:: #mythtv

Daily chat history

Current users (64):

aloril, amessina, Anssi, benklop, blahdodo, brfransen, ChanServ, Chutt, clever, CyberJacob, davic, dblain, dekarl, dym, eharris, ElmerFudd, enyc, frobnic, ghoti, gigem, gregbert, gregl, GreyFoxx, hampton, Hydr0p0nX, ijc, ikevin, J-e-f-f-A, jarle_, jedix, jheizer, jpabq, jpharvey, jpharvey_, jya, kc, knowledgejunkie, libsci, mad_enz, markspieth, MitchCapper, MythBuild, MythLogBot, MythNotifyBot, nephyrin, ooshlablu, peper03, pppingme, ramshadow, rhpot1991, sdrik, sheedy, sphery_, sraue, stuartm, taylorr, Tobbe5178, toeb, tonsofpcs, tris, Warped, XDS2010|AFK, xris, _charly_
Friday, December 1st, 2017, 00:39 UTC
[00:39:11] hampton (hampton! has quit (Ping timeout: 240 seconds)
[00:48:20] dekarl (dekarl!~dekarl@mythtv/developer/dekarl) has joined #mythtv
[00:50:05] dekarl1 (dekarl1!~dekarl@mythtv/developer/dekarl) has quit (Ping timeout: 276 seconds)
[02:11:45] hampton (hampton! has joined #mythtv
[02:17:52] sphery_ (sphery_!~mdean@mythtv/developer/sphery) has joined #mythtv
[02:22:42] sphery (sphery!~mdean@mythtv/developer/sphery) has quit (*.net *.split)
[02:33:11] Guest10049 (Guest10049! has joined #mythtv
[02:35:11] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 248 seconds)
[03:33:11] peper03 (peper03!~peper03@mythtv/developer/peper03) has quit (Ping timeout: 240 seconds)
[03:38:13] peper03 (peper03!~peper03@mythtv/developer/peper03) has joined #mythtv
[06:58:06] Roklobster (Roklobster! has joined #mythtv
[07:03:06] Warped (Warped!~Warped@unaffiliated/warped) has quit (Quit: ChatZilla 0.9.93 [Firefox 56.0.2/20171024165158])
[07:04:27] Guest10049 (Guest10049! has quit (Ping timeout: 240 seconds)
[07:13:03] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[07:14:06] SteveGoodey (SteveGoodey! has joined #mythtv
[08:31:36] Warped (Warped!~Warped@unaffiliated/warped) has joined #mythtv
[08:44:13] willcooke (willcooke! has joined #mythtv
[08:44:13] willcooke (willcooke! has quit (Changing host)
[08:44:13] willcooke (willcooke!~willcooke@ubuntu/member/willcooke) has joined #mythtv
[09:15:26] nephyrin (nephyrin!~neph@ has quit (Ping timeout: 255 seconds)
[09:23:53] Roklobster (Roklobster! has quit (Read error: Connection reset by peer)
[09:24:52] nephyrin (nephyrin!~neph@ has joined #mythtv
[09:35:05] SteveGoodey (SteveGoodey! has quit (Quit: Konversation terminated!)
[10:29:40] Roklobster (Roklobster! has joined #mythtv
[10:32:57] Merlin83b (Merlin83b! has joined #mythtv
[11:13:28] ** stuarta yawns **
[11:44:06] SteveGoodey (SteveGoodey! has joined #mythtv
[12:05:10] Roklobster (Roklobster! has quit (Read error: Connection timed out)
[12:05:59] Roklobster (Roklobster! has joined #mythtv
[12:14:09] SteveGoodey (SteveGoodey! has quit (Quit: Konversation terminated!)
[13:30:28] Roklobster (Roklobster! has quit (Read error: Connection timed out)
[13:31:08] Roklobster (Roklobster! has joined #mythtv
[13:55:38] pvr4me (pvr4me! has joined #mythtv
[13:57:08] pvr4me: #13185
[13:57:08] ** MythLogBot **
[13:57:57] Roklobster (Roklobster! has quit (Read error: Connection timed out)
[13:58:42] Roklobster (Roklobster! has joined #mythtv
[14:08:11] stuarta: pvr4me: that's going to be a challenging thing to fix
[14:08:30] SteveGoodey (SteveGoodey! has joined #mythtv
[14:08:31] pvr4me: Yeah, I’m afraid of that
[14:10:11] pvr4me: I just updated the ticket. It is possible that the problems stem from MacPorts packaging of Qt5. I’ve asked for input from the guy responsible
[14:11:28] stuarta: i could also be the way we use the toolkit
[14:12:11] stuarta: main problem, nobody apart from jya has a mac on which to do any testing / development
[14:12:41] jya: everyone hsould use a mac
[14:12:44] jya: for dev
[14:12:49] pvr4me: It is also strange that these issues are only present in macOS 10.13. This release was supposed to be primarily bug fixes; not breaking changes.
[14:12:50] jya: or linux if you use rr
[14:13:00] jya: rr, best invention this bread
[14:13:03] stuarta: rr ?
[14:13:16] jya:
[14:13:38] jya: stuarta: use your code within a rr session, almost no impact on performance.
[14:13:54] stuarta: looks interesting, can see how that would be useful
[14:14:06] jya: if it ever crash, you will *always* be able to reproduce it, debug, rewind, reproduce the race in a deterministic fashion
[14:14:24] stuarta: very hand
[14:14:25] stuarta: very handy
[14:14:29] jya: you just have to use the program until it crashes
[14:14:44] jya: if I had that at the time I was debugging mythbackend that would have been awesome
[14:15:14] jya: microsoft have produced something similar, but the impact on initial running speed is enormous
[14:15:44] jya: I can run firefox in rr, multiprocess and all, and barely notice the impact on speed
[14:15:51] stuarta: pvr4me: so issue 2, could be potentially as simple as the way fullscreen is signalled via qt
[14:17:59] pvr4me: stuarta: I think so. I think #4 may also be related—the Dock should automatically reappear when the app quits or even goes into the background.
[14:17:59] ** MythLogBot **
[14:18:32] stuarta: might be a side effect of not going properly into fullscreen
[14:21:54] stuarta: pvr4me: do you see the log message from the frontend "Using Full Screen Window"
[14:22:36] pvr4me: give me a minute to launch the VM
[14:24:16] stuarta: . . . ow.cpp#L1060
[14:24:30] stuarta: this bit of code is where it tells Qt to use fullscreen
[14:25:04] stuarta: in fact i have some simple test code to demonstrate this
[14:25:07] stuarta:
[14:25:25] stuarta: specifically the stuff here . . . get/main.cpp
[14:25:48] stuarta: if you comment out line 16 and uncomment line 18, you should get a fullscreen triangle rendered via opengl
[14:28:22] pvr4me: Log from last night:
[14:30:48] pvr4me: stuarta: I’m not a developer/don’t have a build environment to test that fragment
[14:37:03] stuarta: it's a very simple qt app, qmake in the directory, make, run
[14:38:52] stuarta: basically if you can build mythtv, you can build that
[14:39:25] stuarta: erm, that's a very odd screen resolution
[14:39:32] stuarta: UI Screen Resolution: 1039 x 726
[14:39:56] pvr4me: trying to build now
[14:40:32] pvr4me: Donno how the VM gets its screen resolution.
[14:40:45] stuarta: makes it up clearly
[14:41:03] stuarta: you should be able to pull that with xdpyinfo
[14:44:10] Roklobster (Roklobster! has quit (Read error: Connection timed out)
[14:44:47] Roklobster (Roklobster! has joined #mythtv
[15:00:32] pvr4me: stuarta: I got it to build but I can’t seem to get it to execute:
[15:00:34] pvr4me:
[15:01:12] stuarta: mine builds an `openglwidget` binary, so what you want is probably the
[15:01:26] stuarta: if i remember my osx correctly
[15:07:38] pvr4me: … ;)
[15:08:28] stuarta: sadly my first gen mac mini, isn't capable of running any of the 64bit OSX releases
[15:11:17] pvr4me: You mean the Core Duo mini? There was a PPC (G4) version before that. Still 10+ years and still chugging away isn’t bad
[15:11:36] stuarta: yep, the core duo mini. ie the first intel based one
[15:12:09] stuarta: think i've got arch linux on it atm, might try sticking ubuntu on there instead
[15:12:17] stuarta: been months since i turned it on
[15:13:37] pvr4me: We still have a few (vocal) people wanting to use MacPorts on G5 Mac Pros with 10.5 after all these years.
[15:14:46] jpabq (jpabq! has quit (Remote host closed the connection)
[15:14:49] stuarta: hah, crazy people
[15:15:01] stuarta: that's like mythtv 0.18
[15:15:07] pvr4me: Does something need to be added to the .pro file to create a Mac app?
[15:15:23] stuarta: hmmm.... nothing in the .app folder?
[15:15:56] pvr4me: I did have 0.25 build and run on a G5. One lady was using it with Firewire for years
[15:16:40] stuarta: there is some extra stuff in the file specific to mac
[15:17:29] stuarta: i'd start with adding `TEMPLATE = app` to the .pro file for my test app
[15:17:52] jpabq (jpabq! has joined #mythtv
[15:24:26] pvr4me: Duh, sorry. It did build properly earlier. Neede to ‘open openglwidget’ to get it to run. Filled the screen completely, Dock hidden. Dock properly reappeared when I quit.
[15:24:52] stuarta: okay, so that proves that the full screen hint works properly on that install
[15:25:09] stuarta: and we are supposedly using it, definitely in master
[15:26:52] stuarta: doesn't look like the relevant code has changed between fixes/0.28 and master
[15:34:22] Roklobster (Roklobster! has quit (Ping timeout: 260 seconds)
[15:46:41] Roklobster (Roklobster! has joined #mythtv
[16:10:40] peterbennett (peterbennett!~Peter@mythtv/developer/peterbennett) has joined #mythtv
[16:12:54] peterbennett: I see much discussion on #13185 going on here while I have been typing into the ticket
[16:12:54] ** MythLogBot **
[16:13:13] peterbennett: Need to tie the ticket system into IRC :)
[16:13:47] peterbennett: jya: I cannot afford to buy an apple machine
[16:14:28] jya: peterbennett: christmas is near :)
[16:14:56] jya: peterbennett: sorry I haven't answered your email yet
[16:15:20] peterbennett: jya: Never mind I figured it out, did you see my second email?
[16:15:25] jya: oh...
[16:15:34] jya: I was about to answer that... don't worry about the refcount
[16:15:41] stuarta: ENOBUDGET :(
[16:15:54] jya: in fact for our use (and future) I'm not sure the refcount serves any purpuse
[16:15:57] jya: purpose
[16:16:04] peterbennett: stuarta: jya: I am working on fixing the deprecated calls to ffmpeg
[16:16:13] stuarta: nice
[16:16:28] jya: peterbennett: even avutils is making deprecated calls anyway
[16:16:44] jya: so why bother that no matter what you won't be able to resolve all of those warnings
[16:16:54] peterbennett: stuarta: jya: It seems to be not that hard. I went through the ffmpeg tutorial which stupidly still uses the deprecated calls and then changed them to the new calls
[16:17:28] ** stuarta chuckles at that definition of "not that hard" **
[16:17:54] peterbennett: Well my fear is that in a future version deprecated calls will be removed. Is that not how it is supposed to be done?
[16:18:15] stuarta: ffmpeg has actually removed stuff before
[16:19:19] jya: peterbennett: I mean, ffmpeg ref counting isn't thread safe...
[16:19:20] peterbennett: One concern is the cryptic comment in ffmpeg "The new API does not handle subtitles yet."
[16:19:30] jya: so it may be useless for us in the future anyway
[16:19:52] jya: peterbennett: there's plenty of code in ffmpeg that has been removed but we kept
[16:20:03] stuarta: jya: you original wrote mythframe to wrap AVFrame, so we should leverage that but with the new api
[16:20:19] jya: stuarta: that was why i did it in the first place
[16:20:55] stuarta: back to your comments recently that decode should be done in it's own thread, then it should be irrelevant that ffmpeg ref counting isn't thread safe
[16:21:08] jya: peterbennett: I have a mac mini 2012 somewhere (AMD device, with intel quad core)
[16:21:25] jya: stuarta: but even the packet out is refcounted
[16:21:38] jya: so you decode in one thread, you then want to paint that somewhere
[16:21:41] peterbennett: If ref counting isnt thread safe that should only be an issue if you use the same packet or frame in two threads
[16:21:43] jya: that won't be done in the same thread
[16:23:14] stuarta: could prove interesting
[16:24:03] peterbennett: So I figure – fix the deprecated calls first – then start looking at separate threads with a queueing mechanism.
[16:24:33] peterbennett: threads for demux, decode, render, and user interactions
[16:25:36] stuarta: unless the new api requires a rewrite anyway
[16:26:39] peterbennett: Actually it does not seem to require a rewrite. It does give some extra opportunities for improvement but you dont need to use them right away.
[16:27:46] peterbennett: The decode call which passes packets in and gets them out all in one call is changed to two calls so that should make it easier to use separate threads
[16:27:59] stuarta: nice
[16:28:17] peterbennett: For the first step I just put both calls next to each other to be like it was with the combined call
[16:28:57] stuarta: good idea
[16:30:49] peterbennett: One thing that may be problematic is fixing the calls in places like nuppelvideo, and crystal HD that I do not use and dont know how to use, dont even build.
[16:32:00] peterbennett: also privatedecoder_vda I dont know what that is but it uses ffmpeg deprecated stuff
[16:32:30] natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has joined #mythtv
[16:40:27] gigem: I don't think anybody still uses crystal hd. Not many did even when it was new. It was a very short term solution to a problem that no longer exists. IOW, I say remove it.
[16:40:47] gigem: The same probably goes for nupplevideo, though, not for the same reasons.
[16:41:03] stuarta: personally i'd replace nupplevideo recorder with matroska format containers
[16:41:17] stuarta: not that i know *anybody* who still uses something that requires such a thing
[16:41:19] gigem: Hmm, that doesn't sound right. The first 'same' is in reference to 'remove it'.
[16:41:45] stuarta: although maybe things like web cams do need some encoding to disk
[16:42:29] gigem: Right. I can't imagine anyone still has an analog recorder for creating nupplevideo files. And anybody who still has nupplevideo recordings can transcode them into some more modern format.
[16:45:45] peterbennett: I am wondering what will happen when cable cards disappear. There are USB capture devices for HDMI, and John Poet said he should be able to make them work, that would likely require encoding at the recorder
[16:46:06] stuarta: they will be like webcams i suspect, which just throw out frames
[16:46:32] peterbennett: Those USB devices that support Linux, as you say, are like webcams.
[16:47:15] gigem: Really? I thought they generated h.264.
[16:47:23] peterbennett: But probably the nuppelvideo should rather be x264 encoded TS
[16:47:35] stuarta: gigem: suspect it depends on the device
[16:47:42] stuarta: webcams tend to be a v4l2 devices
[16:47:49] stuarta: -s
[16:50:56] peterbennett: Those ones I saw on Amazon, they say they support UVC, not usre what that entails, but since they are all USB3 it may be uncompressed video
[16:52:53] peterbennett: s/usre/sure/
[16:53:15] stuarta: i would file that whole section under "future work"
[16:55:14] gigem: That would be a lot of data, and probably require a reasonably beefy cpu to encode in real time. That probably also explains why they only provide stereo audio (they're very simple devices).
[17:02:59] peterbennett: gigem: I didnt see where it specified stereo audio, but I wonder what happens if the hdmi port is transmitting AC-3, will the usb device just lose the audio?
[17:04:27] stuarta: isn't ac3 encoded into a single bitstream that then gets split out anyway?
[17:04:34] stuarta: maybe they just downconvert
[17:04:41] stuarta: probably pretty trivial in silicon
[17:05:50] peterbennett: So why would the simple device go to the length of downconverting rather than just passing through? That does not make sense.
[17:06:13] stuarta: i'm speculating
[17:06:23] stuarta: probably better if we had actual information
[17:08:16] peterbennett: anyway, I would take HD with stereo rather than nothing at all.
[17:09:11] ** stuarta finds the "virtual video test driver (vivid)" . . . s/vivid.html **
[17:10:43] stuarta: peterbennett: this may also be of interest
[17:10:50] stuarta: using the rpi as a hdmi capture device
[17:20:52] peterbennett: Not sure I understand. Does it somehow turn the hdmi output into an hdmi input?
[17:21:27] stuarta: that's what i understood from skim reading it
[17:24:19] peterbennett: That would be excellent if it works.
[17:25:21] gigem: peterbennett: I was just repeating what some on the mailing list said. Some said it was a deal breaker for them. I'm with you. Stereo audio is better than no audio (and no video). My speculation is this. If the device looks lie a UVC camera, it's likely very simple and stupid. Ergo, it tells whatever it's connected to by hdmi that it's a simple device and only accepts stereo audio which it then passes
[17:25:23] gigem: on over usb. Again, that's all speculation on my part. I have no idea how any audio is passed over hdmi.
[17:26:38] stuarta: that would make sense
[17:26:49] peterbennett: Possibly you would have to set your output device (e.g. cablebox) to stereo to get any audio.
[17:27:06] stuarta: hdmi (as a protocol) is capable of negotiating the audio parameters that are acceptable
[17:27:14] stuarta: iirc
[17:28:24] stuarta: right, i'm off
[17:38:55] peterbennett (peterbennett!~Peter@mythtv/developer/peterbennett) has quit (Quit: Leaving.)
[17:43:28] Merlin83b (Merlin83b! has quit (Quit: Leaving)
[17:58:50] willcooke (willcooke!~willcooke@ubuntu/member/willcooke) has quit (Quit: Do your hobbies)
[17:58:57] benklop (benklop! has quit (Ping timeout: 240 seconds)
[18:08:11] benklop (benklop! has joined #mythtv
[18:34:09] benklop (benklop! has quit (Ping timeout: 250 seconds)
[18:35:45] benklop (benklop! has joined #mythtv
[19:10:42] gregl (gregl! has joined #mythtv
[21:50:34] SteveGoodey (SteveGoodey! has quit (Quit: Konversation terminated!)
[22:36:03] pvr4me (pvr4me! has quit (Quit: pvr4me)
[23:26:30] amessina (amessina!~amessina@unaffiliated/amessina) has joined #mythtv
[23:47:02] Roklobster (Roklobster! has quit (Remote host closed the connection)

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