Back to Home / #mythtv / 2007 / 02 / Prev Day | Next Day
#mythtv IRC Logs for 2007-02-22

---Logopened Thu Feb 22 00:00:13 2007
---Daychanged Thu Feb 22 2007
00:00<xris>I'm on the verge of having this ready to deploy.
00:00<Chutt>i have no idea what you're talking about :p
00:01<Chutt>well, yes.
00:01<xris>intelligent channel icon lookup
00:01<Chutt>but vs storing the url or the 'current cache value'?
00:01<xris>just wondering if it's worth adding an additional field to store the URL to the icon, as well as the current field that stores the local path.
00:02<Chutt>easy enough to requery it if the client wants to update the url
00:02<xris>yeah. URL actually seems a bit redundant...
00:03<xris>the path var will eventually need to be changed, but that will require the backend managing the files (pointing to a path in the local FS doesn't work so well with multiple frontends, different users, etc).
00:03<xris>anyway, I just wanted to see if you had any thoughts about wanting to keep track of the URL, or the icon id, etc.
00:04<Chutt>those should be automatically getting sent from the master backend to any frontends that request it
00:04<Chutt>ie, that's part of the network code already. =)
00:04<xris>maybe I missed that with mythweb and the perl bindings, then.
00:04<xris>I think I always just grabbed them locally
00:05<xris>just a normal file transfer, with the full path as the file?
00:05<Chutt>naw, it's a special one, iirc
00:05<Chutt>i don't remember the details right now.
00:06<xris>np. I'll dig around.
00:06<xris>that solves some of my concern, though
00:09|-|cattelan [n=cattelan@] has quit ["This computer has gone to sleep"]
00:22|-|LLyric [] has quit ["Chatzilla 0.9.77 [Firefox]"]
00:55<xris>Snow-Man: ping
00:56|-|turbo [] has joined #mythtv
00:56|-|briand [] has quit [Read error: 110 (Connection timed out)]
01:01|-|madfactor [] has joined #mythtv
01:12|-|okolsi [] has joined #mythtv
01:14<okolsi>Morning.. change 12847 doesn't compile for me (Athlon 32)
01:21|-|o_cee [] has joined #mythtv
01:29<eskil>"Warning: size of symbol `write_ts_header' changed from 233 in ts.o to 550 in pes.o" ?
01:30<okolsi>yep, numbers differ but otherwise..
01:36<janneg>make function in both files static fixes compilation
01:37<janneg>I need to wake up to decide if both functions are identical
01:46<eskil>Guess the linker doesn't think so.
01:50|-|gnome42 [] has quit [Remote closed the connection]
01:56|-|Dibblah [] has left #mythtv ["Kopete 0.11.1 :"]
01:57|-|Dibblah [] has joined #mythtv
02:17|-|kvandivo [] has quit [Remote closed the connection]
02:37<janneg>they are not identical. I could see that, I wasn't sure if they might give the same results for the same input
02:43|-|eskil [] has quit [Remote closed the connection]
02:44|-|kvandivo [] has joined #mythtv
03:04<okolsi>now HEAD compiles again
03:13|-|xris [] has quit ["Leaving."]
03:13|-|kormoc [] has quit []
03:17|-|mark_ [] has joined #mythtv
03:17|-|aevil [] has joined #mythtv
03:26|-|LLyric [] has joined #mythtv
03:30<stuarta>Cardoe: you around?
03:32<stuarta>i have something i'd like to get you to test re: your HD content
04:45<dude900>anyone got the "NVP AddAudioData() Audio buffer Overflow: audio data lost!" error from playing a pre-recorded show?
04:46<dude900>starts playing the file at super speed then quits
04:47|-|JohnnyST [] has joined #mythtv
04:49<JohnnyST>Hello, have there been some rewrite of seeking in video streams during the last month? I uppgraded to svn yesterday and I get this error when I try to playback HD content:
04:49<JohnnyST>NVP::AddAudioData():p1: Audio buffer overflow, audio data lost!
04:49<JohnnyST>I get then I try to forward in an HD video.
04:51<JohnnyST>720i seems to work but 1080i dose not.
05:12<dude900>my question first :P
05:55|-|turbo changed nick to briand
06:31|-|MoRpHeUz [n=morphbr@] has joined #mythtv
06:32|-|dude900 [] has quit [Client Quit]
06:35|-|lsobral [n=sobral@] has joined #mythtv
06:50|-|MoRpHeUz_ [n=morphbr@] has joined #mythtv
06:51|-|MoRpHeUz [n=morphbr@] has quit [Nick collision from services.]
06:51|-|MoRpHeUz_ changed nick to MoRpHeUz
07:26|-|cattelan [n=cattelan@] has joined #mythtv
07:35|-|cattelan [n=cattelan@] has quit ["This computer has gone to sleep"]
07:47<Cardoe>stuarta: back
07:51<Cardoe>stuarta: actually... lemme get back to you from work
07:51|-|Cardoe [n=Cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
08:07|-|o_cee [] has quit [Read error: 110 (Connection timed out)]
08:15|-|splat1 changed nick to splAt1
08:15|-|splAt1 changed nick to splat1
08:26|-|Weezey [n=weezey@] has joined #mythtv
08:31<Weezey>I've searched the trac and asked around, and no one's got an answer. In the current SVN does lircrc need to be modified so that it controls the mythfrontend process? After the upgrade it doesn't work anymore, but xine does so irexec seems to be doing it's job. I've tried changing prog = mythtv to prog = mythfrontend but that didn't work either. Also, I'm having random slave backend crashes. There's no log spit out though.
08:38|-|Cardoe [n=cardoe@gentoo/developer/Cardoe] has joined #mythtv
08:39<stuarta>ah there you are....
09:07|-|splat1 changed nick to splAt1
09:07|-|jgarvey [] has joined #mythtv
09:08|-|splAt1 changed nick to splat1
09:11<gbee>Weezey: nothing has changed in SVN with lirc support afaik
09:11<gbee>it works for me but then I'm using the builtin support
09:13<gbee>it's almost certainly a #mythtv-users question
09:16<Snow-Man>Or a "read the fucking commit logs" question
09:21[~]stuarta suspects Snow-Man's been taking lessons from Juski
09:49|-|okolsi [] has left #mythtv []
09:56|-|gnome42 [] has joined #mythtv
10:16<Cardoe>stuarta: ping
10:16<Cardoe>stuarta: hit me with something.. ;)
10:18<sphery>Looks like #3109 ( ) is the same non-Myth issue as --only difference being GNOME versus KDE.
10:19<sphery>And shows how to work around it in GNOME.
10:31<stuarta>Cardoe: pong!
10:32<stuarta>right, after reading what you guys were talking about last night
10:32<stuarta>i suspect you may be seeing what i found before.
10:32<stuarta>if your cpu is borderline for the content you are trying to decode
10:32<stuarta>the a/v resync code overcompensates.
10:33<stuarta>to test it, run the frontend with --verbose playback
10:33<Cardoe>It probably is.. 3000+ AMD64
10:33<Cardoe>I do see the comments about a/v resync in there
10:33<stuarta>3200+ here and i see it...
10:33<stuarta>what i found is.
10:33<stuarta>1. it'll be out of sync and drop audio packets to catch up
10:34<Cardoe>the sound is choppy and the video can be do
10:34<Cardoe>er to
10:34<stuarta>2. it goes to far and then has to drop video packets to catch up
10:34<stuarta>repeat 1 & 2.
10:34<Cardoe>But I go back to the main menu and play the file... so I'm like 2 seconds behind... and it works fine
10:34<Cardoe>So the resync does make sense
10:35<stuarta>i'm more after confirmation that your logs show the same sort of thing as mine
10:35<stuarta>dropping audio .... dropping video....
10:36<stuarta>starts dropping video pretty quickly after dropping audio
10:36<stuarta>which would exhibit as audio drops and video stutter.
10:51<janneg>stuarta: can you check if configure had enabled mmx?
10:53<janneg>it should enable it now, but before my last commit it was disabled for 64 bit cpus running a 32bit kernel
10:53|-|JoeyBorn [] has joined #mythtv
10:54<stuarta>yup i have MMX with the latest update
10:56<janneg>that should help h264 playback a little
11:00|-|czth_ [i=dbrobins@nat/microsoft/x-e9529326de4ada1d] has joined #mythtv
11:14|-|czth [i=dbrobins@nat/microsoft/x-ef7ca1f1ce794afd] has quit [Read error: 110 (Connection timed out)]
11:21|-|goao [] has joined #mythtv
11:53|-|mark_ [] has quit [Read error: 110 (Connection timed out)]
11:56|-|xris [] has joined #mythtv
12:06|-|dverzolla [] has joined #mythtv
12:08|-|mark_ [] has joined #mythtv
12:23|-|cattelan [] has joined #mythtv
12:33<Snow-Man>xris: Done, btw.
12:33|-|robthebob [] has joined #mythtv
12:34<xris>Snow-Man: thank you muchly. :)
12:36<xris>odd, it still seems to be having issues. but not package-related now.
12:37<xris>ah, files not fully up to date
12:37<xris>yay, it works
12:43<xris>it's so weird with mod_perl, though. first time the script is loaded after making a change, it fails (my system and every other time, it works fine
12:52<xris>I think I need to move the sqlite db, though. it apparently needs read privs on the directory to handle writes, since it writes to a copy/tmp file
12:52<xris>better to give those privs only to a db directory rather than where it's sitting now
12:52<Snow-Man>Sounds good to me.
12:57|-|xris [] has left #mythtv []
12:57|-|xris [] has joined #mythtv
13:05|-|kormoc [] has joined #mythtv
13:21<gbee>about to commit Paul's patch which moves destruction of the AudioOuput from PlaybackBoxMusic::stop to the destructor, any reason why this shouldn't be done?
13:22|-|legistrate [] has joined #MythTV
13:22|-|brianw_ [] has joined #mythtv
13:25<gbee>practical benefits include mute not being turned off when it starts playing a new track :)
13:29|-|cattelan [] has quit ["This computer has gone to sleep"]
13:30<gbee>hmm, just found the downside is that it plays out what's left in the buffer when you hit stop, so it can actually take a couple of seconds to stop
13:30[~]gbee looks for a solution
13:30<xris>I think amarok has that same issue sometimes.
13:31|-|dverzolla [] has quit ["User pushed the X - because it's Xtra, baby"]
13:38|-|JoeyJoeJo [] has quit [Read error: 110 (Connection timed out)]
13:43<Chutt>stop has to be immediate :p
13:44<xris>yes, I also agree that's generally a good requirement.
13:44<xris>pause, too.
13:46|-|goao [] has quit ["9) nanoseconds"]
13:49|-|eskil_ [] has joined #mythtv
13:50<gbee>ok solved that by making pause() the first thing we do in stop() ;)
13:51<gbee>pause always has been immediate or as close to immediate as audiooutput currently allows
13:51|-|eskil_ changed nick to eskil
13:51<xris>sounds like an ugly kludge.. or a really smart fix. I can't figure out which, though
13:52<gbee>ugly actually, doesn't empty the remaining buffer so when you hit play you get a fragment of the previous track
13:53<gbee>I'll dig through audiooutput for a suitable method or write something to do the job
13:59<gbee>KillAudio sounds promising, although it looks like it closes/stops a lot of stuff which we probably want to keep - it's the only way, other than pause, to actually interupt the loop
14:01<xris>no way to just clear the buffer?
14:01<gbee>not currently
14:02<gbee>I mean it shouldn't be hard to do, but I'd probably rather not touch the audiooutput code at all
14:02<gbee>KillAudio seems to be out of the question :(
14:07|-|Netsplit <-> quits: mark_, jgarvey, Nem^, olds
14:10<gbee>ah-hah, wonder if Reset() is what I'm looking for
14:11|-|cattelan [] has joined #mythtv
14:12<gbee>yep that's it :D
14:13|-|Netsplit over, joins: mark_, jgarvey, Nem^, olds
14:13|-|CDev [] has quit ["Trillian ("]
14:19|-|Tbonius [] has joined #mythtv
14:19[~]xris points at the topic before Tbonius asks an invalid question
14:19<Tbonius>i see it .. thank you
14:22<Tbonius>thank you
14:29<gbee>joy, found another bug
14:30<gbee>whilst playing music, press 'pause', then 'stop', after which 'play' doesn't work
14:30|-|legistrate [] has left #mythtv []
14:32|-|CDev [] has joined #mythtv
14:33|-||Torg| [n=Torg@] has joined #mythtv
14:34<eskil>gbee, I actually remember running into that and writing a fix.
14:36<gbee>well it would save me 20 minutes debugging if you've got a patch or can remember what the problem was :)
14:36<sphery>gbee: would your AudioOutputBase::Reset() thing also benefit playback of recordings? It seems when I use ALSA output, I get a little (half second or so) audio output after muting or pausing (and can see video paused, but audio contining).
14:36<sphery>It's easiest to see when music is playing
14:36<sphery>I guess I could test it someday...
14:37<eskil>gbee, I'm looking, I think it got lost when my laptops drive crashed, and I guess I never filed a bug with it...
14:37[~]eskil sucks...
14:37<gbee>sphery: no, unfortunately not
14:37<sphery>Oh. OK.
14:38<sphery>I've been using OSS output--since there's no reason not to-- so I never looked into it.
14:38<gbee>sphery: the problem you see there is down to the speed at which we iterate through the audio buffer and send fragments to the card
14:39<gbee>well, thinking about it, it's worth a try but you might get the opposite, a little period without audio after unpausing/muting
14:39<sphery>OK. Sounds like you have a handle on it (I've never looked at the code), so I'll trust your conclusions.
14:40<sphery>I won't have a chance to play with it for a few weeks, but I'll put it on my list.
14:40<gbee>nah I don't really, I've no idea of the interaction between NVP and AOB
14:40<gbee>reseting may working but it could also screw up things like audiosync
14:42<eskil>gbee, ah, found it, but it's old and probably applies like shit ;
14:42<sphery>So the hard part--assuming it proves beneficial to me--is testing whether there are negative side-effects elsewhere. May be better to leave as is. After all, for those of us who are bothered by it, there's still OSS output.
14:43<gbee>eskil: thanks, doesn't matter if it won't apply at least I can see what needs doing
14:43|-||Torg| [n=Torg@] has left #mythtv []
14:43<sphery>thanks for the info gbee
14:44<eskil>gbee, at least I hope it was for that case, it sounds very familiar and looks related.
14:45<gbee>sphery: if you want some good advice on solving the problem best ask someone else, I've spent a little more than 20 minutes looking at AOB and only unstand enough to fix the problem I had with mythmusic
---Logclosed Thu Feb 22 14:54:41 2007
---Logopened Thu Feb 22 14:54:48 2007
14:54|-|mikegrb [] has joined #mythtv
14:54|-|Ekipa kanalu #mythtv: Wszystkich: 80 |-| +op [0] |-| +voice [0] |-| normalnych [80]
14:54|-|Kanal #mythtv zsynchronizowany w 3 sekundy
14:57<sphery>gbee: No aggressive sound card buffering (AggressiveSoundcardBuffer), but I do have extra audio buffering (DecodeExtraAudio) enabled
15:06<Chutt>gbee, just don't break gapless playback.
15:07<gbee>gapless playback meaning normal progression to the next track in the playlist?
15:07<Chutt>entirely seamlessly, yes.
15:09<gbee>has it ever been entirely seamless? either way it shouldn't have any effect, instead of destroying and creating a new audiooutput we're just reset()'ing the current one which should actually be faster and even more seamless
15:09<Chutt>basically, not doing anything to the output class unless it *needs* to on a track change.
15:09<Chutt>and reusing it between tracks.
15:13<MoRpHeUz>when using backend's playbacksock to answer requests, is there any default format like everything uppercase ? (for example, "ok" and "OK") ?
15:13<MoRpHeUz>or can be both ?
15:21|-|AlienX [n=theanswr@unaffiliated/alienx] has quit [Remote closed the connection]
15:21[~]sphery wonders if his is the only system on which F9 no longer mutes
15:22[~]stuarta has never pressed F9
15:23<sphery>Hmmm... Seems the C++ escape for the 2nd default mute key "\" is getting written to the DB... We end up with '|,\\,F9', and it chokes on '\\', so it never gets to 'F9'.
15:24<sphery>Setting it with MythControls, it started showing only '|', so I added two new ones '\" '|,\,F9' and now F9 works
15:24|-|lcase [] has joined #mythtv
15:25<sphery>oops added, '|' and 'F9' and mythcontrols changed it to '|,\,F9' and now F9 works
15:26<sphery>Escaping is also in there for Music and Stream (MythStream) contexts...
15:28|-|packetscan changed nick to PacketScan
15:28<sphery>Oh, and \ didn't work, either.
15:33|-|orzel [] has joined #mythtv
15:33|-|orzel [] has left #mythtv ["Konversation terminated!"]
15:35[~]sphery thinks someone changed from a .arg() call in MythMainWindow::RegisterKey() to a .bindValue(), so the '\\' is getting escaped by the MySQL driver...
15:51|-|lcase [] has quit []
15:57|-|MoRpHeUz [n=morphbr@] has quit ["Leaving..."]
16:04<gbee>eskil: the pause, stop, play bug turned out to be a problem in my current tree with calling reset - reset doesn't reset the paused state in AOB
16:06<gbee>already fixed, just calling pause() after reset() if AOB::getPause is true sorts things out
16:07<sphery>I found the change at , old lines 476-506, new lines 505-539.
16:07<sphery>Went from .arg() to .bindValue()
16:08<sphery>But that one (r6456) has inskey.replace('\\', "\\\\"); but removes inskey.replace('\"', "\\\"");
16:08<sphery>but then adds it back (new line 694), making it seem the escaping is in fact necessary--although AIUI, escaping of strings should be done by bindValue()...
16:10<sphery>I think lines 830 and 831 should be removed ( ).
16:10<sphery>I'll have to wait until I can create a new DB to test it. Seems it's only a problem with the escaped literal values in the REG_KEY macros.
16:20<briand>sphery: did u see that i added your theme changes, as requested (along with the others)?
16:23|-|lsobral [n=sobral@] has quit ["Leaving"]
16:23|-|xri1 [] has joined #mythtv
16:24|-|xri1 [] has quit [Client Quit]
16:24|-|xris [] has quit ["Leaving."]
16:24|-|xris [] has joined #mythtv
16:29|-|mark_ [] has quit ["Konversation terminated!"]
16:36|-|aevil [] has quit [Read error: 110 (Connection timed out)]
16:41|-|aevil [] has joined #mythtv
16:43|-|luna6 [] has quit [Connection reset by peer]
16:44|-|luna6 [] has joined #mythtv
16:47|-|eskil [] has quit [Read error: 110 (Connection timed out)]
16:49<sphery>briand: Didn't notice (wasn't following up--I trusted you ;), but thanks.
16:55|-|jgarvey [] has quit ["Leaving"]
17:00|-|Cardoe [n=cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
17:25|-|cattelan [] has quit [Read error: 110 (Connection timed out)]
17:29<gbee>heh, Anduin's 'slightly modified' version of Juski's patch looks nothing like it
17:30<Anduin>It is identical code :)
17:30<gbee>the changeset has one new line
17:34<janneg>100% more lines
17:34[~]gbee goes trout fishing
17:45<gbee>1 minute 49 seconds saved for 7800 tracks isn't fantastic, think I might be able to scrape another minute but it's still too slow
17:46<gbee>maybe with some inspiration and luck I can do better
18:07|-|aevil [] has quit [Remote closed the connection]
18:15|-|Machtin [] has joined #mythtv
18:16|-|Machtin [] has left #mythtv []
18:30|-|robthebob [] has quit [Read error: 104 (Connection reset by peer)]
18:59|-|luna6 [] has quit [Read error: 104 (Connection reset by peer)]
18:59|-|luna6 [] has joined #mythtv
19:04|-|Erroneous_ [n=Erroneou@] has joined #mythtv
19:08<Erroneous_>just out of curiousity, has anyone thought about the default method for matching series by seriesid and not just by title?
19:11<Erroneous_>that way, for instance, you don't record the movie "House" if you record the series "House". It is easy enough to do with a power search, but it seems like that would be something that should be default behavior
19:11<kormoc>Yeah, I've run into that one a few times actually
19:12<kormoc>in my opinion, a flag to set the behaviour would be nice
19:12<Erroneous_>sorry if it is an old and beaten dead horse
19:13<Erroneous_>I can see if your listings source doesn't properly list that information having that not enabled
19:18<Erroneous_>also, how about an option to match episodes by programid in addition to subtitle and subtitle and description
19:19<Erroneous_>because I do have the problem of say Family Guy episodes getting recorded on TBS when I have already watched it from Cartoon Network because the description is different
19:21<Erroneous_>I'm fluent in C and mySQL, so I'd be willing to add these options myself, but I've never contributed to a project before
19:22<kormoc>well, the offical way is to get the patch ready and post it on and see what happens
19:24<Chutt>the seriesid stuff isn't entirely accurate, even with datadirect.
19:27<Erroneous_>isn't accurate as in more that one series has the same ID or more than one series has multiple IDs?
19:34|-|[SA]Pooky [] has joined #mythtv
19:35|-|[SA]Pooky [] has left #mythtv []
19:47|-|cattelan [n=cattelan@] has joined #mythtv
19:49|-|Erroneous_ [n=Erroneou@] has left #mythtv ["Leaving"]
20:45|-|xris [] has quit [""]
20:46|-|kormoc [] has quit ["Done with work, yay! ( )"]
20:51<CDev>What is the best way to send a message from a worker thread to a GUI thread (in a modal dialog)?
20:52<CDev>I was going to use signals and slots, but it doesn't look like it is the best choice.
20:59<Chutt>CDev, events
20:59<Chutt>customEvent(), MythEvent
21:00<CDev>Thanks. I'll take a look.
21:08|-|Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv
21:19|-|kormoc [] has joined #mythtv
21:20|-|JoeyBorn [] has quit [Read error: 110 (Connection timed out)]
21:30|-|xris [] has joined #mythtv
22:18|-|cmorgan [] has joined #mythtv
22:18|-|brianw_ [] has quit [Client Quit]
22:39|-|Nem^1 [] has joined #mythtv
22:39|-|cmorgan [] has quit [Remote closed the connection]
22:46|-|purserj [] has quit [Read error: 131 (Connection reset by peer)]
22:47|-|purserj [] has joined #mythtv
22:56|-|Nem^ [] has quit [Read error: 110 (Connection timed out)]
22:56|-|Nem^1 changed nick to Nem^
---Logclosed Fri Feb 23 00:00:45 2007