:: #mythtv

Daily chat history

Current users (83):

aloril, Anduin, anykey_, Beirdo, ben1066, brfransen, brtb, CaCtus491, Captain_Murdoch, cattelan, Chutt, clever, coling, Cougar, damaltor, danielk22, Dave123, dekarl, dlblog, ElmerFudd, ertyu-m, foobum, foxbuntu, frankster, ghoti, gigem, gregL, GreyFoxx, highzeth, J-e-f-f-A, j-rod|afk, jams, jarle, jcarlos, joe_, joki, jpabq-, jstenback, justinh, jwhite, jya_, k-man, knightr, kormoc, kurre2, kwmonroe, laga, mag0o, MavT, mike|3, mrand, MythBuild, MythLogBot, mzanetti, peitolm, pheld, poptix, purserj, rhpot1991, skd5aner, Slasher`, sphery, sraue, stuarta, stuartm, superm1, sutula, taylorr, tgm4883, The_Ball_, ThisNewGuy1, tomimo, tris, Unhelpful, vallor, wagnerrp, wahrhaft, wseltzer, xavierh, xris, yb0t, zCougar, _charly_
Thursday, March 15th, 2012, 00:04 UTC
[00:04:05] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 248 seconds)
[00:13:25] superm1 (superm1!u4318@ubuntu/member/superm1) has quit (Ping timeout: 265 seconds)
[00:16:06] jya: sphery: did you get the chance on trying the OpenGL default profile ?
[00:19:51] Dave123 (Dave123! has joined #mythtv
[00:20:14] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[00:27:35] jya: stichnot: IMHO, this is a "bug" worth fixing
[00:27:50] gigem (gigem!~gigem@mythtv/developer/gigem) has quit ()
[00:30:12] tester68 is now known as gigem
[00:30:59] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection)
[00:34:34] knightr: wagnerrp, We got a question from the one of translators... What is still required in " Note, in IPv6 setups, this is still required for certain extras such as UPnP." (that an backend setup string)? My guess is it's still required to provide an externally accessible address but I just want to be sure of it before I tell him that...
[00:34:39] knightr: Thanks!
[00:35:04] wagnerrp: knightr: UPNP is an IPv4-only protocol
[00:35:07] superm1 (superm1!u4318@ubuntu/member/superm1) has joined #mythtv
[00:35:28] wagnerrp: autodetection operates on 239.something
[00:35:46] wagnerrp: which is used to then determine address and ports for machines to communicate with each other
[00:35:49] knightr: ah, I didn't know... thanks
[00:35:55] wagnerrp: for mythtv's internal purposes, it doesnt care
[00:36:03] knightr: it's in the range for multicast stuff IIRC?
[00:36:18] wagnerrp: if its told the backend wants to communicate on an ipv6 address, it will do so, and backend autodetection will continue to operate as normal
[00:36:41] wagnerrp: for little network media players, they will require an IPv4 address to communicate on
[00:37:11] wagnerrp: yes, 224/8–239/8 is multicast
[00:37:12] sphery: jya: not yet, will try a bit later tonight
[00:37:51] wagnerrp: knightr: anyway, its just a notice that certain external functions may not work correctly if an IPv4 address is not specified
[00:38:07] jya: ok… let me know how it works.. I would expect OpenGL Standard to give good result for most. This is what I was using as profile before VDPAU came along, that was using a 9400GT card and a 2.6GHz Core 2 Duo
[00:39:19] knightr: wagnerrp, thank you! I re-read the string in question in context and now I understand perfectly what you mean. Thank you!
[00:39:48] jya: sphery: according to my notes from 2008, using that profile, when I was playing 1080i mpeg2 + deinterlacing, the system as 83% idle
[00:41:35] Guest15634 (Guest15634! has joined #mythtv
[00:41:50] gigem: sphery: i should just have to put the droid fonts and Apache license in the themes/fonts or themes/MythCenter-wide directory, right? does anyone have a preference?
[00:42:46] gigem: fyi all, i should always be seen as just gigem from now on. you might still davide lurking for a few more days, though.
[00:43:09] kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has quit (Quit: kormoc)
[00:44:55] sphery: gigem: we use Droid Sans for ATSC captions at . . . een.cpp#L102 (but in a style that's seldom--if ever--used by broadcasters)... Other than that (and because caption font selection will be changing after 0.25), the only use I know of is in the theme, so I'd suggest putting it in the theme.
[00:45:11] sphery: and it can be anywhere under the theme dir--so feel free to put in a subdir for fonts or whatever
[00:46:02] gigem: sphery: okay.
[00:46:21] sphery: should it be this license? . . . s/README.txt
[00:46:31] sphery: and/or . . . fonts/NOTICE
[00:46:38] sphery: (or perhaps that's what you meant?)
[00:46:38] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has quit (Remote host closed the connection)
[00:47:14] stuartm (stuartm! has joined #mythtv
[00:47:14] stuartm (stuartm! has quit (Changing host)
[00:47:15] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has joined #mythtv
[00:47:51] sphery: And, btw, thanks for taking care of it. It was on my "before 0.25" list, so one less thing outstanding for me.
[00:54:16] gigem: sphery: yes, it's just the Apache 2.0 license. i'm not terribly fluent with our font support, so bear with me for two more question. do we only need the DroidSans.ttf file or do we need all of the other DroidSans*.ttf variants, too? and should i include the DroidSerif*.ttf variants for completeness?
[01:02:43] Guest15634 (Guest15634! has quit (Read error: Connection reset by peer)
[01:03:22] sphery: I think with the DroidSans.ttf, it will mostly work, but if the theme uses bold or other variants, it will either "fake" it with that font or use a substitute font
[01:04:33] sphery: and, fwiw, MythCenter-wide /does/ use bold (but not italic, it seems)
[01:04:42] sphery: so I'd say include the bold one, too, at least
[01:05:15] davide (davide! has quit (Remote host closed the connection)
[01:05:38] davide (davide! has joined #mythtv
[01:06:02] sphery: your call on the Armenian, Ethiopic, Georgian, Hebrew, Japanese, and Thai ones
[01:07:04] sphery: seems Italic and BoldItalic and Mono aren't necessary, though
[01:09:57] stuartm: fwiw, the core of the change in means we can now start tagging audio tracks with "(Commentary)" in the track selection menu, it's a pretty safe change with a good WAF but I don't want to make it during the feature freeze without a consensus
[01:19:54] jya: stuartm: how is that not committing without a consensus ? That seems like a commit on master to me :)
[01:23:12] jason (jason! has joined #mythtv
[01:23:39] jason is now known as Guest32780
[01:24:19] stuartm: jya: huh?
[01:25:20] stuartm: that commit fixes a bug, but as I said, the code I had to add to fix said bug means that we can now easily build on it to add that information in the UI
[01:25:46] jya: looks good to me for DVD.. it stays the same for other types of video, which I wouldn't be surprised makes now the large majority of media being b=played. it seems widely accepted that DVD are on their way out
[01:26:43] stuartm: the patch to add the information in the menu looks something like this –
[01:27:18] jya: that too i think is a worthy change
[01:27:29] pheld (pheld! has quit (Quit: Leaving.)
[01:29:20] jya: one about if a DVD has incorrect tag, and when you filter for kAudioTypeNormal you don't get an empty list, but a list containing say just commentary track. How would you know select one of the other track?
[01:30:19] jya: hum… forget that question.. this is only for the autoselection
[01:30:23] Guest32780 (Guest32780! has quit (Read error: Connection reset by peer)
[01:31:36] gigem: sphery: i went with the bare bones. DroidSans.ttf and DroidSans-Bold.ttf only.
[01:31:54] stuartm: jya: so 'normal' tracks described as 'commentary' and 'commentary' tracks described as 'normal'? I'd assume that's not going to happen because this metadata is what a commercial DVD player uses to determine the correct tracks, but this only affects the auto-selection and doesn't stop someone changing the tracks via the menu
[01:32:06] sphery: gigem: looks good... thanks
[01:34:05] stuartm: all the DVDs I tested had the commentary tracks correctly identified, I'm happy that it would have fixed the disc(s) where the wrong audio track was being selected but I can't test that one because they were rental and I no longer have them
[01:34:59] stuartm: but if anyone in the UK has the last disc of Dexter, series 4 and wants to check then I'd love to know
[01:35:44] stuartm: I think it was series 4 ... commentary with John Lithgow
[01:36:55] jya: stuartm: so you are the person renting DVD in the UK :)
[01:37:46] sphery: stuartm: do you happen to know if the transcoding profiles setup got lost when you moved recording profiles to mythtv-setup? re:
[01:37:49] jya: I actually vote to get the audio description text amended so it shows what types of tracks they are
[01:38:05] stuartm: jya: less hassle than blu-ray when it comes to playback via mythfrontend
[01:38:34] sphery: I haven't looked, so I'm not saying they were lost--just that it seems Steve isn't going to find them without more help
[01:38:40] jya: uhoh, my new iPad has shipped !
[01:39:17] stuartm: sphery: they are still there
[01:41:37] stuartm: sphery: that reminds me, we need to get some description strings into the mythtv-setup menu before the translation freeze
[01:41:39] jya: I think the question is where… Steven Adeff on the mailing list, seems to have been unable to locate them
[01:41:56] stuartm: they were missed when we did mythfrontend however many years ago that was now
[01:48:00] stuartm: jya: under Recording Profiles, in mythtv-setup
[01:48:42] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv
[01:49:22] jya: stuartm: yes.. that's where he was told to look into, but he says it's not there
[01:53:36] kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has joined #mythtv
[01:55:12] sphery: jya: those opengl profiles actually work... it seems that when I last tried it it was unable to load the hw deinterlacers (some problem loading shader program) so it was falling back to a software deint, but it works now, so I'm guessing my nvidia driver upgrade fixed it
[01:55:49] sphery: I get like 22% CPU use at 1x or 34% CPU use at 1.75x
[01:56:28] taylorr: danielk22: I found a pretty serious flaw in the ringbuffer for livetv transitions – if you look at RingBuffer::WaitForAvail() you'll notice that the loop inside that waits for data doesn't check livetv is set to switch – it stays in there 16 secs
[01:56:36] jya: sphery: which one did you use?
[01:56:39] sphery: (actually up to about 63% CPU at 1.75x)
[01:56:42] sphery: I'm doing slim, now
[01:56:43] jya: Best is actually a software deinterlacer
[01:57:01] jya: Standard and Slim are all hardware accelerated
[01:57:07] jya: Slim actually doesn't look to bad
[01:57:16] jya: it's actually better than what VDPAU Slim would do
[01:57:18] taylorr: danielk22: if you checkout fixes/0.23 and look at RingBuffer:ReadFromBuf() you'll see that it handles that properly and the refactor doesn't
[01:57:46] sphery: I think I'll stick with the vdpau renderer, but these seem to be usable, so should go in as examples
[01:58:02] jya: oh, if you have vdpau , for sure, stick with it
[01:58:19] jya: vdpau high or standard are way better than the opengl ones
[02:00:11] stichnot: jya: I agree that #10161 should be "fixed", but my current patch is still a bit too icky for most people's taste (including mine)
[02:00:54] sphery: seems high works, too... about 60% at 1x and 96–105% at 1.75x
[02:01:21] jya: stichnot: on my iMac , which is a screamingly fast machine, when I enter the watch recoding screen I have to wait about 6s.. With the CPU going through the rood
[02:01:48] stichnot: how many recordings in the recording group it's trying to display?
[02:01:49] sphery: greedy is definitely a good deint
[02:02:53] stichnot: and jya, have you tried the patch, and how much does it help?
[02:03:07] jya: sphery: kernel and linear HW accelerated have tiny artifacts on scrolling texts. it adds pixels here and there. this is due to the motion detection algortihm used. for most stuff, I find Greedy and Kernel 2X virtually indistinguishable. But Greedy doesn't have those wierd pixel.. it is eavy on the CPU though
[02:03:33] jya: stichnot: I haven't yet… I'm trying to sort out why since I compiled Qt 4.8 , my mac version doesn't compile in 32 bits
[02:04:53] stichnot: jya: how many recordings? I'm curious how the performance compares against my 2500 recordings
[02:05:19] sphery: jya: thanks for coming up with those profile groups... I think it will be very useful being able to tell users "try OpenGL Slim" or whatever once they go in
[02:05:19] jya: stichnot: let me check.. I have 6TB worth :)
[02:05:27] stichnot: ok, me too :)
[02:05:58] jya: sphery: did you read my post a looong time ago comparing them all? . . . #post1814111
[02:06:05] jya: that was before VDPAU and all
[02:07:20] sphery: yeah, that was definitely interesting
[02:08:16] jya: stichnot: can I see the number of recordings somewhere? mythweb gives me: 7,587,089 MB worth
[02:08:19] sphery: I didn't check a 720p recording... should I do that?
[02:08:43] jya: sphery: well, that was a bug somewhere.. because it shouldn't have even mattered as it's not interlaced to start with
[02:08:55] jya: that was 4.5 years ago… it's probably no longer relevant toda
[02:08:57] sphery: jya: watch recordings should tell you how many... or in mythweb, bottom of the recorded programs page
[02:09:05] sphery: ok, I'll assume it's good now
[02:09:13] jya: and FWIW, I've watched the HD channel that gave me problem back in those days on the mac
[02:09:16] jya: and it was all fine
[02:09:56] jya: :) my recorded mythweb page is still loading after 2 minutes :)
[02:10:18] jya: ah done: 2690 programs
[02:10:36] jya: 2690 programs, using 5.7 TB (3 months 12 days 18 hrs 44 mins) out of 7.3 TB (42 GB free).
[02:11:10] jya: sphery: ok.. I'm going to commit the changes then.. thanks for testing them
[02:26:50] Yancho (Yancho!~mpulis@unaffiliated/yancho) has quit (Remote host closed the connection)
[02:44:24] wriley (wriley! has joined #mythtv
[02:44:45] wriley (wriley! has left #mythtv ()
[02:45:01] The_Ball (The_Ball!~The_Ball@ has quit (Ping timeout: 260 seconds)
[02:56:34] stichnot: jya: you're the winner! 2398 programs, using 4.2 TB (2 months 14 days 12 hrs 17 mins) out of 5.4 TB (587 GB free).
[02:57:16] jya: I've fixed my compilation issue… I'll be able to test your patch in a couple of minutes
[02:58:35] stichnot: cool. There's some extra logging, search for "LZY", that is designed to reveal some rough timings. Part of it is waiting for the data from the backend, then the part the patch addresses is to defer the massive amount of string processing
[02:58:39] Beirdo: hehe
[02:58:51] taylorr: danielk22: ok, I triggered that problem because I increased the livetvchain safe_read timeout to be a few seconds – probably not an issue
[02:59:25] Beirdo: 1294 programs, using 2.9 TB (1 month 15 days 5 hrs 45 mins) out of 11 TB (7.6 TB free).
[02:59:40] Beirdo: I'm getting close too. Bet sphery's still the weiner
[03:00:12] jya: f…ck, a git pull --rebase screwed up, I had to redo 3 commits, and now trying to pull just give me You are not currently on a branch, so I cannot use any
[03:00:12] jya: 'branch.<branchname>.merge' in your configuration file.
[03:00:37] Beirdo: you did the git rebase --continue?
[03:00:54] jya: I tried… but it would say I didn't do any git add
[03:01:03] jya: and git status showed me that I had no modified files
[03:01:11] Beirdo: oO
[03:01:25] jya: so I couldn't do git add on anything
[03:01:50] jya: oh well.. look like I'll have to export as patch my existing ones, and git am them
[03:01:56] Beirdo: heh, you're talented. I've never made it do that
[03:02:07] Beirdo: that could work
[03:02:10] Beirdo: or git stash it
[03:02:23] Beirdo: oh wait
[03:02:26] Beirdo: already committed
[03:02:27] jya: I can stash an existing commit ?
[03:02:44] jya: basicually, my git rebase put me into a non branch mode.
[03:02:48] jya: I hadn't noticed
[03:02:53] jya: so I continued my commit
[03:03:10] Beirdo: the other thing to try
[03:03:16] jya: mercurial you could convert a commit into a stash
[03:03:27] Beirdo: git commit -b new-branchname
[03:03:29] jya: that was pretty cool, you could easily modify an existing commit
[03:03:35] Beirdo: that will put you back onto a branch
[03:03:44] Beirdo: so you can get back here should you need ot
[03:03:48] Beirdo: to rather
[03:03:58] jya: I think I like the patch option better, at least I know eactly what I'm doing
[03:04:00] Beirdo: then checkout master and try merging from that branch?
[03:04:07] Beirdo: heh, fair enough
[03:04:28] jya: I never know how to call format-patch though
[03:04:51] Beirdo: git format-patch SHA1-before-the-first-patch-you-want
[03:04:54] Beirdo: :)
[03:05:04] Beirdo: or git format-patch master
[03:05:18] jya: easy !
[03:05:24] jya: 6 patches now
[03:05:26] Beirdo: as long as it creates teh files ya want.
[03:05:44] jya: I think my rebase issues crashed because I had done an git am earlier
[03:05:52] jya: and it thought I had a git am on the way
[03:05:57] Beirdo: oooh
[03:06:07] jya: as I don't know how to cancel the git am ...
[03:06:11] Beirdo: yeah, if the am was still in progress... it shoulda told you
[03:06:17] jya: (didn't even know I had to)
[03:06:47] Beirdo: git am --abort to cancel
[03:06:47] jya: git checkout master
[03:06:47] jya: Warning: you are leaving 6 commits behind, not connected to
[03:06:48] jya: any of your branches:
[03:06:56] jya: bye bye commits
[03:06:59] Beirdo: heh
[03:07:06] Beirdo: well, you have em as patches
[03:07:11] Beirdo: and they are still there anyways
[03:07:14] jya: ah !
[03:07:26] jya: I did git checkout master, and all my commits prior to the rebase are there !
[03:07:44] jya: I know I'm going to have a conflicts as wagnerrp updated the ABI
[03:07:52] Beirdo: git am --continue to continue after fixing a conflict
[03:09:52] sphery: hehe, actually I'm well below jya and stichnot ... I only have 1894 programs (but then again, I watch and delete everything--never archive), so only reason I have a lot is because I record more than I can watch
[03:10:22] Beirdo: wow.
[03:10:31] Beirdo: I'm catching ya fast
[03:10:33] jya: I never delete anything :) my wife does.. but that's annoy me.. What if I wanted to watch that show in say a year from now
[03:11:26] Beirdo: heheh
[03:11:39] Beirdo: you're insane :) not that there's anything wrong with that
[03:12:45] jya: I had to rent external storage space the day my wife forced me to throw away all my boxes and old manuals… I just keep everything
[03:12:54] jya: it did free up a lot of storage
[03:14:09] Beirdo: jya, le collectionneur
[03:14:47] Beirdo: not that I'm much different
[03:16:47] jya: Well, that way I'm sure I get a better price when I sell stuff on ebay!
[03:17:00] jya: always come with box and all original manuals
[03:18:37] Beirdo: hehe
[03:19:00] jya: amazing how people can prostitute themselves… . . . 226298374932
[03:25:11] jya: taylorr: here is what I'm now seeing watching liveTV
[03:25:12] jya:
[03:26:08] jya: if I pause for a 3–4s, and play again, it's all fine. This is quite recent.. never used to do that
[03:26:23] jya: it seems to match the problem you were mentioning that it doesn't buffer enough data
[03:27:25] jya: when I start livetv, I'm only 2s before actual time.
[03:29:38] taylorr: jya: yes, probably related – I'm more interested in the corruption than anything
[03:30:10] taylorr: probably could increase fill_min in RingBuffer and it would help
[03:31:36] jya: I don't see any corruptions… it just stutters
[03:41:44] jya: stichnot: before your patch: entering watch recordings takes 6.7s, 0.9s with… great job !
[03:42:05] jya: on my mac mini frontend, it takes much more time than that
[03:50:11] cattelan is now known as cattelan_away
[04:30:20] The_Ball (The_Ball!~The_Ball@ has joined #mythtv
[04:35:00] taylorr (taylorr!~taylorr@unaffiliated/elmojo) has quit (Remote host closed the connection)
[04:36:27] taylorr (taylorr! has joined #mythtv
[04:36:28] taylorr (taylorr! has quit (Changing host)
[04:36:28] taylorr (taylorr!~taylorr@unaffiliated/elmojo) has joined #mythtv
[05:42:57] stichnot: jya: astounding that it takes 6.7s unpatched on a reasonable frontend. The only thing I can think of is that something is seriously inefficient in the QDateTime implementation there.
[05:43:53] stichnot: I think my unpatched ION frontend takes about the same amount of time
[05:44:05] jya: wagnerrp: i wouldn't have removed mythmessage IMHO, maybe rewrite it to ust be a call to mythutil…
[05:44:16] jya: stichnot: it does take a while...
[06:36:31] skd5aner (skd5aner! has quit (Ping timeout: 276 seconds)
[06:48:42] skd5aner (skd5aner! has joined #mythtv
[07:15:31] jpabq- (jpabq-! has joined #mythtv
[07:21:41] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (*.net *.split)
[07:21:41] abqjp (abqjp! has quit (*.net *.split)
[07:21:42] k-man (k-man!~jason@unaffiliated/k-man) has quit (*.net *.split)
[07:22:06] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[07:27:06] k-man (k-man!~jason@unaffiliated/k-man) has joined #mythtv
[07:58:33] warped (warped!~piotro@ has joined #mythtv
[08:14:06] dekarl: stuartm: wrt the don't autoselect audio description (might actually be nice to have an option to toggle it in the other direction), does the change apply to mythtranscode, too? Ending up with AD after lossless transcode is always a surprise :)
[08:26:04] natanojl (natanojl! has joined #mythtv
[08:47:29] pheld (pheld! has joined #mythtv
[08:50:25] natanojl (natanojl! has quit (Ping timeout: 260 seconds)
[08:51:38] warped (warped!~piotro@ has quit (Quit: warped)
[08:56:50] kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has quit (Ping timeout: 260 seconds)
[09:14:28] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Read error: Connection reset by peer)
[09:14:49] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[09:19:33] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Read error: Connection reset by peer)
[09:20:10] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[09:40:37] stuartm: dekarl: I'll consider adding that after the release of 0.25 – I don't believe mythtranscode uses the same code, I've ended up with the AD track after transcoding even when the frontend played the correct track, I can look into that though
[09:41:27] stuartm: atm I've only added support for AD tracks on DVD, I know the information is available for DVB but I've not yet figured out how to pass that through to where it's needed
[09:57:52] ** stuarta yawns **
[10:05:03] mike|2 (mike|2! has quit (Remote host closed the connection)
[10:05:57] mike|3 (mike|3! has joined #mythtv
[10:07:44] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[11:09:26] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 272 seconds)
[11:10:05] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[11:14:52] joki (joki! has quit (Ping timeout: 255 seconds)
[11:15:20] joki (joki! has joined #mythtv
[11:16:41] map7 (map7! has joined #mythtv
[11:33:10] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 260 seconds)
[11:41:05] wagnerrp: jya: mythmessage has not actually existed in any release version of mythtv
[11:41:13] wagnerrp: it was added during the 0.25 development cycle
[11:41:20] jya: ah ok cool...
[11:41:41] wagnerrp: hence, if its getting replaced... better to never have it show up in a release at all
[11:41:50] map7 (map7! has quit (Quit: Leaving)
[11:42:12] wagnerrp: capt'm has planned to remove it for weeks (or months), just never got around to it
[11:42:30] wagnerrp: i only forced it last night as someone was starting to document it on the wiki
[11:44:58] jya: stuartm, stuarta could either of you provide with two samples from BBC, h264 content, one interlaced, and one progressive ?
[11:46:03] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[11:48:31] jya: in ts format ...
[11:51:12] map7 (map7! has joined #mythtv
[11:56:15] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[12:04:58] davide (davide! has quit (Remote host closed the connection)
[12:05:22] davide (davide! has joined #mythtv
[12:11:12] danielk22: taylorr: I'd be very surprised to see data corruption issues in RingBuffer proper. In {DVD,BD}RingBuffer sure, but RingBuffer was looked at and looked at again and then looked at again before 0.24.
[12:27:55] stuartm: jya: any particular length?
[12:29:26] jya: stuartm: not really.. It's to debug the VDA decoder… I've found one already that play crappily in myth (awful with VDA, poor with ffmpeg)
[12:29:48] jya: 20s will do
[12:30:45] jya: BBC HD interlace would be nice… the VDA code doesn't seem to do any testing in regards to the video being interlaced, and VDA doesn't support interlaced video
[12:31:39] stuartm: I've a 2 minute sample I produced for Mark ready to download, it's interlaced and best watched without the audio because the dialog would make your skin crawl –
[12:32:00] stuarta: quality stuff eh?
[12:32:24] jya: no ts ?
[12:32:29] stuartm: recorded during the 'preview' period, this one being a preview of 'Land Girls'
[12:32:35] stuartm: jya: that is a ts
[12:32:52] jya: ah ok.. i went by the extension .mpg
[12:33:38] stuartm: Input #0, mpegts, from '/home/gbee/Videos/Test Samples/13940_20100117150843.mpg':
[12:34:06] stuartm: jya: that's just what mythtv names them, but everything we record is TS, we stopped recording PS a few years back
[12:34:39] stuartm: that particular sample should be unfiltered too, i.e. mheg and teletext (if any) included
[12:34:59] jya: cool thanks
[12:35:55] jya: i don't get it… xcode shows me two compilation error, yet there are no error there (trying to compile xbmc)
[12:38:34] stuartm: this is the mythtv development channel, you want #xbmc :p
[12:39:16] jya: hehe.. sure enough.. but I'm having a look at their VDA code to find out how they have resolved that particular issue
[12:40:05] wagnerrp: i thought they outright bypassed it
[12:40:15] wagnerrp: used some undocumented API underneath VDA
[12:40:26] wagnerrp: or was that just for the appletv/ipad?
[12:40:54] jya: nope… Actually, our code is based on their code… it's even in the header of our file
[12:41:18] jya: / Based upon CDVDVideoCodecVDA from the xbmc project, originally written by
[12:41:18] jya: / Scott Davilla ( ) and released under the GPLv2
[12:41:51] wagnerrp: must just be the appletv
[12:42:17] jya: probably...
[12:43:41] jya: stuartm: perfect sample… it makes mythavtest hang completely
[12:43:44] jya: excellent
[12:45:18] stuartm: heh
[12:45:22] danielk22: wagnerrp: We used to have some pre-VDA code that hooked into the undocumented MPEG-2 playback engine.. But that code was removed a some time ago.
[12:47:02] jya: danielk22: you're sure it was removed?
[12:47:38] jya: I was looking at the playback settings, there is still a Quartz decoder that hook up to the mpeg2 decoder.. at least that's what the help states, and that it can crash your mac
[12:48:07] The_Ball_ (The_Ball_!~The_Ball@ has joined #mythtv
[12:49:29] danielk22: jya: I think that's just stale help. It had libav hooks and I'm pretty sure it got functionally removed.
[12:49:57] The_Ball (The_Ball!~The_Ball@ has quit (Ping timeout: 248 seconds)
[12:57:47] stuartm: hmm, with extended testing it seems not all DVDs include audio language extension codes, I just got very lucky with those I did test
[12:58:24] stuartm: I really wonder why anyone bothers writing standards if no-one is going to follow them
[13:00:03] taylorr: danielk22: then why does the decoder complain and produce corrupted video when running to close to the end
[13:00:35] stuarta: danielk22: there is a lot of crap legacy video out osx code
[13:01:00] stuarta: funnily enough i mentioned it to markk and 2 weeks later it all disappeared from torc
[13:03:34] jya: stuarta: was probably on his todo list
[13:04:17] stuarta: not until i had a whole discussion with him about if the opengl side was sufficient these days
[13:09:01] cattelan_away is now known as cattelan
[13:15:40] jya: actually, I'm thinking of ripping off the whole carbon . quartz stuff for 0.26
[13:15:43] danielk22: taylorr: Does this affect all decoders? Maybe the decoder has a buffer problem.
[13:16:14] jya: opengl works well, and when buinlding against the Qt frameworks from Qt web site, you don't get carbon anyway
[13:17:50] jya: danielk22: currently, when I start LiveTV it stutters until I pause for a few seconds or rewind
[13:18:21] stuarta: jya: mark already did the work, we should just merge the relevant commits
[13:18:31] stuarta: no point redoing it
[13:19:01] jya: stuarta: I think if we do that, it's going to open a whole can of worms… he has changed so many things, renamed so many files that it's going be an absolute mess
[13:19:04] danielk22: jya: As I understand it that's the problem taylorr is trying to track down, we start playing before we have enough buffering and we play too close to the end of the file.
[13:19:31] jya: plus it wouldn't take me more than an hour, I've been all over the mac source code in the past few weeks, I'm starting to know it (too) well
[13:20:09] stuarta: hehe, well now you now it's already been done anyway
[13:20:29] jya: danielk22: yes, we talked about it… I'm guessing this reduced the channel change time… but with nasty consequences
[13:21:17] jya: stuarta: actually, most of the late compilation issue I've had has been about getting the carbon code to compile
[13:21:49] jya: not much love on the mac side… a lot of stuff seems like a hack of a hack..
[13:21:58] jya: especially the Qt makefile
[13:22:00] stuarta: aye, much hackery
[13:23:49] Jordack (Jordack! has joined #mythtv
[13:44:59] j-rod|afk is now known as j-rod
[13:51:38] danielk22: stuartm: I haven't been able to reproduce the frontend not responding to any keystrokes, but I have seen the backend not responding to anything but escape.
[13:51:58] danielk22: s/backend/frontend/
[13:53:21] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[13:53:53] stuartm: even after ad7ee3fdb ?
[13:56:38] stuartm: any thoughts on why 5.1 audio tracks are being identified as 7 channel?
[13:56:44] stuartm: jya: ^^
[13:56:49] taylorr: danielk22: I'm not sure if it's all decoders or just H.264/MPEG-TS – you can't blame the decoder if you stay 5 seconds behind and it plays back perfectly
[13:57:53] taylorr: danielk22: I've spent a good chunk of time looking into the frontend ringbuffer code but everything looks fine – so if the frontend is clean that means the backend is sending bad data
[13:58:04] taylorr: fyi, this is on a remote frontend
[13:58:45] danielk22: taylorr: The question is does the RingBuffer deliver bad data or are we playing too close to the end of the buffer because the decoder fails when you do. I don't think the RingBuffer delivers bad data.
[13:58:53] taylorr: danielk22: any ideas on how the backend could get bad data when reading the end of a file that is still being written – caching, driver bugs, race conditions
[13:59:51] taylorr: the decoder always reads 32-kbyte chunks and we always give it 32kbyte chunks
[13:59:53] danielk22: taylorr: I know some decoders make a 'best effort' to deliver a frame if it doesn't get all the data within a certain time frame. I don't think the libav mpeg-2 decoder does that, but I know for instance that the libmpeg2 mpeg-2 decoder does that.
[14:00:13] taylorr: that could be a possibility
[14:00:29] danielk22: taylorr: yep, and we make the decoder wait if we don't have a 32 kb chunk.
[14:00:35] taylorr: I figure it would log that info though
[14:00:59] stuartm: jya: nevermind, I see the bug, it only applies to DVDs and I'm just trying to understand it
[14:01:42] danielk22: taylorr: I believe with -v file --loglevel debug the ringbuffer computes checksums of each block so you can verify they are good.
[14:02:10] taylorr: danielk22: as long as the checksum was generated against good data :)
[14:02:12] danielk22: taylorr: That's one of the tools I used to verify everything was ok before 0.24.
[14:03:49] danielk22: taylorr: true, but I really think playing too close to the end of the file is the cause of both pauses which shouldn't be there and decoder errors.
[14:04:00] taylorr: danielk22: looking at libavformat/mpegts.c the read_packet() function wouldn't time out – so you think the h264 decoder would?
[14:04:35] taylorr: danielk22: right, it's related to being too close to the end but I'm looking at it from the point that we shouldn't have corruption
[14:04:37] danielk22: taylorr: That's the hypothesis.
[14:05:30] taylorr: not an easy one for me to prove :) – the h264 decoder is not for the weak
[14:05:56] danielk22: taylorr: I know it's a behaviour some decoders exhibit, I have no idea about libav h264.
[14:06:21] danielk22: taylorr: black box test it, add some sleeps in safe_read when you know you aren't close to the end of the file..
[14:06:28] taylorr: danielk22: just thought of this.... the audio never has any issues so your hypothesis is likely correct
[14:07:12] taylorr: I didn't even think about the audio being fine
[14:07:45] danielk22: taylorr: Audio is very low bitrate compared to video, so errors happen maybe 40x – 100x more often in the video.
[14:08:37] taylorr: ok, true
[14:12:04] taylorr: jya: any dangers from switching from your backports branch to the -fixes branch?
[14:12:15] taylorr: will the audio config get really messed up?
[14:18:09] jpabq_ (jpabq_!~jpabq@mythtv/developer/jpabq) has joined #mythtv
[14:24:48] taylorr: danielk22: so Mark had is audio completely stop playing when he played too close to the end – so that adds an interesting twist
[14:39:08] danielk22: taylorr: The design I remember was 1/ wait for a lot of buffering before starting video playback, 2/ if we get close to the end of the file slow down playback to 1x, 3/ if we get really close to the end of the file pause video playback for a second or so to allow the recorder to catch up.
[14:39:23] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[14:40:48] danielk22: taylorr: Of those #1 was always getting broken, and #3 was occasionally getting broken. I'm not terribly familiar with the current state of the code except I remember mark playing around a bit with #3.
[14:41:29] danielk22: #1 is often broken when someone wants to make livetv channel changes faster.
[14:46:43] xavierh: I noticed that in LiveTV the current channel group is not displayed, which might confuse the user on why some channels are not browsed, would it be considered as a bug ?
[15:09:10] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 246 seconds)
[15:16:50] stuartm: I don't suppose anyone is sitting on a copy of "Blu-ray ROM2 Part 3 : Audio Visual Basic Specifications version 2.4"?
[15:21:02] davide (davide! has quit (Quit: Konversation terminated!)
[15:29:47] Anssi_ (Anssi_!hannulaa@mandriva/developer/anssi) has joined #mythtv
[15:32:39] Anssi (Anssi!hannulaa@mandriva/developer/anssi) has quit (Ping timeout: 245 seconds)
[15:32:46] Anssi_ is now known as Anssi
[15:35:48] j-rod is now known as j-rod|afk
[15:37:23] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[15:38:09] wagnerrp: one of the VLC guys might have their hands on one
[15:39:19] stuartm: I've asked in #libbluray and we'll see what comes of it
[15:41:21] stuartm: it's ridiculous that you can find a torrent for any Blu-ray disc you want but that the techinical specs for BD-ROM remain a closely guarded secret
[15:41:45] stuartm: as though the licensing fee would really bring in that much cash
[15:42:08] stuarta: i really wish this whole IP licensing thing would disappear
[15:45:55] stuartm: it's outside the reach of what a single hobbyist developer is willing to spend, but still chicken feed to pretty much any small business – the logical flaw is that the number of businesses interested in those specs is tiny relatively speaking and they can't bring in that much cash, so why bother at all? If the specs were free then you'd see many more applications which would only increase demand for their main income stream i.e. their small cut
[15:45:56] stuartm: on disc sales
[15:47:47] wagnerrp: in theory, you shouldnt be able to do anything with it unless you have a license key anyway, right?
[15:47:51] j-rod|afk is now known as j-rod
[15:47:56] wagnerrp: (assuming AACS hadnt been cracked yet)
[15:48:58] stuartm: aye, they charge you at every turn, they'd invoice for the paper the invoice was printed on
[15:51:00] stuartm: anyway, no support for ignoring/labelling commentary tracks on BD unless I can see the spec, and even then there is no guarantee
[15:51:40] wagnerrp: does libbluray itself not expose any such information?
[15:51:44] stuartm: I'm switching my focus to DVB instead, the spec of which is freely available
[15:53:08] stuartm: wagnerrp: not that I can see, the only unexplained bit of the bd_stream_info struct is char_code and I suspect that's related to the character encoding of text subtitles
[15:53:52] stuartm: in fact I'm almost certain of that
[15:57:56] danielk22: Beirdo: 42dfafea .. the logStop() needs to be run before we close the databases and the MythContext dtor won't be run when we only instantiate a MythCoreContext..
[15:59:11] danielk22: We could just shut down the DB logger, but there is no API for that at the moment.
[16:03:21] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv
[16:11:00] stuartm: how difficult would it be to support AD tracks? It means decoding two audio streams and then combining them before output
[16:15:15] xavierh: Edit channel in livetv mode is always blank
[16:39:26] danielk22: stuartm: Should the new mythwelcome type functionality default to off? If it only allows the backend to shut itself off then I think it should default to off since that isn't the typical setup, but maybe it does more?
[16:41:48] stuartm: it only allows the backend to shut down, for now, but then only if the backend is configured to shut-down – the behaviour before this point was that the backend would never shut down even if configured to do so long as a frontend was left running (intentionally or unintentionally)
[16:43:11] stuartm: if the general opinion is that it should default to off then so be it, personally I'm of the opinion that you shouldn't have to modify multiple settings just to get one feature to work and I wouldn't have bothered adding the setting to disable it at all, but ...
[16:43:46] wagnerrp: can you make the default option be dependent on whether the backend is set up to shut down?
[16:43:51] stuartm: i.e. if you toggle the setting allowing the backend to shutdown when idle, then it should just shut down
[16:44:02] wagnerrp: if it isnt, nothing happens, if it is, you have to tell it you dont want the frontend to go idle automatically
[16:45:00] wagnerrp: theres some new screen that pops up when it goes idle, correct?
[16:45:54] stuartm: wagnerrp: yep, although if that's the issue it can be themed to be much less intrusive than it is now
[16:47:24] stuartm: fwiw, I wasn't intending to leave the default theme version as it is now, which is pretty crappy looking
[16:50:09] stuartm: I hope I'm not coming across as too defensive, I'd like to improve the behaviour if possible in line with user expectations rather that simply throwing settings at it because there are problems with it
[16:50:56] stuartm: I can for example easily prevent it going idle if the backend isn't configured to shut down
[16:51:14] stuartm: that's a pretty obvious improvement
[16:51:33] kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has joined #mythtv
[16:51:56] wagnerrp: right, if the backend isnt going to do anything anyway, dont bother with that potentially obtrusive popup
[16:52:39] zCougar (zCougar!~cougar@2001:67c:32c:600:8ca9:6434:8767:d93b) has quit (Ping timeout: 260 seconds)
[16:59:38] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 240 seconds)
[17:00:28] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[17:02:50] stuartm: obviously in the future there will be the option to automatically shut down (or put to sleep) an idle frontend but I haven't taken it that far yet
[17:04:15] zCougar (zCougar!~cougar@2001:67c:32c:600:c1aa:ad5d:3f67:8b35) has joined #mythtv
[17:05:51] kenni: jya, taylorr: FWIW, I created a ticket 5 days ago about ringbuffer messages in master. It might be related to what you're seeing:
[17:06:07] kenni: the symptoms I'm seeing are a bit different though
[17:18:09] danielk22 (danielk22!~danielk@ has quit (Ping timeout: 255 seconds)
[17:18:37] stuartm: ok, I'm stumped on how to rename 'File browse' mode, it seems metadata still works if you enable it so the earlier suggestion of 'Show/Hide metadata' isn't going to work
[17:18:46] danielk22 (danielk22!~danielk@ has joined #mythtv
[17:20:03] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 252 seconds)
[17:20:15] dblain (dblain! has joined #mythtv
[17:20:16] dblain (dblain! has quit (Changing host)
[17:20:16] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[17:23:59] stuartm: "Enable/Disable Library Browsing"? (database being the library in this case) ...
[17:25:17] Guest10049 (Guest10049! has joined #mythtv
[17:26:32] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 246 seconds)
[17:26:45] j-rod is now known as j-rod|afk
[17:26:49] stuartm: or just "Enable File Browse mode (slow)"
[17:27:09] Guest10049 is now known as dblain
[17:27:41] dblain (dblain! has quit (Changing host)
[17:27:41] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[17:29:37] j-rod|afk is now known as j-rod
[17:43:01] taylorr: kenni: I'm actually running a 0.24-fixes backend so it might be a different issue
[17:48:54] Beirdo: danielk22: hmmm, good point. How many places do we use a mythcorecontext without a mythcontext?
[17:52:01] warped (warped!~piotro@ has joined #mythtv
[17:53:23] danielk22: Beirdo: I don't know of any places. Captain_Murdoch created that class, I'm unsure of his future plans for it.
[17:53:53] Beirdo: K. have you seen a case where logStop() didn't get called, or something?
[17:54:27] wagnerrp: stuartm: better to make that 'Tree Structure', since there are different view modes that can build alternate trees
[17:54:37] Beirdo: it all seems to work OK here, but I'm sure there are situations I may have missed, and if so, be nice to fix it pronto :)
[17:54:55] wagnerrp: or maybe 'hierarchical'
[17:55:05] danielk22: Beirdo: not in the current code. There was an issue earlier, but then we had a lot of threads running wild and some were perhaps restarting logging between the MythContext dtor and the MythCoreContext dtor.
[17:55:17] Beirdo: ahh, K.
[17:55:24] Captain_Murdoch: mythcontext has UI components. that's why we had to have a mythcorecontext which doesn't have UI code.
[17:55:43] Beirdo: I guess we could just make logStop() smarter, so if it's already stopped, it just exits
[17:55:55] Beirdo: then it can be called in both places with no worries
[17:56:44] Captain_Murdoch: we needed to be able to access parts of the core context from libmythbase so I moved as much as I could at the time from libmyth/mythcontext to libmythbase/mythcorecontext
[17:56:46] Beirdo: if that's something we want to do, I can look at that later today
[17:57:39] danielk22: Captain_Murdoch: But one would never create just a MythCoreContext? It's just for visibility to libmythbase classes?
[17:58:17] Captain_Murdoch: I think we might do that for non-gui apps. can't recall if that's the way we do it or not. running a "git grep" now.
[17:59:05] danielk22: Captain_Murdoch: I just grepped and didn't see anything, but I thought it would make sense for something like mythutil.
[17:59:23] Beirdo: or perhaps mythtranscode, yeah
[17:59:32] Captain_Murdoch: we don't have text versions of the "can't find the DB" screens, so maybe we do use mythcontext everywhere still.
[17:59:43] Beirdo: heh
[17:59:47] Beirdo: wunderbar
[18:00:31] Beirdo: so we are probably OK for the moment then
[18:00:39] danielk22: Beirdo: so I think making logStop() safe to call twice and calling it in MythCoreContext is ideal.. but obviously not critical as we're ok for the moment.
[18:00:46] Beirdo: right
[18:01:01] Beirdo: I'll add it to my list of things to tweak in logging after releas e:)
[18:01:03] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Quit: ChatZilla [Firefox 10.0.2/20120215223356])
[18:01:19] Captain_Murdoch: my grep finds no direct use of mythcorecontext either, so we should always have mythcontext and mythcorecontext. a big reason for separation was not having circular loops in dependencies. libmythbase couldn't depend on libmyth to get to the corecontext, so all non-ui dependent parts were moved to libmythbase.
[18:01:56] Beirdo: gotcha
[18:02:00] Captain_Murdoch: no direct use of "new mythcorecontext" that is...
[18:03:16] kenni: taylorr: Ahh, okay
[18:03:46] kenni: I didn't have the issue before I upgraded from 0.24-fixes to master
[18:08:21] stichnot (stichnot!~chatzilla@ has joined #mythtv
[18:08:21] stichnot (stichnot!~chatzilla@ has quit (Changing host)
[18:08:21] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv
[18:08:42] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Read error: Connection reset by peer)
[18:09:00] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[18:12:44] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 246 seconds)
[18:13:30] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[18:18:10] Beirdo: danielk22: why rename threadRegister/Deregister? The new names really don't fit in.
[18:20:41] Beirdo: a visibility change doesn't require a naming change
[18:22:55] danielk22: Beirdo: mostly to make sure no outside plugins get caught out, but I also wanted to make it a bit more descriptive and avoid the mixedCase used for method naming.
[18:24:07] danielk22: logging{Register,Deregister} would meet the first two of those goals if ya prefer?
[18:24:35] Beirdo: we use mixedCase all over the place for functions. loggingRegisterThread maybe?
[18:24:55] Beirdo: makes sense though, to make sure none got missed :)
[18:25:32] danielk22: Yeah, we don't have a rule on the mixedCase for functions, it was brought up once or twice but it's not a policy.
[18:25:33] stuartm: wagnerrp: ugh, that's not good, since one of the reasons for picking 'directory structure' was to direct users to that option instead of using 'file browse'
[18:25:48] stuartm: what a mess
[18:25:49] warped (warped!~piotro@ has quit (Quit: warped)
[18:25:54] danielk22: loggingRegisterThread/loggingDeregisterThread works for me, I'll make the change.
[18:26:11] Beirdo: K  :) nice and descriptive too
[18:28:34] Beirdo: oooh, nice catch on the jobqueue
[18:29:21] Beirdo: the metadata one's usually fast, but there's no harm in preemptively closing the db and having to reopen
[18:30:18] danielk22: That is what inspired me to address the issue. I was restarting the backend often to test ceton recorder changes and had to wait forever for the unclean mysql shutdown.
[18:30:32] Beirdo: if you have VB_SYSTEM on, it could well reopen immediately though to log what's happening in myth_system itself
[18:30:48] Beirdo: yeah
[18:30:55] danielk22: I thought about adding it for all myth_system, but thought there are plenty of short program runs using myth_system.
[18:31:10] Beirdo: previewgen for example :)
[18:31:36] Beirdo: or even worse... the check for pulse audio :)
[18:33:05] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[18:33:15] danielk22: Grrr. It's unbelievable that we need to jump through that hoop for pulse audio. P|ss poor API design.
[18:33:32] Beirdo: heh, so agreed
[18:36:02] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 255 seconds)
[18:36:58] dblain (dblain! has joined #mythtv
[18:36:59] dblain (dblain! has quit (Changing host)
[18:36:59] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[18:51:19] wagnerrp: Beirdo: do you still use mythlcdserver?
[18:51:31] Beirdo: yeah
[18:51:43] wagnerrp: noticed any problems recently?
[18:51:49] Beirdo: not more than usual
[18:52:03] Beirdo: it's always been a finnicky bitch
[18:52:09] xavierh: Would be nice if mythlcdserver was a bit more generic
[18:52:18] wagnerrp: ah, user complaining its recently been crashing
[18:52:31] wagnerrp: or if not crashing, at least not continuing to run
[18:52:36] Beirdo: I haven't seen it crash other than when I upgraded it while it was running
[18:52:42] xavierh: I would love to see the current button list on a phone or something
[18:53:00] Beirdo: how is that an LCD?
[18:53:34] Beirdo: the lcdserver stuff is for 2-line or 4-line LCDs, not graphics displays
[18:53:48] wagnerrp: and specifically for LCDProc
[18:54:38] Beirdo: aye
[18:54:46] wagnerrp: getting a current buttonlist would be something to be added to the frontend Services API server
[18:54:56] Beirdo: that can be changed, but then we need something better to control the LCD modules
[18:55:06] Beirdo: and I haven't heard of a better tool yet
[19:04:04] jcarlos (jcarlos! has quit (Read error: Connection reset by peer)
[19:05:15] jcarlos (jcarlos! has joined #mythtv
[19:09:24] wagnerrp: xavierh: if youre actually considering writing something, take a look at Frontend/GetContextList and Frontend/GetActionList?Context=
[19:15:01] warped (warped!~piotro@ has joined #mythtv
[19:21:57] sphery: stuartm: could just get rid of *LoadMetaData settings (since they only encourage users to use file browse mode)--or, better, just get rid of *NoDB and *LoadMetaData  :)
[19:22:13] sphery: that said, I like the names you used
[19:25:01] stuartm: right, but not changes we can make for 0.25 and without some consensus among devs
[19:26:43] stuartm: I do actually think that the current wording will do for 0.25 despite the problem that wagnerrp pointed out, in the long run I think the whole thing will get changed so that we won't even need the menu option (no file browsing, automatic scans of connected media etc)
[19:29:20] stuartm: 'hierachy' or 'tree structure' aren't clear enough, they'd require some explanation which is little better than what we had with 'flat view'
[19:30:33] warped (warped!~piotro@ has quit (Quit: warped)
[19:33:50] sphery: stuartm: Yeah, current wording seems fine to me. Though do you want to also change theme/default{,-wide}/video-ui.xml to change "Enable metadata in file browse mode"?
[19:35:37] sphery: seems it's also in Terra and MythCenter-wide
[19:41:46] stuartm: ah, yeah, I forgot to include *.xml in my grep
[19:42:35] xavierh: wagnerrp: yes I have then this but I would need to know the current context. Also the current context are too vague, It doesn't really tell you what action are available
[19:43:50] xavierh: wagnerrp: I am considering to something to be able to do that.
[19:44:00] xavierh: s/to/to add
[19:47:32] cattelan (cattelan! has quit (Quit: Terminated with extreme prejudice - dircproxy 1.2.0)
[19:50:51] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[19:52:47] cattelan (cattelan! has joined #mythtv
[20:08:33] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 252 seconds)
[20:10:03] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[20:35:37] Jordack (Jordack! has quit (Quit: Three ducks and a cheese burger walk into a bar...)
[20:40:11] dblain (dblain!~dblain@mythtv/developer/dblain) has quit ()
[20:45:12] stuartm: it seems to be one step forward, two steps back whenever we fix livetv bugs :(
[20:49:20] skrock (skrock! has joined #mythtv
[20:58:38] map7 (map7! has quit (Ping timeout: 246 seconds)
[21:01:54] stichnot_ (stichnot_!~chatzilla@mythtv/developer/stichnot) has joined #mythtv
[21:03:32] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 246 seconds)
[21:03:35] stichnot_ is now known as stichnot
[21:04:06] jya: taylorr: no… it should work just fine… the audio configs are fully compatible
[21:07:36] jya: kenni: you're ticket is almost identical to what I'm seeing… at least in the error message
[21:08:51] jya: Beirdo danielk22 : to be fair, the reason we have to check if pulse audio is running as little to do with pulse itself and more with ALSA and that for digital devices only one application can open the device at a time. So to check for free alsa device we have to make sure nothing is using them. that could be pulse..
[21:11:14] pclark (pclark! has joined #mythtv
[21:11:46] jason (jason! has joined #mythtv
[21:11:53] jason is now known as jasonnz
[21:12:17] danielk22: jya: What I mean is pulse could have exposed some saner way to check if they are providing the "ALSA" device when emulating alsa.
[21:13:04] jya: danielk22: tbh, there's probably one :)
[21:13:42] jya: the pulse alsa emulation is actually from alsa…
[21:14:14] pclark (pclark! has quit (Remote host closed the connection)
[21:14:51] jya: the main issue with pulse is its asynchronous nature, and with alsa being fundamentally synchronous, it's a bad match. They should have just copied Apple's CoreAudio and not bother with a new API
[21:15:30] jya: I never see the point of reinventing new code just so its yours. if something works well, no shame in copying it...
[21:16:46] danielk22: jya: I honestly respect Lennart's role in finally dealing with audio issues in Linux even if he didn't do it perfectly. I think the disto's jumped the gun adopting it before it was ready and now we're stuck with an API that wasn't properly vetted.
[21:17:48] jya: I totally agree.
[21:18:32] jya: I'm just baffled on some of the bugs I'm seeing and that seems to be unique to using mythtv: like havng to set the pulse server sampling rate to 48kHz otherwise you often get white noise
[21:18:48] jya: issue has been there for years...
[21:19:21] jya: I've been over the code over and over again… It's been rewritten from scratch a few times… yet there's always the same issue
[21:22:19] jya: that is weird:
[21:22:48] jya: why would it try to load the debugging info from the .o file? shouldn't it be in the application itself?
[21:25:47] jya: crash is in the ringbuffer by the look of things: 0x00000001076e90aa in QReadWriteLock::lockForRead ()
[21:26:04] jya: rwlock seems to be null
[21:26:27] jya: danielk22: /\/\ that would be your specialty no ?
[21:26:35] jya: i mean the ringbuffer
[21:27:47] danielk22: I don't see the ringbuffer in that backtrace..
[21:28:06] danielk22: looks like the guidegrid to me
[21:28:51] danielk22: I need to run in a sec.. but was there a video playing in the guidegrid?
[21:29:18] jason (jason! has joined #mythtv
[21:29:18] jasonnz (jasonnz! has quit (Read error: Connection reset by peer)
[21:29:45] jason is now known as Guest61985
[21:31:22] skrock (skrock! has quit (Quit: WeeChat 0.3.7)
[21:35:10] danielk22 (danielk22!~danielk@ has quit (Ping timeout: 260 seconds)
[21:42:15] danielk22 (danielk22!~danielk@ has joined #mythtv
[21:45:12] j-rod is now known as j-rod|afk
[21:50:15] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[21:58:33] Guest61985 (Guest61985! has quit (Read error: Connection reset by peer)
[22:14:38] jya: danielk22: according to the report of the user on the mailing list, he only just attempt to start livetv, from there it crash… the reason I mentioned the rungbuffer, is that all insteand of lockforRead() seems to be in one of the ringbuffer
[22:19:38] jya: danielk22: sphery has asked a question that makes the bug similar to #9845, supposed to be fixed in master… but looks like it's not if he has the EPG showing up
[22:21:16] sphery: jya: #9845 is still open:
[22:21:45] sphery: #9846 is one of the ones closed by danielk's recent Live TV fixes
[22:22:10] jya: yeah, I'm referring to 9845, danielk22 states it has been fixed in master
[22:23:42] sphery: ah, ok, so it's waiting on testing by someone who was affected by it on 0.24-fixes and may be fixed
[22:24:41] sphery: skd5aner: haven't you recently upgraded to master/0.25beta? If so, can you re-test (the one where entering Live TV causes segfault if you enable "Show the program guide when starting Live TV")
[22:27:54] danielk22: Yeah, this one should be fixed in master. There was an issue where Guide didn't check if the video player was still there if it was started with a player.
[22:29:35] jason (jason! has joined #mythtv
[22:30:01] jason is now known as Guest73219
[22:33:20] jya_ (jya_!~jyavenard@ has joined #mythtv
[22:33:21] jya_ (jya_!~jyavenard@ has quit (Changing host)
[22:33:21] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[22:41:58] stuartm: is anyone else going to express an opinion on committing to 0.25? Per the email sent earlier today to the list?
[22:42:25] ** sphery needs to catch up on the list **
[22:44:20] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection)
[22:47:07] stichnot: jya: on your frontend, do you by chance override the default settings of GetSetting("ShortDateFormat", "ddd d") GetSetting("DateFormat", "ddd d MMMM") or GetSetting("TimeFormat", "h:mm AP") ?
[22:47:23] gigem: stuartm: i don't mind.
[22:48:34] stichnot: stuartm: it seemed too trivial for me to comment on. :) (i.e., no problem here)
[22:49:25] sphery: stichnot: I'm using DateFormat = ddd MMM d yyyy , ShortDateFormat = ddd M/d , and TimeFormat = h:mm AP (and, FWIW, I'm not on a system with a Qt that spins wildly in time-related functions--I'm still blaming (without proof) eglibc on that :)
[22:50:00] stuartm: thanks, I've just realised I didn't consider the time difference and it's only been 12 hours, it can wait another 12, it just needs to go in before the final string freeze if it goes in at all
[22:52:31] danielk22: stuartm: I didn't look at the actual patch, but it sounds pretty safe.
[22:55:20] stichnot: humph. the defaults in mythmiscutil.cpp are different than in mythfrontend/globalsettings.cpp.
[22:56:14] Guest73219 (Guest73219! has quit (Read error: Connection reset by peer)
[22:57:29] sphery: defaults in globalsettings are populated into the db, meaning the "local" defaults wouldn't apply (though I'm not excusing the inconsistency)
[22:58:48] stuartm: fwiw I'm looking at properly identifying AD tracks for DVB, there is code in mpegts.c to do this but it's commented out and wouldn't work anyway, it wouldn't even compile – perhaps not significant to those in ATSC land but properly labelling AD tracks and ignoring them when auto-selecting would be a nice improvement
[23:00:06] stuartm: if only we didn't use libmpeg2 for lossless transcoding it might also have lead to a fix for a long standing transcoder bug where it selects the wrong audio track to keep
[23:00:34] stuartm: s/lead/led/
[23:01:50] stichnot: sphery: yeah, so I revise my question for jya to ask if the settings are different from the default values in globalsettings.cpp. My settings are the defaults in globalsettings.cpp.
[23:03:31] dekarl: stuartm: cool, I've been looking at where the AD flags would be for DVB this morning, but as its in the playback part I didn't get anywhere :(
[23:04:18] sphery: stichnot: are you also on Ubuntu (an eglibc-using distro)? I wonder if any users on Red-Hat- (glibc-)based distros see the slowness.
[23:06:12] stuartm: they are in the component descriptor, quite a few component types actually described as AD relating mostly to the different audio codecs and whether they are 'mixed' AD tracks or standalone ones
[23:07:03] stuartm: dekarl: see ETSI EN 300 468, aka DVB Bluebook, the component descriptor table
[23:07:41] dekarl: stuartm: I can (more or less) find the bits blindly in the stream, but I couldnt wrap my mind about the avfdecoder and ringbuffer stuff :D
[23:08:37] stuartm: there is also apparently such a thing as "Audio for the hearing impaired", I'm not sure if that's where they crank the volume to 11 or just drop the volume of background noise/music so that the vocal track is easier to hear
[23:09:04] stuartm: I've never seen it in the wild
[23:09:19] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 245 seconds)
[23:10:18] stichnot: sphery: Ubuntu 10.04, but the slowness is fully explainable by the ION frontend. I probably tried it on the fast i7 backend (also 10.04) and found it plenty fast.
[23:11:43] sphery: isn't jya saying that it's 6.7s on his i7?
[23:11:44] dekarl: I've only seen audio for visual impaired and text overlay for the hearing impaired yet
[23:12:04] stuartm: dekarl: even I'm working in the dark, a little bit anyway, I've identified where I believe the additions need to be made but I don't know if it will actually work
[23:12:12] jya_: stuartm: I answered about your patch… Seems that one answer only seems to be the rule in the develop list :)
[23:12:57] stichnot: sphery: yes, which makes me believe it's either the Mac Qt or something weird about QDateTime formatting.
[23:13:10] jya_: stichnot: this is a brand new machine, the only settings I've changed are related to the playback profile and the use of OpenGL vs Qt for drawing
[23:13:38] stuartm: I've only been looking at this for two days after all, it never even occurred to me to work on it until I started looking at the commentary bug for DVDs
[23:14:56] stuartm: which it turns out won't fix every DVD, since I've discovered that some don't bother to flag the tracks as AD/Commentary :/, still it should reduce the hit rate for that bug
[23:15:05] dblain (dblain! has joined #mythtv
[23:15:05] dblain (dblain! has quit (Changing host)
[23:15:05] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[23:15:06] dekarl: stuartm: btw, while we're at doing sensible stuff automatically, is there still some amoung of content in 14:9 in the UK? (preferably with valid AFD signaling) Since I started using VDPAU I have to watch older midsomer murder episodes in passe partout :( (I can manually zoom in though)
[23:15:15] stichnot: jya_: is it a Release build of Qt? anything odd about your Qt installation?
[23:15:32] jya_: stichnot: It's 4.8.0 that I compiled myself.
[23:15:50] jya_: but having said that, it does the same with vanilla 4.7.4 and 4.6.4 install
[23:15:55] stuartm: dekarl: automatic zoom doesn't work?
[23:16:14] dekarl: stuartm: hmm, I tend to fixup the rips with tsdoctor / tsmuxer etc. and manualy fix missing signaling
[23:16:24] sphery: dekarl: I hope you mean "manually select half zoom" and not "zoom in using manual zoom"
[23:16:25] dekarl: stuartm: aye, no automatic zoom with VDPAU for me
[23:16:43] dekarl: sphery: no, I mean manualy selecting "fill horizontal"
[23:17:14] dekarl: the display is 16:10 and the content is 14:9 picture packed in anamorph 16:9 frames
[23:17:20] stuartm: I'm really not sure on the quantity of 14:9 stuff, it was all signalled as ... damn, what's the term, my mind has gone blank ... anyway, it was all supposed to be stretched to 16:9
[23:17:24] sphery: I think auto zoom is non-functional with vdpau since we can't inspect the frames without inefficiently moving them to RAM
[23:17:27] stichnot: jya_: so many billions of cycles to compute 2000 items... at this point, all I can say is that the answer is sure to be obvious if you can throw a profiler at it
[23:17:27] stuartm: anamorphic ... righ
[23:17:29] stuartm: ty
[23:17:35] sphery: dekarl: yeah, that's what half zoom is meant for
[23:17:55] sphery: MENU|Video|Zoom|Half Zoom, IIRC
[23:18:07] jya_: stichnot: I've been wanting for a while to set a proper xcode projects to compile mythtv.. this would make debugging and profiling so much easier...
[23:18:08] stuartm: sphery: we should be stretching it if it's signalled as 16:9
[23:18:13] sphery: (and not selecting the aspect stuff)
[23:18:34] stuartm: although there would still be thin black bars on a 16:10 display ...
[23:19:06] dekarl: stuartm: streching the anamorphic part works like a charm. its the "we only use 14:9 of the available 16:9 space and pillarbox the content" which gets an additional letterbox by the player thats not nice
[23:21:09] stuartm: heh, I'm not getting my head around that right now, but I can picture the end result, I actually have the same issue with an old DVD where it's signalled incorrectly
[23:22:17] stichnot: jya_: one other thought is to temporarily hack MythDateTimeToString() to return a fixed string. That ought to confirm whether it's a QDateTime problem.
[23:22:18] stuartm: dekarl: seems BBC One HD and BBC HD are still 1440x1080, anamorphic
[23:22:37] jya_: stichnot: I can definitely try that
[23:23:01] dekarl: the transmitted picture has one picture aspect but the signalling contains the information how much of that is actually used. The purpose is that a tv can do something sensible automatically (without looking for black bars)
[23:23:17] jya_: stuartm: I downloaded the BBD HD test video, it's like 1960x1088
[23:23:29] stuartm: no content is originally 14:9 though and broadcast as such, some 4:3 but even a lot of older stuff was 16:9
[23:23:29] dekarl: stuartm: anamorphic or square pixels is another topic :)
[23:23:58] stuartm: dekarl: right ... you can tell that I'm not at my best tonight ;)
[23:24:08] dekarl: stuartm: I have recordings with a picture 14:9 (midsomer murder was produce like that in the transition period between 4:3 and 16:9)
[23:24:19] dekarl: me, too :D
[23:25:24] stuartm: I can't remember if I've seen 14:9 then, certainly not recently
[23:26:26] dekarl: might be because we germans have like 4–5 years of backlog to catch up. (and we started out with a best of collection of the first years which has been backfilled in parallel to the newer episodes, which are still old for you)
[23:29:18] jason (jason! has joined #mythtv
[23:29:43] jason is now known as Guest77967
[23:31:03] stuartm: yeah, Midsomer Murder's has been going for years and years, it's a running joke now that this tiny fictional village has a higher murder rate than the entire country and there can't be anyone left alive ;)
[23:31:15] jpabq_ (jpabq_!~jpabq@mythtv/developer/jpabq) has quit (Quit: jpabq_)
[23:32:02] stuartm: it's almost 15 years old now, and it's been almost that long since I last watched an episode
[23:34:04] stuartm: actually, looking a wikipedia it seems 'Midsomer' is supposed to be a small county rather than a single village
[23:36:06] stuartm: suprisingly popular abroad too, was it Beirdo who has a Midsomer Murder collection? One of the US devs anyway
[23:36:22] dekarl: over here we are already at the 2007 episodes, with the backfill having reached the 2004 season. makes for quite some jumps
[23:36:45] Beirdo: not me, but that sounds liek something I'd record
[23:40:12] stuartm: hmm, think it was jpabq- then, when he posted his steppes screenshots iirc
[23:44:20] danielk22: stuartm: The early Dr. Who reboot episodes where in that intermediate aspect ratio... If you can grab some of those off the BBC I'd like to look at it. There are AFD flags in the stream we're not looking at. If they broadcast I could add support fairly easily and it wouldn't involve looking at the actual video content.
[23:45:26] danielk22: I know some networks in the US are sticklers about getting the AFD stuff right. They don't want re-aired 4:3 material to get letterboxed on all sides on 4:3 TV sets.
[23:46:03] stuartm: danielk22: I'll keep a look out, I'm one of those rare few who doesn't watch Dr Who so I've no idea if they are showing repeats of the earlier episodes
[23:46:10] stuartm: stuarta: ^^
[23:47:27] stuartm: nothing in the schedules for the next couple of weeks, someone might have an old recording around ... justinh perhaps?
[23:54:38] Guest77967 (Guest77967! has quit (Read error: Connection reset by peer)

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