Back to Home / #mythtv / 2008 / 02 / Prev Day | Next Day
#mythtv IRC Logs for 2008-02-18

---Logopened Mon Feb 18 00:00:35 2008
00:22<Captain_Murdoch>rooaus: I haven't given it a whole lot of thought (just got home from a day with about 12 hours combined driving and flying), but theoretically you could have a virtual storage group that returned a Myth playable URL when you searched using StorageGroup::FindRecordingFile().
00:23-!-kormoc [n=kormoc@unaffiliated/kormoc] has quit []
00:37-!-Hannibal- [] has quit [Read error: 104 (Connection reset by peer)]
00:48-!-janneg [] has quit [Remote closed the connection]
00:48-!-janneg [] has joined #mythtv
01:28<xris>GreyFoxx: sounds like my bug.
01:34<rooaus>Captain_Murdoch: Cool, will think some more about it, enabling storage groups in mythmusic is the first step I guess. Take it easy :)
02:02-!-patdk-lap [] has joined #mythtv
02:10-!-gnome42 [] has quit [Remote closed the connection]
02:21-!-anykey_ [] has joined #mythtv
02:31-!-Andui1 [] has joined #mythtv
02:32-!-Anduin [] has quit [Nick collision from services.]
02:32-!-Andui1 is now known as Anduin
02:40<hads>Anduin: Oops, wrong componenet on that ticket sorry.
02:41<hads>I blame Monday.
03:01-!-xris [] has quit []
04:00-!-SpAc [] has joined #mythtv
04:01-!-SpAc [] has quit [Client Quit]
04:02-!-mick_laptop [n=mweiss@clamwin/admin/mickhome] has joined #mythtv
04:03<mick_laptop>hi everyone
04:04<mick_laptop>is there a plan on making thumbnails for videos that aren't searchable online?
04:05<mick_laptop>i've seen thsi:
04:05<mick_laptop>which didn't even work for me -- but that is besides the point
04:05<mick_laptop>the poster on the bottom made a point, it would be cool if images would be handled like that
04:05<mick_laptop>i', just not sure how one would go about doing that in the best way :)
04:17-!-sigger_ [] has joined #mythtv
04:33-!-gardz [] has joined #mythtv
04:43<gbee>mick_laptop: already written the code
04:44<gbee>won't be committed until after 0.21 is released and then it still need Anduin to help integrate it into mythvideo
04:50-!-grokky [] has joined #mythtv
04:54-!-MaverickTech [] has joined #mythtv
05:00-!-XChatMav [] has quit [Read error: 60 (Operation timed out)]
05:07-!-norcal_ [] has joined #mythtv
05:18-!-mykeul [n=mykeul@] has joined #mythtv
05:27<hads>heh, multirec is being weird. Changes channel to a completely different one than I request.
05:34-!-norcal_ [] has quit ["Leaving"]
05:47<gbee>yeah, it does that - thought it was fixed, but clearly not
05:48<gbee>hads: what it's actually doing is changing to the last channel you requested, so it's always one step behind
05:51<hads>gbee: Intersting, first time I've seen it. It wouldn't let me change to another input that was available to tune to another multiplex either.
05:51-!-ahbritto [] has joined #mythtv
05:52<gbee>is this with the latest revision?
06:09-!-sigger__ [] has joined #mythtv
06:13-!-sigger_ [] has quit [Read error: 104 (Connection reset by peer)]
06:29-!-mkasson__ [] has joined #mythtv
06:31-!-grokky [] has quit []
06:32-!-sigger__ [] has quit [Read error: 110 (Connection timed out)]
07:06-!-ma9mwah|tired [] has joined #mythtv
07:10-!-janneg [] has quit [Remote closed the connection]
07:15-!-janneg [] has joined #mythtv
07:15-!-ma9mwah|c2d [] has quit [Read error: 110 (Connection timed out)]
07:28-!-fstxx [] has joined #mythtv
07:34-!-MrGandalf [] has quit ["Leaving"]
07:56-!-renatofilho^ [n=renato@] has joined #mythtv
07:58-!-lsobral [n=sobral@] has joined #mythtv
08:04-!-czth__ [n=dbrobins@nat/microsoft/x-6bf08469c5fcb2d5] has joined #mythtv
08:17-!-davilla [] has quit ["Leaving"]
08:18-!-ma9mwah|c2d [] has joined #mythtv
08:20-!-czth_ [n=dbrobins@nat/microsoft/x-ee4e332183eaec83] has quit [Read error: 110 (Connection timed out)]
08:29-!-ma9mwah|tired [] has quit [Read error: 110 (Connection timed out)]
08:36-!-dekar1 [] has joined #mythtv
08:52-!-dekarl [] has quit [Read error: 110 (Connection timed out)]
08:57-!-sgm [] has joined #mythtv
08:58<sgm>hey guys, im using Fedora Core 4 and when i run ./configure it works ok but the make fails, the configure spits out "cat: /usr/lib/qt-3.3: Is a directory" about 10 times and then carries on as normal... any ideas what I've done wrong?
09:23<sgm>the make ends with errors from util-opengl.cpp
09:25-!-JoeBorn [] has quit ["Leaving"]
09:28-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv
09:31<sgm>the end of the make is here, if someone could have a look I'd appreciate it
09:38-!-grim[GameOp]_ [] has joined #mythtv
09:45-!-fstxx [] has quit [Connection timed out]
09:47-!-gardz [] has quit [Read error: 110 (Connection timed out)]
09:56-!-sgm [] has quit ["Leaving"]
10:13-!-Agrajag- [] has quit [Read error: 110 (Connection timed out)]
10:16-!-TelnetManta [n=benwilli@] has joined #mythtv
10:26<gbee>The change to mythweather script paths I made a while ago is causing problems for people who have already configured mythweather. Any objections to a DB update which truncates all the weather tables, wiping existing setups? It's by far the easiest solution - the alternative is to spend time making mythweather handle removed scripts better, which I intend to do anyway but could take longer than I really want to spend on mythweather befo
10:30-!-jgarvey [] has joined #mythtv
10:54-!-jmk_ [n=jmk@] has joined #mythtv
11:01-!-JoeBorn [] has joined #mythtv
11:18-!-xris [] has joined #mythtv
11:25-!-xris [] has quit []
11:26-!-jwhite [] has joined #mythtv
11:35-!-beavis [] has joined #mythtv
11:52-!-czth_ [n=dbrobins@nat/microsoft/x-5d994f23f62f0bc0] has joined #mythtv
11:59-!-xris [] has joined #mythtv
12:08-!-czth__ [n=dbrobins@nat/microsoft/x-6bf08469c5fcb2d5] has quit [Read error: 110 (Connection timed out)]
12:19-!-mykeul [n=mykeul@] has left #mythtv []
12:22-!-gnome42 [] has joined #mythtv
12:24-!-fstxx [] has joined #mythtv
12:30-!-ma9mwah|tired [] has joined #mythtv
12:32-!-saxin [] has joined #mythtv
12:32<dhr>does anyone here understand mythreplex or mythtranscode? I'm having serious problems with one of them.
12:32-!-saxin [] has left #mythtv []
12:42-!-ma9mwah|c2d [] has quit [Read error: 110 (Connection timed out)]
12:50-!-fstxx [] has quit [Read error: 110 (Connection timed out)]
12:51-!-danielk22 [] has left #mythtv []
12:53-!-danielito [] has joined #mythtv
12:53-!-danielito [] has left #mythtv []
12:56-!-Balachmar [] has joined #mythtv
13:04-!-danielk22 [] has joined #mythtv
13:14-!-feiner [] has quit [Remote closed the connection]
13:17-!-feiner [] has joined #mythtv
13:18-!-patdk-lap [] has quit [Read error: 113 (No route to host)]
13:21-!-otwin_ [n=otwin@] has joined #mythtv
13:22-!-otwin [n=otwin@] has quit [Read error: 131 (Connection reset by peer)]
13:22-!-jamesd [] has quit [Remote closed the connection]
13:24-!-mick_laptop [n=mweiss@clamwin/admin/mickhome] has quit [Read error: 104 (Connection reset by peer)]
13:24-!-sc00p [] has quit [Read error: 110 (Connection timed out)]
13:28-!-sc00p [] has joined #mythtv
13:34-!-danielk22 [] has left #mythtv []
13:54-!-Kaukasus [n=Alwin@] has joined #mythtv
13:55<Kaukasus>MOi Moin
13:55<Kaukasus>KAnn mir jemand beim Konfigurieren von Mythbuntu behilflich sein ?
13:58<Balachmar>Kaukasus: That question belongs in #mythtv-users
13:58-!-_gunni_ [] has joined #mythtv
13:59<laga_>Kaukasus: or #mythbuntu
13:59<laga_>Kaukasus: english preferred
14:06<GreyFoxx>janneg: You around ?
14:06-!-Kaukasus [n=Alwin@] has left #mythtv ["Verlassend"]
14:10-!-meshugga [] has quit [Remote closed the connection]
14:10-!-meshugga [] has joined #mythtv
14:10<dhr>Are Geoffrey Hausheer or Paul Harrison here?
14:12<Balachmar>Can anyone help me with my coding problems? I am trying to add LastFM scrobbling to MythMusic
14:13<Balachmar>But I can't find out how to take an md5 hash of a string...
14:14-!-Octane [] has joined #mythtv
14:17<hads>I don't know personally but grep the source for md5, it's bound to be used somewhere else.
14:22-!-mick_laptop [n=mweiss@clamwin/admin/mickhome] has joined #mythtv
14:27<Balachmar>You mean in MythTV?
14:27-!-mykeul [] has joined #mythtv
14:27<Balachmar>There are two md5's in mythtv none of them are to proces strings, one is for video or something the other for a number.
14:28<Balachmar>I hate feeling so stupid... (Since I know how to do it in say, Python )
14:30-!-MavT [] has joined #mythtv
14:31-!-Geshka1963 [n=MirandaL@nat/cisco/x-1fe678a856bc3239] has joined #mythtv
14:34<Tr1p> is "Search IMDB" working with u guys
14:35<Anduin>Tr1p: read the response to the same question in #mythtv-users
14:36-!-jgarvey [] has quit [Read error: 110 (Connection timed out)]
14:37-!-jgarvey [] has joined #mythtv
14:37-!-mick_laptop [n=mweiss@clamwin/admin/mickhome] has quit [Read error: 104 (Connection reset by peer)]
14:38-!-MaverickTech [] has quit [Read error: 148 (No route to host)]
14:49-!-danielk22 [] has joined #mythtv
14:50<GreyFoxx>danielk22: Do you know anything about our custom av_add_stream() code in avformat/utils.c ?
14:53<onyxsoft> Balachmar: I've used a version of this for md5 in C programs:
14:53<onyxsoft>Balachmar: This page may help:
14:55<danielk22>greyfoxx: a little bit
14:56<GreyFoxx>I'm trying to understand why at the beginning we drop the stream we are suppose to be adding
14:56<GreyFoxx>av_remove_stream(s, id, 0); is causing havoc with an h264 mp4 file I encoded
14:56<GreyFoxx>the video stream is dropped and then the player flipps out as it can't find any video :)
14:56<danielk22>if the stream used to be audio and is now video, or any other change of type, we need to get rid of the old one..
14:57<GreyFoxx>This is right at playback startup, mov_read_trak calls av_new_stream , which calls av_add_stream
14:57<GreyFoxx>av_add_stream is dropping the video stream
14:58<GreyFoxx>if I comment out that av_remove_stream() it plays without error
14:58<GreyFoxx>can't seek proper (just like mkv files) but at least it plays :)
14:58<danielk22>where is mov_read_trak?
15:00-!-xri1 [] has joined #mythtv
15:01<danielk22>hmm, from the comment at the top of av_add_stream it seems like this was never intended for non-ts streams to use..
15:02-!-xris [] has quit [Read error: 113 (No route to host)]
15:02-!-xris [] has joined #mythtv
15:04<GreyFoxx>I'm playing with encoding some of my DVD's to H264 and it wouldn't play in myth so off I went trying to find out why :)
15:04<danielk22>greyfoxx, why don't you just replace our av_new_stream() with the one from ffmpeg? It looks like someone was a little overambitious in trying to eliminate redundant code...
15:06<danielk22>the only place I could see this being a problem is the av_new_stream() in mpegts.c for the pcr stream, but that can be fixed if this resolves the problem with non-ts streams...
15:07<GreyFoxx>works fine for me using the one from ffmpeg
15:08<danielk22>Can you put the patch in one of the appropriate tickets and send me the ticket #, I'll try it out on some of my TS test files...
15:09<GreyFoxx>then if I can get it to seek properly (and hopefull mkv's too) then I can start mass reencoding my DVD collection
15:09<gbee>I'd be interested in checking it against a couple of problematic files
15:09<janneg>GreyFoxx: #4275 is probably related
15:09<gbee>that's one of the ones I wanted to test :)
15:10<GreyFoxx>janneg: hmmm that does look like it;'s the same issue
15:11<gbee>I've held off committing a patch which allows the internal player to be used in mythgallery because of #4275, probably too late to get that into 0.21 now though :(
15:12-!-danielk22 [] has left #mythtv []
15:18-!-mattwire [] has joined #mythtv
15:18-!-xri1 [] has quit [Read error: 110 (Connection timed out)]
15:21<GreyFoxx>gbee: patch is on that ticket if you wanna try it out
15:21<gbee>GreyFoxx: compiling now, thanks :)
15:22*GreyFoxx goes to learn about avformat seeking
15:22<mykeul>hi all
15:23<mykeul>I'm not happy with current implementation of "intelligent ordering" in mythmusic
15:24<mykeul>so I wrote a new one which is more closer to what I expect it to do
15:24-!-danielk22 [] has joined #mythtv
15:25<mykeul>should I fill a new ticket in track for review ? or is there any dev to look at my patch ?
15:26<Anduin>All patches should go through trac
15:26<mykeul>ok, thanks
15:27-!-otwin_ is now known as otwin
15:28<gbee>mykeul: could you sum up what's wrong with the current shuffle? Just curious
15:28<mykeul>current implementation seems to have side effects
15:29<mykeul>I have many 8 to 10 rated songs, but since a few day, I only get 4 rated songs
15:33<mykeul>what I did (and what I expected it to do) is: compute weights for all songs, then compute a random order based on that weights
15:33<gbee>yeah, that's probably because the playlist is linear and not quiet random enough that the high rated stuff is mixed in with the low rated stuff, so when you get down the bottom of the playlist you've only got low rated stuff left
15:33<mykeul>the higher the weight is, the higher chance tha song has to be played
15:35<mykeul>that song
15:35<gbee>mykeul: playcounts etc also factor into the equation, so if the higher rated tracks are played more often then the lower rated tracks will start to appear more frequently in the playlist - you can adjust the playcount/rating weighting in the settings
15:36<mykeul>I tried this, but that is not what I expect it to do
15:36<gbee>but I look forward to seeing your patch in trac, it may well perform better than the existing Intelligent shuffle
15:36<mykeul>If I rate a song with 8 to 10 stars, I expect it to be played more often
15:37<mykeul>but after a while, playcount and time since last play makes stars useless
15:37<gbee>mykeul: it is - until it's been played so often that it's playcount is much higher than the low rated songs and then the playcount rating comes into effect
15:38<gbee>mykeul: have you tried reducing the weighting on play count and last play to 0?
15:39<gbee>or at least much lower than they normally are? I'm not saying what you want is wrong, just checking that you can't get the desired behaviour from the existing code
15:41<mykeul>I want/need some sort of random order, so I kept a little weight on playcount and last play of 5 each, and 90 for rating
15:42-!-kormoc [n=kormoc@unaffiliated/kormoc] has joined #mythtv
15:42<mykeul>I submit the patch, may be easier to discuss it...
15:44-!-jams [] has quit [Read error: 104 (Connection reset by peer)]
15:44<xris>gbee: did you do the new weather stuff?
15:45<gbee>xris: no, was a SoC project by Lucien Dunning, I've just got stuck as the maintainer because no-one else will touch it :/
15:45-!-jams [] has joined #mythtv
15:46<xris>I tried to set it up yesterday and can't get it working.. it won't let me pick any scripts.
15:46<gbee>I've made several changes in an attempt to make it a more usable, but it still needs plenty of work
15:47<xris>it was briefly stuck on a couple of bbc scripts that no longer exist (looks like they were moved), so I wiped the db tables, but now I can't pick anything at all.
15:47<gbee>xris: I made a much needed change prior to 0.21 which might mean you have to truncate all the weather tables and set it up again
15:48<xris>I must have missed some of those tables, then.
15:48<kormoc>xris, also, I had to run the perl scripts by hand to see what modules they required. If it's missing a perl module, it won't show up in the ui at all
15:48<gbee>TRUNCATE weathersourcesettings; TRUNCATE weatherscreens; TRUNCATE weatherdatalayout;
15:49<xris>kormoc: interesting
15:49<gbee>I wanted to find a way to make that change without breaking existing installs, but all the alternatives were complex and too big this close to a release
15:49<xris>rpm should have caught that.. I'll check to see if it actually caught the perl requirements or not
15:49<xris>gbee: gotcha
15:50<gbee>the scripts directory was a mess and we needed to sort them out before 0.21, after 0.21 too many people would have been using the scripts in their old location and the change would have been even more disruptive
15:51<Balachmar>janneg are you there, because I think you just replied on the md5 problem
15:51<xris>gbee: yeah, I see that you can't just call the scripts without a chdir, either.
15:52<gbee>after 0.21 I want to stop it using the full script path in the database and instead have it search for the script in each available location - ala storage groups
15:52<mykeul>gbee: the patch is attached to ticket #4708
15:52<gbee>but like I said, there is a _lot_ of work that needs doing on mythweather
15:52<gbee>mykeul: ok, thanks
15:52<xris>kormoc: yup. definitely missed a couple of packages..
15:53<mykeul>gbee: thanks to you
15:54<gbee>I'd appreciate whatever help you or kormoc might offer on the perl side, simplifying installs etc
15:54<gbee>not sure anything can be done, but if you have any ideas :)
15:54-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
15:55<Balachmar>Just out of curiosity, I thought everything in mythtv was C++ or is it easy to use other languages as well?
15:56<Balachmar>Because I would have finished programming yesterday if I had used python :)
15:56<xris>gbee: so is libmythappearance ever going to be a separate piece of mythplugins, or is it officially part of libmyth now?
15:57<gbee>officially part of mythfrontend
15:58<xris>going to remove the refs from my specfile, then
15:58<gbee>libmythappearance technically doesn't exist any more since it's no longer a loadable library but a part of the mfe application
15:59<xris>I briefly had to create a package for it when it was inside of mythplugins
16:01<xris>ok, found the missing perl dep for the US scripts.. will test when I get home to see if it works
16:02<xris>hmm, or I could just vnc from here. heh
16:03<gbee>whole setup of mythweather sucks, I'm going to try my best to make it easier before 0.21 but my hearts not really in it
16:04<xris>knowing about the missing perl dep would have been a big help
16:04<xris>but that's solved in the packages noow
16:05<xris>I'll test the rest of this when I get home tonight.
16:07<kormoc>Balachmar, there are bindings for other languages (perl, php ones are being deved...), there's also projects like mythpython to add mythfrontend support via python
16:08<hads>And Python bindings
16:08<kormoc>ahh, didn't realize they were in yet
16:08<hads>Just, still a couple of issues pending.
16:09<gbee>the core parts of mythtv are c++ (and will probably remain that way), perl and python are both used for scripts and some addon functionality
16:09<hads>(and not as featureful as the Perl ones as yet.).
16:10<Balachmar>how would I use those python bindings? Or do you guys prefer that most of the stuff stays in C++
16:10<kormoc>Balachmar, you're writing for mythfrontend right?
16:10<Anduin>The plugin in python stuff isn't in there.
16:11<Balachmar>Because I could also just stick with C++ and learn to like it :)
16:11<Balachmar>Kormoc I am trying to add lastfm scrobbling to mythmusic at the moment
16:11<kormoc>did the mythpython project die?
16:11<kormoc>Balachmar, yeah, the python bindings won't help ya
16:11<gbee>depends who you ask, some of us are snobs about using scripting languages for stuff it wasn't really designed for
16:12<Anduin>I think that would be easier to argue with perl :)
16:12<gbee>perl/python are well suited for stuff which may change frequently - grabbers, scrapers, external website interaction etc
16:13<Balachmar>My problem is that I am spoiled with python, for instance with the md5 problem I am having now I would just use the md5 lib from python :)
16:14<Balachmar>But I will use this to make myself learn more on C++
16:15<kormoc><language war>md5 lib? bah, it's all built in with php! :P
16:15<Balachmar>It is better for me if I also learn a proper language :)
16:15<gbee>the problem with c++, if you can call it that, is that there are too many libs - c++ doesn't have CPAN etc
16:15<hads>Balachmar: The bindings are for interaction with the backend/database/etc. from external scripts and the likes, not actual frontend development as such.
16:15<kormoc>Yeah, it's given rise to the monster libs like qt and friends rather then small ones
16:16<Balachmar>hads: I had figured as much
16:16<danielk22>greg, you there? The patch on #4275 doesn't compile...
16:16<Balachmar>Well, also I find the docs like this one hard to understand:
16:16<gbee>that's why QT is great though, it offers a good set tools in a common package with documentation which rivals that of PHP
16:17<gbee>QT 4 has an MD5 function, but QT 3 doesn't unforunately
16:17<Balachmar>yes I had already found that one...
16:17<GreyFoxx>danielk22: hmmm compiles for me on 2 machines
16:17<GreyFoxx>I tested it on a second machine before I put it up there
16:18<GreyFoxx>still using trunk though
16:18<danielk22>av_add_stream is missing...
16:18<gbee>../../libs/libavformat/ undefined reference to `av_add_stream'
16:19<GreyFoxx>ahhh ok
16:19<gbee>whilst building the mythtv app
16:19<GreyFoxx>sorry fixing that now
16:19<Balachmar>hads: that one looks great!
16:19<GreyFoxx>strange that mine hasn't com,plained
16:21-!-mattwire [] has quit [Read error: 113 (No route to host)]
16:24<GreyFoxx>sorry, a differet patch up now
16:24<Chutt>anyone noticed any brokenness with the lcd support?
16:25<GreyFoxx>For some reason that other one does compile and run for me
16:25*GreyFoxx runs to get his kidlet from daycare
16:26<Balachmar>ok, I at least got it to compile, now let's see if it prints the right stuff!
16:28<Balachmar>it crashes... with this error: mythfrontend: symbol lookup error: /usr/local/lib/mythtv/plugins/ undefined symbol: _ZN3MD5C1Ev
16:28<Balachmar>any ideas?
16:34<Balachmar>I think I might have found the problem...
16:36<Balachmar>mmm nope didn't find it...
16:37-!-terr1 [] has joined #mythtv
16:44-!-mattwire [] has joined #mythtv
16:45<hads>Anduin: Which component would you prefer I create tickets for the Python bindings for? I forgot they aren't really mythvideo anymore.
16:47<Anduin>hads: I'm moving them to mythtv, just set me as the owner for now. Maybe one day there will be a real component.
16:48<hads>Anduin: OK, I can't do owner though, I think it must be registered users only.
16:48<Anduin>hads: and that was =G that did it, thought I caught all the moved returns but apparently not.
16:48<danielk22>GreyFoxx, that patch doesn't appear to cause any problems for my streams.
16:49<Anduin>hads: hmm, oh yeah :) Just let me know if I miss one, or keep doing them as mythvideo (eventually I'll figure out how to move them to mythtv without ijr being set to owner so I can skip the double change)
16:49-!-feiner [] has quit [Remote closed the connection]
16:50<hads>Anduin: Good as, thanks.
16:51<Anduin>hads: Do you have any test sitting around for the module? (If not I'll try to get one in there soon, clearly me running things manually isn't reliable)
16:52-!-mattwire [] has quit [Read error: 113 (No route to host)]
16:53<hads>Anduin: No real test, but I picked up on the last brokenness from this little status script that I use quite a lot; After I saw that I went through manually and looked at all the indentation, at least it's had two sets of eyes now.
16:54-!-feiner [] has joined #mythtv
16:55<Anduin>hads: Hmm, ok, I'll add some quick unit test to the bindings directory, just simple tests like I have for the various grabber scripts.
16:55<hads>Anduin: I could try and write some tests if you like (I don't have any unittest experience really though).
16:57<gbee>GreyFoxx: patch partially works for #4275, video plays though the sound is messed up which is probably another issue entirely
16:58-!-djc_ [n=djc@] has quit [Read error: 110 (Connection timed out)]
16:59-!-djc_ [n=djc@] has joined #mythtv
16:59<gbee>sounds like a helicopter - it's fine with ffplay
17:00-!-Balachmar [] has quit ["Ex-Chat"]
17:00<janneg>gbee: try ffmpeg -i $infile -vcodec copy -acodec copy outfile.avi
17:02<gbee>makes no difference, outfile.avi has the same audio problems
17:02-!-_gunni_ [] has quit ["KVIrc 3.2.4 Anomalies"]
17:04-!-mattwire [] has joined #mythtv
17:05<Cardoe>so.. what does --compile-type=profile do exactly?
17:05-!-TelnetManta [n=benwilli@] has quit ["Ex-Chat"]
17:06<danielk22>it creates a release compile with debugging symbols
17:07<danielk22>so that you can profile mythtv in v-tune or oprofile
17:08<Cardoe>so basically it doesn't strip it..
17:08<Cardoe>cause right now stripping happens to early..
17:09<Cardoe>i.e. make install shouldn't strip.. make install-strip should
17:10-!-feiner [] has quit [Remote closed the connection]
17:10-!-feiner [] has joined #mythtv
17:11-!-terr1 [] has quit [Remote closed the connection]
17:12<gbee>janneg: I think we need to supress some of the output from avcodec for mjpeg - it's _very_ noisy ~ 4-8 lines per thousand'th of a second
17:12<janneg>Cardoe: that's an qt3 qmake issue
17:12<Cardoe>janneg: oh how I hate qmake
17:12<danielk22>well normally --compile-type=release doesn't even use "-g" so it frees up some registers too..
17:13<Cardoe>-g doesn't necessary eat registers
17:13<danielk22>depending on the -O level..
17:13<Cardoe>it's just extra symbols in the result file
17:13<janneg>gbee: libav should be almost silent
17:14<danielk22>it also makes (at least certain versions of) gcc use a stack pointer.
17:14<janneg>gbee: is that with -v libav?
17:14-!-beavis [] has quit ["Verlassend"]
17:14<danielk22>and it makes the compile take much longer to complete
17:14<gbee>mythtv, which may use -v libav by default
17:15<gbee>I'll check what it's like from within the frontend
17:17<janneg>gbee: no, mythtv, doesn't enabled libav as default
17:18<danielk22>cardoe: the ubuntu packages use the profile target, except for compiling the mythtv ffmpeg libs
17:18<gbee>janneg: well no sign of that output when I play the video from within mythfrontend (mythvideo)
17:19<gbee>just the odd error from libav
17:19<gbee>so I don't think we need to worry about it after all
17:19<danielk22>this lets them get more useful stack traces, but still get most of the performance gains that release gives you.
17:20<danielk22>obviously the stack traces aren't 100% accurate, because code moves around at -O3...
17:20<Cardoe>danielk22: how do you change it for ffmpeg?
17:21-!-renatofilho^ [n=renato@] has quit ["Ex-Chat"]
17:21<danielk22>they wrote a little patch
17:21<danielk22>hmmm, I think it was only for alpha processors though.
17:22<janneg>danielk22: has PCR in a seperate PID. playback and av sync seems to be ok. hard to judge by that sample though
17:22<Anduin>I patch up the .pro file to get the same thing.
17:22<Cardoe>Anduin: patch?
17:23<Anduin>Cardoe: just to modify the QMAKE_CFLAGS_DEBUG
17:23<Cardoe>Anduin: right. got a link?
17:23<danielk22>janne: I think we need to test a stream where we append several streams together where the PCR pid of one stream == an a/v stream of another stream. And then see if it switches successfully.
17:24<Anduin>Uh, no, but all it does it -= the release stuff and += -O3 and fomit-frame-pointer
17:24<Cardoe>shouldn't need omit-frame-pointer on amd64 for the extra registers
17:25<Cardoe>but on x86, with ffmpeg you do if I remember ffmpeg correctly
17:25<Anduin>Yeah, not needed on x86 either anymore (comes with -Oanything
17:26<gbee>janneg: you need a test sample of this mjpeg file? I can probably record a short sample
17:27<janneg>gbee: yes, for the audio issue
17:28<Cardoe>--enable-dts isn't needed anymore? does Myth no longer link to dts/dca?
17:30<danielk22>i think mythtv uses some ffmpeg functionality now, but janne would know
17:31<danielk22>cardoe: I checked the ubuntu package, they just use a profile build. The patch I was thinking of is something else and only for alpha processors.
17:32-!-siXy [i=siXy@] has joined #mythtv
17:33<janneg>Cardoe: no, --enable-dts used to enable the dts passthrough which didn't require libdts. so I removed --enable-libdts
17:33<janneg>ffmpeg has native dca support and the wrapper is removed
17:38-!-jmk_ [n=jmk@] has quit ["Leaving"]
17:38<Cardoe>I'm trying to pass a user's -march and -mtune parameters directly in..
17:38<Cardoe>when --arch=athlon64 (because -march=athlon64 in my system configuration...)
17:38<Cardoe>ERROR: CPU specific ./configure options failed compile test
17:38<Cardoe> Removing CPU specific compilation options. ()
17:40<Cardoe>are you guys using --cpu for -march and --tune for -mtune?
17:40<danielk22>k8 ?
17:41<Cardoe>athlon64 is more appropriate, but k8 would work
17:41<Cardoe>athlon64 and k8 will be different in the future
17:41<janneg>Cardoe: read ./configure --help. use --cpu for -march and --tune for -mtune/-mcpu
17:41<Cardoe>janneg: I am.. it's not very clear.
17:42<janneg>--arch is for the general architecture only (x86, ppc, x86_64, ...)
17:44<janneg>hmm, I thought I had change the --arch help text
17:47-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
17:48<Cardoe>janneg: alright well I got it going
17:48<Cardoe>thanks for the help.
17:48<Cardoe>I'm sure I'll be around a bit more..
17:48<Cardoe>trying to package up some "beta" quality builds
17:48<Cardoe>rather then totally "meh" builds
17:48-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
17:49<gbee>janneg: did you get the PM?
17:53<janneg>yes, got the file already thanks
17:54<janneg>audio sounds like we are playing the audio too fast
17:59-!-Agrajag- [] has joined #mythtv
18:04-!-mykeul [] has left #mythtv []
18:24-!-jgarvey [] has quit ["Leaving"]
18:31-!-TelnetManta [] has joined #mythtv
18:52-!-greend139 [n=greend13@] has joined #mythtv
19:09-!-Geshka1963 [n=MirandaL@nat/cisco/x-1fe678a856bc3239] has quit [Remote closed the connection]
19:17<xris>GreyFoxx: moving a conversation in here... I'm curious why we've never worked to get our ffmpeg mods pushed upstream.
19:24<Vaelys>s/win 30
19:26<gbee>some of them have gone upstream, but I think the rest were considered to be hacks and/or no-one wanted to spend time fighting the cause with the ffmpeg devs
19:26<gbee>danielk22 can probably list some of the reasons
19:26-!-greend139 [n=greend13@] has quit []
19:30<janneg>gbee: can you open a new ticket for the audio problem
19:34-!-sc00p_ [] has joined #mythtv
19:37<xris>gbee: possibly also because the ffmpeg devs can be not-much-fun to deal with.
19:40-!-sc00p [] has quit [Read error: 113 (No route to host)]
19:41-!-feiner_ [] has joined #mythtv
19:43-!-feiner [] has quit [Read error: 104 (Connection reset by peer)]
19:45-!-lsobral [n=sobral@] has quit [Read error: 110 (Connection timed out)]
19:49-!-adante [] has joined #mythtv
19:53<danielk22>xris: it's mostly because our mods aren't upstream worthy yet, or are hopelessly tied into not quite upstream worthy code (like all the mpegts code)
19:54<xris>anyway, time for me to head home. :)
19:54-!-xris [] has quit []
19:55<danielk22>but there are some like the caption/subtitle/mheg stuff which could be pushed upstream if someone felt like pushing the ffmpeg devs.. I've managed to push a few bug fixes upstream, but only when the upstream bug was really obvious. Features are harder to push.
20:06<Chutt>kormoc, the new listings page quite often puts the popup description out of the visible window when mousing over near the right and bottom of the screen
20:07<Chutt>seems to be more often when it's first loading the info
20:07<kormoc>Chutt, if you wiggle the mouse, does it auto fix itself?
20:07<Chutt>no, but if i move and come back, it does
20:07<Chutt>but not alwasys
20:07<kormoc>I'll have to play with the position code a bit more then.
20:07<Chutt>i'd assume the position was being set before the info was loaded
20:07-!-mattwire [] has quit ["Leaving"]
20:07<Chutt>ie, the size of the rect was small cuz of no info
20:08<kormoc>it's not, but it might be before the reflow happens
20:08<Chutt>this does feel considerably faster, though
20:08<kormoc>That's great to hear, that was the goal
20:09-!-sc00p_ [] has quit ["This computer has gone to sleep"]
20:12<Chutt>now if only the 100% load in the backend on the recordings screen were to be fixed.. =)
20:13<Chutt>why does Yeechang Lee on the -users list want to deinterlace 1080p video?
20:13-!-sc00p [] has joined #mythtv
20:19<danielk22>doesn't make any sense to do that in mythtv..
20:19<Chutt>or anywhere?
20:19<Chutt>i just did a merge to trunk, btw.
20:20<danielk22>some players update the OSD more often..
20:21<Chutt>yeah, but, might as well just double the framerate, if you're decoding 24p, then
20:21<Chutt>no need to run an actual deinterlacer
20:23<danielk22>none of these are ideal.. really per fps a film projector gives you the best reconstruction of the moving image since the light is only on for an instant.
20:23-!-davilla [] has joined #mythtv
20:24<danielk22>but people like to run deinterlacers even when they don't have to, because it allows them to use a progressive modeline, which is better for desktop tasks.. plus when you do any software zooming you need to turn on deinterlacing..
20:24<Chutt>but it's already progressive
20:25<Chutt>zoom/crop will still be progressive
20:25<danielk22>yeah, frame doubling 1080p is a real head scratcher :)
20:25-!-xris [] has joined #mythtv
20:26<danielk22>It's in theory very bad and in practice a no-op :)
20:26<Chutt>well, if the deinterlacer doesn't check the flags, in practice it'll get you lesser quality video
20:27<danielk22>maybe he likes the video little blurry, takes the edge off it ;]
20:36<xris>gbee: still awake?
20:43-!-MrGandalf [] has joined #mythtv
20:46-!-siXy [i=siXy@] has quit ["bye!"]
20:51-!-t0ny-p40 [n=t0ny-p40@] has joined #mythtv
20:59-!-eskil [] has joined #mythtv
21:01-!-ma9mwah|c2d [] has joined #mythtv
21:03<Chutt>how do you disable the autodiscovery stuff for a backend?
21:04<hads>--noupnp I think
21:05<hads>Or --disable-autodiscovery if you mean for mythfrontend
21:05<Chutt>naw, need to stop this from thinking it's a slave
21:06<hads>OK, that's about the limit of my UPNP knowledge :)
21:11-!-ma9mwah|tired [] has quit [Read error: 110 (Connection timed out)]
21:15<Chutt>i wonder if the hdmi out on this board works in linux
21:30<Chutt>danielk22, deinterlacing doesn't seem to work if i switch from SD->HD channels
21:30<Chutt>i have a separate profile for > 720x480 than I do for <=, but...
21:30<Chutt>says it can't load the output type or the deinterlacer.
21:31<Chutt>and defaults to xv-blit w/ no filter
21:45<Chutt>sheesh, do we take a huge time to startup or what
21:52<Chutt>wonder why this is recording at 720x576
21:53-!-Tr1p [] has quit [Read error: 104 (Connection reset by peer)]
21:56-!-Tr1p [] has joined #mythtv
22:01-!-jmk [] has quit ["Leaving"]
22:07-!-jmk [] has joined #mythtv
22:09<Chutt>ProfileItem::IsMatch is busted.
22:09<Chutt>if you have rules: <= 720x480 and > 720x480
22:10<Chutt>720x576 matches _neither_ of them.
22:17-!-ma9mwah|tired [] has joined #mythtv
22:17<danielk22>Yeah, it's an implied "and"
22:18<Chutt>i'm changing the rule for > to be an or.
22:18<Chutt>could just rewrite the whole thing to use WxH as one value, too
22:18<Chutt>ie, actually multiply it
22:19<danielk22>well there are times that wouldn't work, like you want to turn off deinterlacing iff the number of lines == the number of lines in the display
22:19-!-davilla [] has quit ["Leaving"]
22:20<Chutt>you'd need to separate out the comparisons for width and height to do that properly.
22:20<danielk22>I figured making the cmp apply to both W & H gave you the ability to do whatever you wanted, with the cost of requiring you too define more matches...
22:20<Chutt>yeah, but, see above
22:21<Chutt>720x572 _is_ > 720x480
22:21<Chutt>i'd have to define it as <= and >= for it to work, which wouldn't really make sense to me
22:21<kormoc>What happens if it == if you define both <= and >=?
22:22<danielk22>well the # lines are but not the width...
22:22<Chutt>but the total image size is greater
22:22<Chutt>and it's one comparison in the ui
22:22<Chutt>kormoc, it'd _work_, but it's really weird logic
22:22<Chutt>i think:
22:22<Chutt> else if (cmp == ">")
22:22<Chutt> match &= (size.width() > width) || (size.height() > height);
22:22<Chutt>works for all situations
22:23<Chutt>since it's last in the list
22:23<Chutt>strike that last, but..
22:24<danielk22>how about a last "sloppy" comparison after the strict ones? would that work?
22:24<Chutt>the ui would still be confusing
22:25<Chutt>how about making == and != treat 0 as a wildcard?
22:25<Chutt>for your 'turn off deinterlacing iff the number of lines = number of lines in display'
22:26<Chutt>yeah, i can't think of a > situation where || doesn't work.
22:27<danielk22>using 0 as a wildcard would give some more power..
22:27-!-ma9mwah|c2d [] has quit [Read error: 110 (Connection timed out)]
22:28<Chutt>i still didn't figure out why it didn't reload the deinterlacer setting properly once it switched to the final video resolution from the dummy res
22:28<danielk22>The lines are matched in order so even if someone comes up with an ambiguous set of matches it will always select the same profile.
22:28<danielk22>I dunno, if you give me the -v playback I can prolly figure it out though..
22:30<danielk22>"same profile." -> "same profile, when given the same input video params."
22:30<Chutt>it's easy to repro, just setup two rules (<= 720x480, and > 720x480)
22:30<Chutt>then switch from SD to HD (with dummy rec)
22:30<Chutt>did it on two different machines
22:30<Chutt>just to make sure it wasn't the settings messed up in the db
22:31<danielk22>those are the only two rules?
22:31<Chutt>(sd for me is a pvr-150, hd is a hdhr)
22:31<Chutt>i need to not use the nicer deinterlacers on 1080i
22:32<Chutt>not enough processing power
22:32<danielk22>yeah, then it is not seeing any matches and the default is not to do any deinterlacing at all...
22:32<Chutt>for the dummy rec
22:32<Chutt>but once it hits 1080i after the channel change
22:32<Chutt>it still didn't do any deinterlacing
22:32<danielk22>ah, when it matches the 1920x1080 on the second rule, gotcha
22:33<Chutt>i'm checking in the || fix :p
22:33<Chutt>if you'd like to change that, that's cool
22:35<danielk22>i'd prolly prefer to just have separate cmp's for W&H, but I'll look at this more when I create some better defaults..
22:36<Chutt>don't want to complicate that screen more, though
22:37<danielk22>that's why i made it one cmp for both..
22:37<Chutt>right =)
22:38<Chutt>this is one screen where having the settings interface just be in the web interface would be a whole lot easier
22:38<Chutt>(not necessarily mythweb, but a internal server)
22:39<kormoc>I've actually been meaning to write one for mythweb, at least to match as much of mythsetup as I can
22:39<xris>kormoc: I've been holding off on that until there's an API
22:40<danielk22>my web 2.0 expertise is non-existent...
22:40<xris>would at least like a way to query the backend to find out what various fields/values should be.
22:40<danielk22>weird it is loading the deinterlacer here
22:40<danielk22>VDP: Accepting: cmp(> 720 480) dec(ffmpeg) cpus(0) rend(xv-blit) osd(softblend) osdfade(enabled) deint(linearblend,linearblend) filt()
22:41<danielk22>hmm something is broken though
22:41<danielk22>VDP: LoadBestPreferences(1920x1088, 25)
22:41<danielk22>VDP: Accepting: cmp(<= 720 480) dec(ffmpeg) cpus(0) rend(opengl) osd(softblend) osdfade(enabled) deint(linearblend,linearblend) filt()
22:41<danielk22>2008-02-18 22:38:11.505 VDP: Accepting: cmp(> 720 480) dec(ffmpeg) cpus(0) rend(xv-blit) osd(softblend) osdfade(enabled) deint(linearblend,linearblend) filt()
22:41<danielk22>that first one should be rejected
22:42<danielk22>i should make that debug a little clearer...
22:43<Chutt>i was confused by it =)
22:43<danielk22>yeah, me too!
22:44<danielk22>I can't believe I printed that and not FindMatch..
22:45<danielk22>k, I'll look at this tomorrow morning.
22:45<Chutt>not a huge deal
22:45<Chutt>the fix i checked in seems to work
22:45<Chutt>good enough for the release, unless it has some side-effects
22:46<danielk22>Still it's one of the bullet points for the release, I'd like to make it a clear improvement from the end user's perspective.
23:09-!-Netsplit <-> quits: adante, gbee, grim[GameOp]_, Cougar, MavT, sc00p, sphery, dlblog, eharris, dekar1, (+7 more, use /NETSPLIT to show all of them)
23:10-!-Netsplit over, joins: sc00p, adante, MavT, otwin, grim[GameOp]_, dekar1, PointyPumper, tomimo, nemik, eharris (+7 more)
23:11-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Connection timed out]
23:51-!-Tanthrix [] has quit [Read error: 104 (Connection reset by peer)]
23:55-!-Tanthrix [] has joined #mythtv
---Logclosed Tue Feb 19 00:00:11 2008