdf@2001:2002:51eb:d24e:8c18:11e7:cd66:ec8b) has joined #mythtv
[08:06:09] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv
[08:06:10] Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey
[08:10:09] Tobbe5178 (Tobbe5178!~asdf@2001:2002:51eb:d24e:8c18:11e7:cd66:ec8b) has quit (Read error: Connection reset by peer)
[09:02:31] willcooke (willcooke!~willcooke@ubuntu/member/willcooke) has joined #mythtv
[09:08:41] Merlin83b (Merlin83b! has joined #mythtv
[09:50:48] stuarta: mark____: so do we think after your commits this morning we should be in a place to cut the branch, or should we wait for some cross testing from peterbennett and other devs?
[10:01:51] markspieth2: mark____: unfortunately yes. unless you cash force int16 as integer which from reading I don't think is possible.
[10:02:51] MythBuild: Build master-ubuntu-18_04–64bit #81 is complete: Failure [failed test (failure)] –
[10:02:51] ** MythLogBot **
[10:03:35] markspieth2: I also changed all the constant renders to 16 digits to minimize rounding errors. I think that was worthwhile.
[10:04:23] markk_ (markk_!~androirc@ has joined #mythtv
[10:06:19] stuarta: hrm, MythLogBot is going to have to ignore MythBuild
[10:08:14] markspieth2: the other thing I noticed that the deinterlacing algos assumed that top left was origin whereas bottom left is. fixing that made a noticeable difference in quality esp y12 chroma sampling which is not deinterlacing.
[10:08:39] markk_: stuarta: I'm pretty comfortable with those commits. The last has had a few hours of testing with as many configurations as my sanity could handle.
[10:09:15] markk_: Am stuck in jury service this week and next – so dev time now limited:(
[10:10:03] markspieth2: for precision issues perhaps provide the same image twice to the gl alo offset by half a line
[10:10:06] stuarta: markk_: okay, thanks for the feedback.
[10:11:25] markk_: markspieth2: yv12 chroma sampling is still borked. I think it is trying to handle chroma upsampling issue but failing
[10:12:12] MythBuild: Build master-ubuntu-18_04–64bit #82 is complete: Success [build successful] –
[10:12:12] ** MythLogBot **
[10:12:12] markk_: I'm going to try and improve yv12 by using 3 textures. Should improve performance
[10:12:46] ** stuarta reduces the number of cpus on the ubuntu-18.04 builder and restarts the build **
[10:12:59] stuarta: well it completed whilst i typed that
[10:27:57] markk_ (markk_!~androirc@ has quit (Ping timeout: 246 seconds)
[11:17:03] markk_ (markk_!~androirc@ has joined #mythtv
[11:20:19] markk_: markspieth2: slightly confused – do you have local changes to the GL code? Some of those changes you talk about dont appear to be in master...
[11:27:49] markk_ (markk_!~androirc@ has quit (Read error: Connection reset by peer)
[11:28:49] markk_ (markk_!~androirc@ has joined #mythtv
[11:32:50] markk_ (markk_!~androirc@ has quit (Read error: Connection reset by peer)
[11:33:16] markk_ (markk_!~androirc@ has joined #mythtv
[11:34:40] blahdodo (blahdodo!~blahdodo@ has quit (Quit: Bye bye)
[11:35:50] blahdodo (blahdodo!~blahdodo@ has joined #mythtv
[11:48:03] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)
[12:36:21] markk__ (markk__! has joined #mythtv
[12:39:34] markk_ (markk_!~androirc@ has quit (Ping timeout: 244 seconds)
[12:41:47] markk__ (markk__! has quit (Ping timeout: 240 seconds)
[12:41:54] markk__ (markk__! has joined #mythtv
[13:08:09] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv
[13:08:10] Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey
[13:14:16] markk__ (markk__! has quit (Remote host closed the connection)
[13:22:11] markspieth: markk__: I have a patch set which works better for android and have been trialling them locally but am not sure whether they are good. I send them to the mailling list a while back. I will repost for your perusal.
[13:30:43] markspieth: see #13230
[13:30:43] ** MythLogBot **
[13:56:37] blahdodo (blahdodo!~blahdodo@ has quit (Quit: Bye bye)
[13:56:58] blahdodo (blahdodo!~blahdodo@ has joined #mythtv
[15:41:32] kwizart (kwizart!~kwizart@fedora/kwizart) has joined #mythtv
[15:41:32] Mode for #mythtv by ChanServ!ChanServ@services. : +v kwizart
[17:16:53] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has quit (Remote host closed the connection)
[17:17:29] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has joined #mythtv
[17:17:29] Mode for #mythtv by ChanServ!ChanServ@services. : +v stuartm
[17:28:19] Tobbe5178 (Tobbe5178!~asdf@2001:2002:51eb:d24e:4c45:87a5:8a03:6021) has joined #mythtv
[17:40:05] rubbergsx (rubbergsx! has joined #mythtv
[17:49:16] rubbergsx (rubbergsx! has quit (Remote host closed the connection)
[17:59:34] Merlin83b (Merlin83b! has quit (Quit: Leaving)
[18:20:21] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)
[18:20:35] Steve-Goodey (Steve-Goodey! has joined #mythtv
[18:20:36] Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey
[18:48:34] willcooke (willcooke!~willcooke@ubuntu/member/willcooke) has quit (Quit: Do your hobbies)
[21:24:45] Steve-Goodey (Steve-Goodey! has quit (Quit: Konversation terminated!)
[21:26:12] mark____: markspieth: thanks for that. do you really think using 16 decimal places is helping? I don't see anything on deinterlacingtop/bottom – only removing the silly yv12 stuff
[21:26:31] mark____: stuarta: how's the release going? :)
[21:27:57] stuarta: assuming nothings broken i could probably cut it now
[21:28:05] kwizart: hi there, any info why vaapi backend are blacklisted in mythtv ?
[21:28:34] mark____: vaapi backend??
[21:28:55] stuarta: kwizart: no vaapi blacklisting to my knowledge
[21:28:57] kwizart: which made me use xv and discover that my adapter is not supported
[21:29:04] kwizart:
[21:31:30] kwizart: stuarta, it's more some entrypoint that are not supported , looking at libmythtv/vaapicontext.cpp
[21:31:33] kwizart: IZZ (UNSUPPORTED)
[21:31:35] kwizart: etc
[21:31:46] stuarta: that's normally the encoding profiles
[21:31:57] stuarta: kwizart: please pastebin `vainfo` output
[21:32:00] kwizart: I exepct that's the reason why I cannot set vaapi, but maybe there is another reason
[21:32:36] stuarta: if you can't select it, it's normally because it was built without it
[21:32:52] stuarta: which if you (as i suspect) are using fedora packages i'd be surprised
[21:32:52] kwizart: stuarta,
[21:33:12] mark____: VAEntrypointVLD is the only entry point that actually works – god knows why they added the others.
[21:33:27] kwizart: stuarta, I can select it, but then it fails to setup a display
[21:33:34] stuarta: kwizart: logs please
[21:33:50] stuarta: VAEntrypointVLD <- are the decoding entry points
[21:34:02] stuarta: VAEntrypointEncSlice <- are the *encoding* entry points
[21:34:30] stuarta: VAEntrypointVideoProc <- i believe is the chain used for de-interlacing
[21:35:09] stuarta: kwizart: ah, also assuming fedora, are you using wayland?
[21:35:30] kwizart: stuarta, no
[21:35:47] stuarta: no? okay, then we need to see some logs
[21:35:47] ** kwizart tries to reproduces the logs **
[21:35:58] stuarta: --verbose playback
[21:37:55] kwizart: I should probably run it with LANG=C
[21:38:03] stuarta: heh
[21:38:26] markspieth: mark____: my theory was rounding where the iterator * max line was quantised so that teh last horizontal pixel was green (0,0,0). It seems to work but depends on width value
[21:39:32] kwizart: stuarta,
[21:39:55] markspieth: Have just attached my full openglvideo.cpp which I got back from reflog/cat-file for full reference as the last diff has some of the line order changed and have not fixed them yet.
[21:40:54] markspieth: for the yv12 I reenabled it and ested with a low res video (see oter dropbox items provided by david)
[21:41:16] stuarta: kwizart: oh, you have a vdpau capable card. it's doing a vaapi -> vdpau translation
[21:41:21] stuarta: why not use vdpau?
[21:41:30] markspieth: the RECT_TEST macro was useful for me to understand what was going on.
[21:41:45] kwizart: stuarta, vdpau doesn't work (green display)
[21:42:08] stuarta: ok, so that's a separate problem
[21:42:14] kwizart: okay so I've disabled libvdpau-va-gl
[21:42:14] markspieth: off to work so if you need to discuss use the mailing list.
[21:42:21] stuarta: kwizart: and it deserves a ticket
[21:42:27] mark____: kwizart: you need to change to the OpenGL ui renderer
[21:42:52] mark____: markspieth: thanks
[21:45:06] stuarta: kwizart: or at least set the ui painter to auto
[21:45:10] ** mark____ thinks this is exactly the sort of problem that should be going away shortly... **
[21:45:30] stuarta: mark____: looking forward to the hacking that can be done post cutting the branch
[21:45:44] stuarta: i have a feeling it may be better off in a branch initially
[21:47:32] mark____: stuarta: keep on meaning to outline some ideas/plans in response to your v31 email. Hopefully I'll have another 5 or 6 hours of tedium hanging around the courts tomorrow :(
[21:48:14] stuarta: i dread ever having to do that
[21:50:07] peterbennett: The current master is looking good as far as opengl is concerned
[21:50:42] peterbennett: mark____: I am not sure what the difference between opengl and opngl-yv12 is
[21:51:10] stuarta: so the consensus is we are good to cut the branch?
[21:51:20] peterbennett: mark____: It looks like the new opengl hquyvy is the slow version of uyvy.
[21:51:56] peterbennett: stuarta: I am happy with it now.
[21:52:20] peterbennett: users of Android may have to change some settings when upgrading because of the OpenGL changes
[21:52:37] kwizart: okay, so it seems better to set ui to auto (instead of qt – I don't remember why I've made this change if not the default)
[21:53:22] kwizart: then vaapi works once libvdpau-va-gl is removed (but I'd like to keep it installed but enforce vaapi still for the long run)
[21:54:56] kwizart: then there is the xv adapter name with modesetting, do you want a bug report for both ?
[21:55:07] kwizart: trac is under maintainance ?
[21:55:16] stuarta: shouldn't be
[21:55:41] stuarta: nope
[21:58:28] kwizart: ?
[21:58:32] stuarta: no
[21:58:38] stuarta:
[21:59:00] stuarta: that was my test url iirc
[22:10:21] jpabq: Re: Android, should qtwebkit work? I just tried building the dependancy libs from scratch for the first time, in a long time, and it bombs when it gets to qtwebkit.
[22:14:04] stuarta: branch cut, builders running a fixes/30 build
[22:16:17] jheizer: oh, new build master
[22:16:23] stuarta: yeah :)
[22:16:24] jheizer: 2 reasons for high fives
[22:16:25] stuarta: i like it
[22:16:46] jheizer: my auto complete url went to a dead page
[22:16:51] jheizer: and was like hmmm
[22:17:03] stuarta: ah yeah they would
[22:17:34] jheizer: looks nice
[22:17:44] jheizer: I'll zip my ubuntu one back to 18.04 now
[22:19:51] peterbennett: jpabq: I have not had any problem, but I have not rebuilt libs for a while
[22:21:02] stuarta: now where the hell do i change the version string in master to v31pre
[22:23:36] jpabq: peterbennett: gigem markspieth: it complains "make: *** qtwebkit/Source: No such file or directory. Stop.". If I do a find, it is in "android/libs/qt-everywhere-src-5.10.1/qtwebkit" and "android/libs/qt-everywhere-src-5.10.1/build/qtwebkit"
[22:23:59] stuarta: think qtwebkit is deprecated
[22:24:34] jpabq: Very. I think it is still used by at least one mythplugin, though. Not that I use any mythplugins.
[22:27:35] stuarta: mythbrowser iirc
[22:27:52] stuarta: which needs to either be pretty much rewritten or removed
[22:27:53] jheizer: says a builder with the same name is already connected and dropping me
[22:28:10] jheizer: ip addr
[22:28:11] jheizer: bah
[22:28:12] stuarta: jheizer: which slave name?
[22:28:31] jheizer: jheizer-ubuntu-lts-64bit
[22:28:37] stuarta: i may have nicked one for testing
[22:29:18] jheizer: ah got ya
[22:29:24] stuarta: lemme check
[22:29:29] peterbennett: jpabq: IIRC there is some special code in libr build ft qtwebkit
[22:29:45] jheizer: Log –
[22:30:07] jheizer: all I did was change the startup script to use buildbot vs buildslave
[22:30:14] jheizer: not sure if anything else was required
[22:31:05] stuarta: erm... dns failure?
[22:31:20] jheizer: not sure, it connects after that
[22:31:29] ** stuarta looks some more **
[22:31:36] gary_buhrmaster (gary_buhrmaster!~gtb@2601:647:4802:4c3:222:4dff:fe51:6728) has joined #mythtv
[22:32:10] peterbennett: jpabq: I beleiev this solves the webkit issue – sudo apt install gperf bison flex gradle
[22:32:47] gary_buhrmaster: stuarta: mythweb fixes/30 branch?
[22:32:55] stuarta: gary_buhrmaster: good point
[22:33:39] jheizer: don't worry about me
[22:33:44] jheizer: poor timing on my part
[22:34:01] stuarta: gary_buhrmaster: done
[22:34:17] SteveGoodey (SteveGoodey!~steve@2a00:23c5:7da3:4501:7a24:afff:fe9d:c233) has quit (Quit: Konversation terminated!)
[22:34:19] jheizer: oh
[22:34:22] jheizer: and those logs are old
[22:34:30] jheizer: just noticed the date/time. Derp
[22:34:36] rubbergsx (rubbergsx! has joined #mythtv
[22:34:37] gary_buhrmaster: stuarta: thanks (I am just being my usual PITA).
[22:34:45] rubbergsx (rubbergsx! has quit (Client Quit)
[22:34:54] jheizer: I'll just swap back
[22:35:00] stuarta: gary_buhrmaster: nah, someone has to check i've not forgotten anything
[22:36:14] stuarta: jheizer: i don't appear to have "borrowed" your slave name
[22:36:16] gary_buhrmaster: stuarta: at one point I thought I saw a script floating around that did the branch on both mythtv and mythweb at the same time, so I am a bit surprised it got missed (or the script has disappeared into some forgotten hole).
[22:36:33] jheizer: running it manually I was having other issues
[22:36:39] jheizer: will check it out another day
[22:36:50] stuarta: gary_buhrmaster: there a release script that does that for the actual release
[22:38:56] gary_buhrmaster: stuarta: sounds like a branch script could be useful for the future forgetful (or the next release wrangler?).
[22:41:38] stuarta: heh
[22:44:55] stuarta: jheizer: oh fyi, with your rpi, the whole slave name should change
[22:45:13] stuarta: s/jessie/stretch as well as s/71/7l/
[22:46:45] MythBuild (MythBuild! has quit (Quit: buildmaster reconfigured: bot disconnecting)
[22:46:56] MythBuild (MythBuild! has joined #mythtv
[22:48:04] rubbergsx (rubbergsx! has joined #mythtv
[22:48:09] jheizer: stuarta: yes true
[22:48:30] rubbergsx (rubbergsx! has quit (Client Quit)
[22:50:16] jheizer: hmm if it is even online
[23:06:04] jheizer: bah yeah, nfsroot is mounting. todo++
[23:22:28] mark____ (mark____! has quit (Quit: mark____)
[23:23:43] jpabq: peterbennett: you mean `sudo dnf install gperf bison flex gradle`  :-)
[23:24:07] jpabq: I tried that, but I already had thst stuff installed.
[23:24:42] jpabq: I am building again, but with qtwebkit commented out of
[23:29:22] jpabq: Which obviously I didn't do correct, since it still tried to build it. Sigh. This would not be so bad if it didn't take so long.
[23:40:05] peterbennett: jpabq: I had the same error before when I first tried library build
[23:41:34] peterbennett: jpabq: I think the solution is to look through the build log file. That is how I found the missing dependencies. There may be an error message, but it doers not stop at the error, it only stops later when something is missing
[23:42:28] jpabq: The error I am seeing seems to indicate a path issue. It claims it cannot find the qtwebkit src directory, even though it is there.
[23:42:46] peterbennett: jpabq: The message "make: *** qtwebkit/Source: No such file or directory. Stop." was misleading
[23:43:26] jpabq: Oh? Okay. If my latest attempts at skipping qtwebkit still don't work, I will try to find the build log file.
[23:45:54] peterbennett: jpabq: Onle time the message from the log file was "The WebKit build was disabled for the following reasons: * Missing ruby from PATH "
[23:46:09] peterbennett: That is found by reading the log file carefully
[23:47:15] jpabq: Ah. I actually do not have ruby installed. I have re-installed the OS since the last time I built this stuff, and missed that.
[23:48:50] peterbennett: To get the build log I think you need |tee mythbuild.log
[23:59:25] peterbennett: stuarta: We also need a fixes/30 branch of packaging repo

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