:: #mythtv

Daily chat history

Current users (57):

aloril, Anssi, blahdodo, brfransen, ChanServ, clever, davic, dizygotheca, dym, eharris, ElmerFudd, enyc, frobnic, gary_buhrmaster, ghoti, gigem, gregl, GreyFoxx, Hydr0p0nx, Hydroponx, ijc, jheizer, jpabq, jpharvey_, jya, kalamaja, knowledgejunkie, libsci, mad_enz, Maliuta_, markspieth, markspieth2, MitchCapper, mkbloke, MythBuild, MythLogBot, MythNotifyBot_, nephyrin, ooshlablu, Panic, peper03_, peterbennett, poptix, pppingme, ramshadow, rhpot1991, sdrik, sphery, stuarta, stuartm, taylorr, tonsofpcs, tris, Warped, xris, zbot`, _charly_
Sunday, January 13th, 2019, 00:02 UTC
[00:02:04] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Read error: Connection reset by peer)
[00:03:52] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv
[01:22:05] ramshadow (ramshadow! has quit (Remote host closed the connection)
[01:23:36] ramshadow (ramshadow!ramshadow@2a02:f98:0:20:5054:ff:fe3f:e989) has joined #mythtv
[02:56:04] hampton (hampton!~dhampton@mythtv/developer/hampton) has quit (Quit: Leaving.)
[05:48:30] gigem (gigem!~david@mythtv/developer/gigem) has quit (Ping timeout: 246 seconds)
[05:48:53] gigem (gigem!~david@mythtv/developer/gigem) has joined #mythtv
[05:48:53] Mode for #mythtv by ChanServ!ChanServ@services. : +v gigem
[06:56:18] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Ping timeout: 252 seconds)
[07:02:09] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv
[07:26:39] RokLobsta (RokLobsta! has quit (Remote host closed the connection)
[08:03:16] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv
[08:03:17] Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey
[08:35:24] Tobbe5178 (Tobbe5178!~asdf@2001:2002:51eb:d24e:c3a:d7e2:4443:98da) has joined #mythtv
[09:04:37] mark____ (mark____! has joined #mythtv
[09:20:24] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Ping timeout: 252 seconds)
[09:22:12] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv
[09:22:53] mark____: stuarta: I suspect you've been asked before. Please don't bite my head off:) Is there a reason we don't use something like travis for CI?
[09:40:49] warpme_ (warpme_! has joined #mythtv
[09:47:45] mark____ (mark____! has quit (Quit: mark____)
[11:50:37] mark____ (mark____! has joined #mythtv
[11:55:17] warpme_: Hi mark___
[11:55:33] mark____: warpme_: morning
[11:55:57] mark____: i sense a question coming... :)
[11:57:07] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Ping timeout: 250 seconds)
[11:57:10] warpme_: I?m wonder do You have plans (in long term) to suuport video pipeline on arm platforms – where display, video and 3d blocks are discrete and can be glued by kernel DRM subsystem?
[11:58:45] mark____: I assume you mean non-android?
[11:59:13] warpme_: yes. I mean linux os running on such SoCs
[11:59:30] warpme_: linux – kind of GNU/Linux
[12:00:23] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv
[12:02:42] mark____: well that depends... I, personally, wouldn't want to maintain that code if every SoC has a different implementation. That is essentially where we are with the Pi support but using the broadcom wrappers.
[12:04:37] warpme_: well – pls look here:
[12:04:37] ** MythLogBot **
[12:05:51] warpme_: this is a bit innaccurate – but I think future is v4l2 m2m and narrowing support only to v4l2 m2m might be resonable....
[12:09:31] mark____: warpme_: first priority is android zero copy. Otherwise if there is FFmpeg support for v4l2 m2m then – then both GLES and DRM prime interop should be feasible – esp. once we're up to speed on the DRM code needed for VAAPI interop.
[12:10:52] warpme_: saying android zero copy – do You mean i.e. dambuf between ffmpeg decoded frames and drm?
[12:12:44] mark____: I'm not sure at the moment – just haven't had time to look at it yet – but I'd be amazed if android doesn't have support for MediaCodec<->GLES
[12:19:18] warpme_: pls look at this: If I?m reading code properly – this set of patches atop of ffmpeg4.02 uses dambuf between ffmpeg and drm. Advantages are imho: 1\zero copy transfers between decoded frames and disaply subsystem; 2\no gles3 reuired; 3\exploits drm (which is imho really nice technology).
[12:19:49] warpme_: this youtube video is nice about subject:
[12:28:21] warpme_: look at video from around 16:12 time
[12:34:09] mark____: warpme_: it'll take me a while to digest that lot. but I think, from the video, the subsequent discussion around egl rendering is key. We also need/want GLES for the whole video/ui presentation.
[12:39:19] warpme_: Indeed – gles is a must. But for me UI is one thing and 4k60p playback is another. Teoretically gles path with dmabuf_import for video and thesame gles for ui will be most consistent – but probably gles itself might be problem. iirc gles v3 is required for this and iirc only mesa provides v3 today
[12:41:27] mark____: warpme_: but don't forget we need a way of displaying overlays i.e. OSD – so UI and video cannot be entirely seperate.
[12:41:36] warpme_: so looking on mesa on arm socs: quite good is freedreno for adreno gfx (qualcom), samsung exynos also is quite good – but both are so expensive that x86 SFC with HD500+ is comparable in price and works perfectly.
[12:45:31] warpme_: other SoC with really good price/features ratio (s905 & rockchip) are with mali450 and only choice for gles is lima (amlogic is expiring this year licenses – so only lima will be chioce). Q is: will gles implementation on mali450 be good enough to display 4k@60p via gles dmabuf_import path? I doubt so this redirect me to arch with drm prime (video via drm plane); ui/osd via gles
[12:47:20] warpme_: both (video and osd) mixed on screen via drm subsystem
[12:58:10] mark____ (mark____! has quit (Quit: mark____)
[13:20:58] SteveGoodey (SteveGoodey!~steve@2a00:23c5:7da3:4501:7a24:afff:fe9d:c233) has joined #mythtv
[13:20:59] Mode for #mythtv by ChanServ!ChanServ@services. : +v SteveGoodey
[14:09:44] jpabq (jpabq!~quassel@mythtv/developer/jpabq) has quit (Ping timeout: 244 seconds)
[14:10:13] jpabq (jpabq!~quassel@mythtv/developer/jpabq) has joined #mythtv
[14:10:14] Mode for #mythtv by ChanServ!ChanServ@services. : +v jpabq
[15:02:44] warpme_ (warpme_! has quit (Quit: warpme_)
[15:48:57] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Ping timeout: 250 seconds)
[15:52:52] nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv
[16:13:39] mark____ (mark____!~mark@ has joined #mythtv
[16:22:18] warpme_ (warpme_! has joined #mythtv
[16:24:48] warpme_: btw: if anybody is interested in drm details: . . . m-subsystem/
[16:43:34] warpme_ (warpme_! has quit (Quit: warpme_)
[16:45:12] warpme_ (warpme_! has joined #mythtv
[16:56:48] warpme_ (warpme_! has quit (Quit: warpme_)
[18:56:58] mark____: anyone have an idea of what is still using the deprecated UI dialogs? (ignoring plugins for now)
[19:10:46] peterbennett: mark____: Hi
[19:23:47] mark____: peterbennett:: evening
[19:25:29] peterbennett: mark____: This NVDEC stuff uses glew and glut
[19:25:29] stuarta: mark____: re travis, it's something that hasn't been looked at for some time
[19:25:48] stuarta: i may revisit in the future, but right now buildbot serves our purposes quite well
[19:25:48] mark____: peterbennett: arrgh
[19:26:32] mark____: staurta: I was only curious really – prompted by the amount of time you seem to spend kicking the bots
[19:27:01] stuarta: mark____: that's mainly because we've moved to a new version which requires pretty significant changes to the config
[19:27:14] stuarta: mark____: trac is the main problem right now
[19:27:29] stuarta: errors out on 90% of the commits rather than processing them
[19:28:14] stuarta: mark____: on the buildbot front i'm pretty happy with where we are, main bounces lately are to add/remove workers and to ensure the pips for the new version were installed
[19:28:15] peterbennett: GL Extension Wranglet and The OpenGL Utility Toolkit
[19:29:09] peterbennett: In spite of using all those tools that purport to make openGL easier it looks very complicated
[19:29:21] mark____: peterbennett: it was just the prospect of adding more libraries – hence the arrgh...
[19:31:51] peterbennett: mark____: I imagine that those are not needed, they just make things a bit easier – but there are other things needed for this NVDEC stuff – nvidia-cuda-toolkit
[19:35:30] peterbennett: I am wondering what the difference is between glGenBuffers, geGenBuffersARB, __glewGenBuffersARB
[19:36:38] peterbennett: At least I don't have to worry about Android for the NVDEC stuff
[19:51:56] mark____: peterbennett: I'm guessing those are all the same – just wrappers
[19:52:37] mark____: stuarta: if working, how long do the trac/github hooks take to work
[19:53:01] peterbennett: mark____: Yes – glut is only used for creating a window and associating it with opengl – we already have that
[19:53:50] mark____: peterbennett: when you say NVDEC uses glew/glut – is that just the examples?
[19:53:53] gigem: mark____: I thought Roger Siddons recent changes removed the last remaining code using the old settings. Or were you asking about something else?
[19:54:51] peterbennett: mark____: Yes – those are used by the examples – I was hoping to steal some code from the examples – so I will have to get rid of the glew and glut stuff out of teh code I steal.
[19:56:01] mark____: gigem: no – that was exactly what I was after. It's just that I removed the deprecated code from mythmainwindow and that then caused issues with mythdialogs – but then various other classes seem to be using mythdialogs... mythwizard etc
[19:56:25] mark____: just not sure if all of those classes are deprecated
[20:04:16] peterbennett: There is some ui function that pops up a deprecated message during compile whenever it is used. I forget what it is. YOu should be able to see it by doing a full make
[20:11:10] peterbennett: mark____: Oh dear, I cannot believe it – the NVDEC sample program uses OpenGL 1 :(
[20:11:24] peterbennett: Those ARB functions are all OpenGL 1
[20:20:02] gigem: mark____: I don't know anything about those dependencies. Sorry.
[20:34:33] stuarta: mark____: trac hooks should work pretty much straight away. i can almost guarantee they've broken again
[20:34:42] ** stuarta goes to kick them **
[20:35:51] stuarta: peterbennett: a lot of the plain opengl examples use glew/glut to get access to opengl. you can do away with that when you use the Qt5 opengl interfaces
[20:36:46] stuarta: peterbennett: i'd suggest comparing some of my opengl test programs with the typical opengl example programs that also draw the triangle
[20:37:02] amessina (amessina!~amessina@unaffiliated/amessina) has quit (Quit: Konversation terminated!)
[20:37:18] stuarta: one of the reasons i worked through those examples, is that i could not find anywhere examples of the opengl triangle, which used Qt5 rather than glew/glut
[20:48:22] stuarta: mark____: apologies for the duplicate emails. 3 attempts and the webhook finally completes successfully
[20:52:49] stuarta: i'm seriously considering trialling redmine as a replacement to trac
[21:00:15] amessina (amessina!~amessina@unaffiliated/amessina) has joined #mythtv
[21:11:05] stuarta: mark____: i believe we should add you new branch to the build network
[21:41:25] peterbennett: stuarta: Do you think we should clean up some of the old unused branches?
[22:03:37] gregl (gregl! has joined #mythtv
[22:35:41] amessina (amessina!~amessina@unaffiliated/amessina) has quit (Quit: Konversation terminated!)
[22:36:53] Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)
[23:02:40] SteveGoodey (SteveGoodey!~steve@2a00:23c5:7da3:4501:7a24:afff:fe9d:c233) has quit (Quit: Konversation terminated!)
[23:07:39] davic (davic!~davic@unaffiliated/davic) has quit (Ping timeout: 268 seconds)
[23:08:17] davic (davic!~davic@unaffiliated/davic) has joined #mythtv
[23:09:17] Tobbe5178 (Tobbe5178!~asdf@2001:2002:51eb:d24e:c3a:d7e2:4443:98da) has quit (Read error: Connection reset by peer)
[23:23:56] mark____: stuarta: thanks – all sorted. re adding branch to builds – maybe give it a week or two until it has settled down
[23:25:53] dizygotheca (dizygotheca! has joined #mythtv
[23:27:02] mark____ (mark____!~mark@ has quit (Quit: mark____)
[23:29:18] dizygotheca: mark____: Yes I have a patch that removes MythDialog. It can go in master. Do you want it in your branch as well. Or instead ?

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