Back to Home / #mythtv / 2008 / 07 / Prev Day | Next Day
#mythtv IRC Logs for 2008-07-30

---Logopened Wed Jul 30 00:00:06 2008
00:06-!-_gnome42 [] has joined #mythtv
00:13-!-reynaldo_ [n=rverdejo@] has quit [Read error: 110 (Connection timed out)]
00:20-!-gnome42 [] has quit [Read error: 110 (Connection timed out)]
00:23-!-_gnome42 [] has quit []
00:47-!-briand [] has quit [Read error: 104 (Connection reset by peer)]
00:59<teprrr>patches incoming!
01:03<teprrr> there you go, if someone is willing to commit it.. makes mythlcdserver to respect dateformat setting when displaying the date on lcd screen
01:05-!-briand [] has joined #mythtv
02:09<teprrr>hey, can someone say whether it's enough if only one client can connect to mythlcdserver at the time?
02:10<teprrr>hrm, yup, currently it does allow multiple connections
02:10<teprrr>though it may be messy..
02:12<teprrr>yeah, my fault.. duh
02:18-!-adante [] has quit [Read error: 101 (Network is unreachable)]
02:28-!-edannenbe [] has joined #mythtv
02:45-!-Dibblah [] has joined #mythtv
02:47-!-famicom [] has quit ["Leaving"]
02:49-!-rooau1 [] has quit [Read error: 110 (Connection timed out)]
02:50-!-rooaus [] has joined #mythtv
02:51-!-famicom [] has joined #mythtv
03:06-!-foxbuntu [] has quit ["Leaving"]
03:13-!-adante [] has joined #mythtv
03:53-!-edannenbe [] has quit [Read error: 113 (No route to host)]
03:53-!-edannenbe [] has joined #mythtv
03:57-!-noisymime [n=josh@nat/ibm/x-834b9ebcfc4816c8] has quit [Read error: 110 (Connection timed out)]
05:54<gbee>of course I now have to rebuild it as a debug build because it segfaults in AvFormatDecoder::InitByteContext on all video
06:23-!-dagar [] has quit [Read error: 110 (Connection timed out)]
06:27-!-dagar [] has joined #mythtv
06:31<gbee>janneg: backtrace of segfault with ffmpeg sync branch -
06:42<MrGandalf>gbee: where's the segfault?
06:44-!-MrGandalf [] has quit ["work"]
06:47<gbee>AvFormatDecoder::InitByteContext avformatdecoder.cpp:808
06:47<gbee>ic->pb->buffer_size = 32768;
06:48<gbee>can't play any recordings (mpeg2 TS), segfaults immediately
07:08<clever>161% transcode progress again
07:16<gbee>always nice to get more for your money
07:30<janneg>gbee: yeah, due to the unnoticed libs it's completely untested
07:32<janneg>+install error
07:39-!-lyricnz_ [] has joined #mythtv
07:50-!-leprechau [] has quit [Connection timed out]
07:50-!-leprechau [] has joined #mythtv
07:55-!-robthebob [] has joined #mythtv
07:59<teprrr> -- can anyone make a quick check to see if I've ported those correctly?
07:59<teprrr>also modified to use QSomeClass style includes
08:00<teprrr>and serversocket.{cpp,h} are not needed anymore at all with QTcpSocket
08:00<teprrr>and that patch also includes more error messages
08:00<teprrr>for lcdproc socket connection
08:01<teprrr>oh.. and accidentally my earlier dateformat fix.. nm about that one
08:01-!-|gunni| [] has joined #mythtv
08:09-!-joobie [] has joined #mythtv
08:15-!-anny__ [n=chatzill@] has joined #mythtv
08:17-!-_gunni_ [] has quit [Read error: 110 (Connection timed out)]
08:28<teprrr>small change for lcd functionality, displays both channel name and program name on separate lines, if there's more than one line available on lcd:
08:34-!-matt_ubuntu [] has joined #mythtv
08:34-!-matt_ubuntu [] has left #mythtv []
08:42-!-joobie [] has quit ["This computer has gone to sleep"]
08:54-!-MrGandalf [] has joined #mythtv
09:09-!-famicom [] has quit [Read error: 110 (Connection timed out)]
09:13-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv
09:21-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit ["Konversation terminated!"]
09:22-!-famicom [] has joined #mythtv
10:17<MrGandalf>janneg: in the ffmpeg merge, did you end up changing filters/ to include ../libs?
10:25-!-gustave_ [] has joined #mythtv
10:28-!-MrGandalf [] has quit [Remote closed the connection]
10:29-!-MrGandalf [] has joined #mythtv
10:31-!-anny__ [n=chatzill@] has quit ["ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]"]
10:35-!-tomimo [] has quit ["Gotta get going ..."]
10:37-!-iamlindoro_ [] has joined #mythtv
10:37-!-iamlindoro_ [] has left #mythtv []
10:43-!-matt_ubuntu [] has joined #mythtv
10:43-!-matt_ubuntu [] has left #mythtv []
10:45<MrGandalf>janneg: Ah, I see why you're saying ac3 passthrough is broken..
10:55-!-cattelan is now known as cattelan_away
11:05-!-cattelan_away is now known as cattelan
11:06<MrGandalf>gbee: there?
11:08<MrGandalf>in ic->pb->buffer_size = 32768; ??
11:09<MrGandalf>I back ported janneg's merge into 16384 (practically -fixes)
11:09<MrGandalf>but I'm also seeing a segfault in AvFormatDecoder::InitByteContext().
11:09<gbee>that's what the backtrace says, I've not looked into it since I don't really have the time
11:10<MrGandalf>ok, just wanted to make sure it wasn't anything I introduced.
11:10<MrGandalf>thanks :)
11:10<janneg>there was an api change in ffmpeg changing a member to a pointer
11:10<janneg>I'll look at it now
11:11<MrGandalf>yes, ic->pb.buffer --> iv->pb->buffer
11:11<MrGandalf>(as an example)
11:12<janneg>problem might be that the user has to allocate the buffer
11:12<MrGandalf>would make sense
11:12<janneg>that would atleast explain the segfault
11:13<stuarta>might make processing more efficient if we can just set the pointer to the relevant bit of packet
11:13<janneg>it's not a simple char array
11:21<sphery>teprrr: Regarding your patch--I'm not sure what the official dev position is--but I've been leaving the Qt3-style "#include <qwhatever.h>" and only use Qt4-style "#include <QWhatever>" for new includes I add. That way a) the include-style change can be done as a separate commit since it's unrelated to my changes and b) we don't have a million patches that all make the same changes (and all become invalid when the first is applied).
11:22<gbee>I've been making the changes as I go along to any files I touch in commits
11:22<sphery>teprrr: Not every patch is applied as quickly as your editable combobox patch (see and count the "patch" rows :).
11:22-!-iamlindoro_ [n=iamlindo@] has joined #mythtv
11:24<gbee>easier to say - there are 89 patch tickets
11:25<sphery>I don't know Trac well enough to figure that out without actually counting... :)
11:25<gbee>there are 378 defect tickets
11:26<gbee>none of which is a clear indication of how long it might take, patches for bugs have a higher priority than bugs without patches unless those bugs are considered severe
11:27<sphery>Yeah, just trying to indicate that there are a lot of patches in there, some of which are rather old, so new functionality patches (like the LCD server changes) can sit for quite a while.
11:27-!-xris [] has quit []
11:28<gbee>if the patch fixes an area that a dev knows well, or is currently working on then it rises up the list - if no-one knows that area or the dev responsible is busy then it could take a while
11:28<gbee>in short there really isn't any way to know how long it could take for a patch to be looked at or committed
11:30<gbee>not enough devs, not enough time
11:31<teprrr>sphery, yeah, just tidied other things up and decided to clean-up useless includes and also modify to use qt4-style ones.. though that patch in my webhost was there just if someone has something to comment on it
11:32<teprrr>sphery, and yes, I can see the problems with merging from different places when those have been changed.. but if you want to do clean-up sometime, it's better to do it once, at least sometimes that's it
11:32<teprrr>s/that's it/that is/
11:33<sphery>Yeah. If nothing else, separate patches could be nice. (And quilt makes it easy.)
11:33<teprrr>sphery, and yes, I've seen that the patches aren't applied always so quickly
11:34<teprrr>hmm, quilt allows you to easily split patches even with svn?
11:34<teprrr>I haven't actually used it ever.. just wondering :)
11:35<teprrr>sphery, though the good thing in patching the lcdserver is that it doesn't affect so many places.. so even though those changes would be commited soon, it shouldn't broke up other functionalities :P
11:35<gbee>allows patches to be layered, from what I've read in the manual, which unfortunately wasn't what I was looking for
11:36<sphery>quilt allows you to easily make patches on top of other patches (even when they change the same file). You don't generate them with svn diff, but they still work like svn diff patches. :) See /usr/{,local/}share/doc/quilt*/quilt.pdf for a nice overview.
11:39<gbee>I want something that lets me keep a big pot of patches whilst managing conflicts, generally if I've two patches which affect the same file they aren't related, so patch B might be committed before the earlier patch A
11:39<sphery> was done with quilt because it modified files modified by other patches in the ticket (quilt new <patchfile> ; quilt add <file> ... ; quilt refresh).
11:39<gbee>mainly the need to predeclare that I'm about to editing a file is the feature which I dislike about quilt
11:39<teprrr>yup, I'll need to take a look at that pdf
11:40<stuarta>gbee: that's easy to work around if you forget
11:41<sphery>gbee: Yeah, it won't help with conflicts. You actually need to manage conflicts yourself and generate multiple series files (or just guess which way they'll be applied and only fix conflicts if you guess wrong).
11:41*stuarta loves quilt
11:41<teprrr>what I've been looking for a while is a way to select only some changes from the diff for a patch
11:41<stuarta>i'm using it to do the MHEG QT4 port atm
11:41<teprrr>but I'll take a look at quilt tomorrow or so
11:42<stuarta>patch 1 is daniel's patch, patch 2 is my additions
11:42<gbee>maybe I'll try again, but although pretty messy I have a system which seems to work for me
11:42<stuarta>it's what works that's important
11:43<sphery>Yeah, stuarta , I forget quite often, too. I just work around it with svn diff > ~/patch ; svn revert <file> ; quilt add <file> ; patch -p0 -i ~/patch (but now I'm wondering if there's an easier way)
11:46-!-bahadunn [n=bahadunn@] has joined #mythtv
11:47<bahadunn>can mythtv do video playlists?
11:47<sphery>bahadunn: #mythtv-users
11:47-!-De_Wr0ng [n=Steven@] has joined #mythtv
11:47-!-bahadunn [n=bahadunn@] has left #mythtv []
11:54<janneg>gbee, MrGandalf the ic->pb issue should be solved but I test it before I push it out
11:54<teprrr>talking about those patches, it'd be nice to get that translation file included for the next release.. Finnish translation of Mythtv is currently quite miserably..
11:55<gbee>janneg: ok, the missing include path issues fixed too? If not I can maybe figure out what mythtranscode needs, the branch doesn't include the plugins which are broken so ..
11:56<gbee>translation fixes can be backported to 0.21-fixes, but they'd have to apply to that branch
11:56<janneg>gbee: the includes are fixed and already on the server
11:58<janneg>ffmpeg changed the cross library includes to "library/header.h" so we have to enclude libs/ in the main mythtv tree
11:58-!-tomimo [] has joined #mythtv
11:59<janneg>for the plugins I might to have to change the ffmpeg header install locations
12:01<gbee>does git have a revert type command?
12:02<sphery>I see git-revert in my /usr/bin...
12:02<janneg>git revert or if your changes aren't public yet you can easily change history
12:02<teprrr>hmm, coolness, someone just added mythtube plugin to trac.. :o
12:02<teprrr>but gotta get some sleep, happy hacking
12:02<gbee>sphery: so there is, it's not listed in the git help output
12:03<teprrr>gbee, hm. no idea how to do the applying, but I'm using the file created for trunk with my 0.21-fixes installation too
12:03<janneg>for example if you want to change an error in the last commit you can use git commit --amend
12:03<gbee>these are uncommitted local changes
12:04<iamlindoro_>janneg: I didn't pay attention to my history on the home computer, so didn't know if you had seen it, but did you see my backtrace for the preview generation segfault w/ your branch? I know it's not anything official so I didn't want to bother you with it if you didn't care.
12:05<janneg>iamlindoro_: that's probably the same problem I'm fixing atm
12:05<iamlindoro_>oh, ok, sorry about that
12:05<janneg>gbee: ah, git reset --hard
12:07<janneg>iamlindoro_: why, the main problem was that I hadn't noticed the installing into wrong directories and published therefor something completely untested
12:07*gbee sobs as his checkout stays at the same revision
12:08<gbee>oh that's just wierd, had to re-run the command a couple of times before getting the updates
12:08<iamlindoro_>janneg: I just didn't want you to think I felt entitled to support on it :)
12:09<gbee>git is going to be the death of me
12:10<gbee>last commit was "prefix fixes in configure"?
12:10-!-edannenbe [] has quit [Remote closed the connection]
12:11<MrGandalf>janneg: did you say your 00-ffmpeg-sync is updated then?
12:13<janneg>yes, but it doesn't ... argh, I forgot to run ./configure
12:14<janneg>MrGandalf: the ic->pb fix is not pushed yet
12:15<janneg>I want to test it before I push it again
12:15<MrGandalf>janneg: if you could point me to where you're fixing it (once you've pushed it), that would help me greatly.
12:36-!-gnome42 [] has joined #mythtv
12:57-!-kormoc [n=kormoc@unaffiliated/kormoc] has joined #mythtv
13:09-!-De_Wr0ng [n=Steven@] has left #mythtv ["Leaving"]
13:13<janneg>MrGandalf: pushed, the top commit fixes playback
13:17-!-PointyPumper [n=pintlezz@] has quit [Connection timed out]
13:18<MrGandalf>thanks :)
13:27<iamlindoro_>janneg: Yep, fixed thumbnail generation, nice job
13:40*MrGandalf is a sudden dislike for git
13:48-!-jgarvey [] has joined #mythtv
13:51-!-gnome42 [] has quit []
13:56-!-gnome42 [] has joined #mythtv
13:58<gbee>Welcome to the club
13:58<iamlindoro_>janneg: Not sure if this has been discussed, but think this AM's changes may have broken mythmusic compilation
14:00<gbee>iamlindoro_: pointed it out this morning
14:00<iamlindoro_>ah, ok
14:01<clever>iamlindoro_: did you see my comment about your OSD bug?
14:01<iamlindoro_>clever: Did I have an OSD bug? :)
14:01<clever>you mentioned how the osd wasnt updating right
14:01<clever>30 00:39:48 < iamlindoro> Anyone noticed in trunk that when skipping in the internal player, it reflects the timestamp of the time *before* the skip rather than the one
14:01<iamlindoro_>ah, no, probably in my logs at home, but I'm at work
14:01<clever>it feels like the timestamp only updates on a key press
14:01<clever>before the seek happens
14:02<gbee>it's a shame since the whole point of my grabbing and building the ffmpeg branch was to see if it fixed the mp3/flac playback bugs
14:02<clever>if i leave the OSD up it doesnt count the seconds away anymore like it normaly did
14:02<janneg>hmm, maybe I should have created a single repo
14:02<clever>its frozen until i hit a key
14:02<iamlindoro_>clever: yeah, same here
14:03<clever>so for once we can be shure its not my config to blame:P
14:03<janneg>and I think the current setup is not optimal for sharing the work
14:04<janneg>but it makes my work much easier
14:07<gbee>ok, seems the magic formula is "git pull; git checkout origin/00-ffmpeg-sync"
14:09*iamlindoro_ is extra-confused... mythmusic just compiled on one machine, but not on the other
14:09<gbee>iamlindoro_: the git ffmpeg branch?
14:10<iamlindoro_>I have no doubt it's me... would just love to know *how* it's me
14:10<gbee>doesn't compile here
14:10<janneg>gbee: git checkout master; git checkout --track 00-ffmpeg-sync origin/00-ffmpeg-sync
14:11<janneg>then a git pull in the 00-ffmpeg-sync should work as expected
14:11<gbee>janneg: I seem to be thanking you a lot lately, but thanks ;)
14:11<gbee>git checkout: --track and --no-track require -b
14:12<janneg>working in remote branches is not intended by git hence no auto update of the checkout
14:12<janneg>gbee: yes, I missed that
14:12<gbee>git checkout: updating paths is incompatible with switching branches/forcing
14:13<janneg>gbee: no need to thank since I gave you the wrong command earlier
14:13<janneg>gbee: with 'git checkout --track -b 00-ffmpeg-sync origin/00-ffmpeg-sync'?
14:14<gbee>yeah, got it - had -b and --track switched
14:18<sphery>iamlindoro_: I was able to compile MythMusic in trunk r17981 . If your theory about system ffmpeg getting in the way is true, my system has an old ffmpeg on it (older than the one in use in Myth), so...
14:18<MrGandalf>anyone tested with ac3 audio? (not passthrough)
14:19<MrGandalf>specifically using enable-liba52bin
14:19<iamlindoro_>sphery: Yeah, I think I found my user error, looks like this AM's compile didn't get installed, so it was compiling properly against the ffmpeg libs from yesterday
14:20-!-famicom [] has quit [Read error: 104 (Connection reset by peer)]
14:21-!-famicom [] has joined #mythtv
14:21<gbee>MrGandalf: native decoder no good?
14:21<MrGandalf>gbee: volume too low
14:22<iamlindoro_>sphery: Think it's likely a case of git being too smart for me
14:25<janneg>MrGandalf: the amplifying is also currently reverted
14:26<janneg>compiling the plugins now
14:33<MrGandalf>hmm, for some reason SSE3 got enabled..
14:33<gbee>SSE3 is bad now? I've been compiling with it enabled for weeks
14:34<MrGandalf>for core2?
14:35<gbee>well obviously if the arch doesn't support it then it's not good ;)
14:35<gbee>although I thought Core 2 did
14:36<MrGandalf>flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 cx16 xtpr lahf_lm
14:37<MrGandalf>also, is it normal for libav* to compile with -march=k8 for a core2?
14:37<gbee>MrGandalf: pni == SSE3
14:38*sphery wonders if the 'echo "$1 IS at $LIBPATH"'/'echo "$1 not at $LIBPATH"' in configure's has_library() in -fixes was accidentally left-in debugging in
14:38<gbee>Prescotts New Instructions or something
14:39<gbee>dunno if k8 is correct for core2
14:40<MrGandalf>k8 looks to be an override for p4, which would make sense for core2
14:40<MrGandalf>just getting a very off segfault in put_h264_qpel8_h_lowpass_l2_ssse3()
14:41<gbee>I'm using k8-sse3 for sse3 instructions with gcc 4.3, but on an AMD 64 cpu
14:41<gbee>k8 doesn't support sse3
14:44<MrGandalf>probably missed something in configure..
14:44<gbee>try "prescott" instead of "k8"
14:44<gbee>or nocona
14:45<gbee>heh, ignore all that - core2
14:45*gbee should have read all the way down first
14:45<gbee>core2 supports ssse3 as well as sse3
14:47<janneg>MrGandalf: h264 seems to work fine here but I just noticed that my source isn't using multiple slices anymore
14:48-!-huhlig [] has joined #mythtv
14:51<iamlindoro_>sphery: Yep, was indeed /usr/include/libav* being picked up first
14:51<janneg>gbee: mythmusic complies fine here but my ffmpeg system headers are new enough
14:52<janneg>uninstalling them now
14:54<iamlindoro_>janneg: yeah, likewise here, I installed my newest ffmpeg onto the other system and mytnusic cooperates now
14:54<gbee>'cept it shouldn't be using the ffmpeg headers, it should be using our copies :)
14:55<sphery>compiles fine here with ffmpeg SVN r6732 (from the dark ages, I think), so it seems to work if ffmpeg is old enough, too
14:56<gbee>actually that's odd, it's pulling in "mythtv/ffmpeg/avcodec.h" but still managing to link against the ffmpeg copies?
14:57<gbee>sphery: this is with the ffmpeg resync branch and as I said,it shouldn't be using the ffmpeg headers but ours
14:57<gbee>mytharchive is currently failing - "libavutil/avutil.h: No such file or directory"
14:57<sphery>Oh. nvm, then. I didn't compile that one.
15:02<MrGandalf>definately sse3 related
15:02<MrGandalf>compiled with pentium4 and no segfault
15:04<MrGandalf>I'll have to study the differences between the older and newer i386/h264dsp_mmx.c
15:05<MrGandalf>could be a compiler thing as well.. only using gcc 4.1
15:12<gbee>MrGandalf: what about a more appropriate arch like core2 ?
15:12-!-Chutt [] has joined #mythtv
15:13<MrGandalf>gbee: I could do that, but core2 only works on later compiler versions. I'm making a patch against -fixes for a few people with differing gcc versions.
15:13<gbee>ahh ok
15:17<MrGandalf>looks like someone else was complaining about the same thing on the ffmpeg-dev list
15:21<MrGandalf>looks like an alignment issue with the compiler.. maybe 4.2 will fix it.
15:25<gbee>install ffmpeg head so I can compile mythmusic for now
15:25<MrGandalf>looks like gcc 4.2 fixes the issue..
15:26<MrGandalf>janneg: something to note
15:26<sphery>danielk_Zzzzzz: I just re-tested -fixes with and without the main patch on #5314, and--because of Nigel's changes in [17166] , -fixes creates the DB without issues now. #5314 is now just cosmentic--simply preventing errors (that don't hurt anything) from occurring/being logged . So, I'd recommend closing #5250 as fixed by [17166] .
15:29<gbee>well the ffmpeg re-sync breaks the avfdecoder in mythmusic, guess there was an API change or something
15:29<gbee>guess_format() is failing
15:29-!-danielk_Zzzzzz is now known as danielk22
15:30<danielk22>sphery: thx
15:31<sphery>Sorry if I woke you up... Meant for you to catch it in scrollback.
15:32<danielk22>i've been awake for a while now, just not IRCing
15:40<gbee>anyone know what the audio device name is in ffmpeg now? used to be "audio_device", then it was changed to "oss", now it's ... ?
15:43<janneg>gbee: avfdecoder uses libav* for playback?
15:44<janneg>then it can't work since I didn't add libavdevice
15:44<gbee>#include <mythtv/ffmpeg/avformat.h> #include <mythtv/ffmpeg/avcodec.h>
15:44<gbee>janneg: that's new?
15:45<janneg>gbee: it's a new library splitted all hardware access from libavformat
15:46<gbee>uses audiooutput actually, but libav* for decoding to a PCM format for audiooutput
15:46<gbee>so I'm not sure it needs libavdevice?
15:47<janneg>I don't think so. I look at it after fixing the plugins build
15:48<gbee>right now it's failing at guess_format (libavformat) which is supposed to provide the output format -
15:50-!-Chutt [] has quit [Remote closed the connection]
15:52<gbee>from looking at some mailing list posts I think we do require libavdevice after all, the hardware output formats have been moved there too
15:52<gbee>so you have to use avdevice_register_all(); to add oss to the list
15:52-!-leprechau [] has quit [Read error: 110 (Connection timed out)]
15:53-!-leprechau [] has joined #mythtv
15:54<janneg>gbee: apply after updating and installing mythtv
15:57<gbee>so pull first?
15:59<janneg>gbee: are you sure that using OSS for guess format is good? wouldn't be one of PCM_* better
15:59<janneg>gbee: yes
16:01<gbee>janneg: I've no idea which one is best, Paul switched to oss 7 months ago - I've switched it to pcm_s16le for now which has resolved the guess_format issue but unfortunately it's still failing to play
16:03<janneg>pcm_s16le would be the same as oss on little endian
16:05<gbee>building now, had to resolve a conflict in
16:16-!-adante [] has quit [Remote closed the connection]
16:16-!-adante [] has joined #mythtv
16:19<gbee>that's the sort of bug that stuarta was talking about the other day
16:19-!-kormoc [n=kormoc@unaffiliated/kormoc] has quit []
16:22<gbee>janneg: bunch of recordings are failing to play :(
16:26<MrGandalf>gbee: of what type?
16:26<janneg>gbee: segfault or just returning to the menu?
16:26<gbee>janneg: either returning straight to the menu or getting stuck in a prebuffering loop
16:27<gbee>MrGandalf: all from the same source - DVB-T, mpeg 2 TS
16:28-!-adante [] has quit [Remote closed the connection]
16:28-!-adante [] has joined #mythtv
16:29<gbee>some recordings from the same channel are fine
16:30<janneg>tried more than ten recent recordings and switched through 10 channels without problem
16:31<danielk22>I've seen this too in head.. something is wrong with signal monitoring, maybe more is broken. Possibly Qt4 related.
16:31<gbee>always seems to be the same ones, and represents about 1 in 20
16:31<danielk22>It only happens with digital, not analog
16:31<gbee>they play fine with trunk
16:33<MrGandalf>gbee: did you look at -v libav?
16:33<gbee>some exit straight away, other get stuck - logs look different for each
16:36<gbee>MrGandalf: doesn't seem to add anything at all to the logs, whether -v libav is broken or whether it just isn't saying anything I don't know
16:36-!-foxbuntu [] has joined #mythtv
16:36<MrGandalf>gbee: I've found that "-v libav" doesn't play well with "-l <log file>"
16:37<gbee>not using -l
16:37<MrGandalf>so I always run it in a shell or redirect it to a file.
16:37<janneg>gbee: might be the allmost unmodified mpegts.c, ignore it for now
16:37<MrGandalf>It should at least show you the streams in the file, if it's TS.
16:38<gbee>MrGandalf: just checked, it shows streams for the files which work but not the ones which don't
16:39<MrGandalf>I would try playing one of those with the latest ffplay
16:40<gbee>:/ I'll have to nfs mount the recording directories first
16:44<gbee>keep hitting "MythSocket(95ca90:37): writeStringList: Error, invalid string list." too
16:44<gbee>not sure if that's also true with trunk, but it's happening a lot with the ffmpeg branch
16:45-!-adante [] has quit [Remote closed the connection]
16:45-!-adante [] has joined #mythtv
16:46<gbee>ok, it's really screwed up - some recordings I get stream info - you'll love this :)
16:46<gbee>Input, mp3, from 'myth://'
16:46<gbee>Input # 0, mp3, from 'myth://':
16:46<gbee>Duration: 13:35:58.05, start: 0.000000, bitrate: 160 kb/s
16:47<gbee>Stream, 1/90000: Audio: mp2, 48000 Hz, stereo, 160 kb/s
16:47<janneg>where comes from?
16:48<gbee>janneg: sorry, that's a feature of my IRC client, it's replacing numbers prefixed with # with a link to the ticket
16:49<gbee>added it to save myself typing the entire url or copy/pasting it - so I type #{nospace}1234 and it expands it to the full ticket address
16:50-!-adante [] has quit [Remote closed the connection]
16:50-!-adante [] has joined #mythtv
16:52<janneg>trac#{nospace}1234 would be saver
16:53<janneg>but I've also thought of adding something like that to my irc client
16:54-!-adante [] has quit [Read error: 54 (Connection reset by peer)]
16:54<gbee>works on incoming and outgoing strings, also have [ 1234 ] set to link to changesets
16:55<janneg>otoh "ticket 1234" opens #1234 in my browser
16:55-!-adante [] has joined #mythtv
16:56<janneg>same with 'change 123' using konqueror's web shortcuts
16:57-!-adante [] has quit [Read error: 104 (Connection reset by peer)]
16:58<gbee>ok, modified the regexp so that four digits are required
16:58-!-adante [] has joined #mythtv
16:59<gbee>plugins patch seems to help the build
17:02-!-gustave_ [] has quit ["ChatZilla 0.9.83 [Firefox]"]
17:02-!-adante [] has quit [Read error: 104 (Connection reset by peer)]
17:03-!-adante [] has joined #mythtv
17:03<gbee>mythmusic still doesn't work, seems it doesn't like pcm_s16le any better than oss
17:07-!-adante [] has quit [Connection reset by peer]
17:07<janneg>doesn't work here either, more things to fix before the merge can happen
17:07-!-adante [] has joined #mythtv
17:07<gbee>oh well
17:08-!-MrGandalf [] has quit ["Leaving"]
17:11-!-adante [] has quit [Read error: 104 (Connection reset by peer)]
17:11-!-adante_ [] has joined #mythtv
17:11-!-adante_ is now known as adante
17:16-!-cattelan is now known as cattelan_away
17:16-!-adante [] has quit [Read error: 54 (Connection reset by peer)]
17:16-!-adante [] has joined #mythtv
17:17-!-cattelan_away is now known as cattelan
17:18-!-adante [] has quit [Remote closed the connection]
17:19-!-adante [] has joined #mythtv
17:23-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
17:24-!-adante [] has quit [Remote closed the connection]
17:24-!-adante [] has joined #mythtv
17:25-!-dekarl [] has joined #mythtv
17:31-!-dekarl1 [] has quit [Read error: 145 (Connection timed out)]
17:47-!-lyricnz_ [] has quit []
17:51-!-jgarvey [] has quit ["Leaving"]
18:01-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
18:11-!-Netsplit <-> quits: jk1joel, phedny
18:13-!-Netsplit over, joins: phedny, jk1joel
18:22-!-MrGandalf [] has joined #mythtv
18:23-!-noisymime [] has joined #mythtv
18:25-!-iamlindoro_ [n=iamlindo@] has quit []
18:25-!-Dibblah [] has left #mythtv ["Kopete 0.12.4 :"]
18:26<MrGandalf>has mythmusic been fixed yet?
18:33<MrGandalf>odd.. mythmusic works for me
18:33<MrGandalf>but I haven't recompiled plugins either
18:33<gbee>it compiles, but it the avfdecoder doesn't work - not a problem unless you have WMA or WAV in your collection
18:34<MrGandalf>I see
18:34<gbee>or unless you are using my patch to move mp3, flac and vorbis decoding to the avfdecoder
18:34<MrGandalf>I haven't patched or recompiled
18:35<MrGandalf>and since this is on 16384, I think I'm safe :)
18:40-!-foxbuntu [] has quit ["Leaving"]
18:49-!-CIDR [] has joined #mythtv
18:49<CIDR>Anyone know if linux supports UVD ?
18:53-!-Chutt [] has joined #mythtv
19:12-!-Dibblah [] has joined #mythtv
19:27<MrGandalf>gbee: seeing that same thing with it not playing certain things:
19:28<MrGandalf>2008-07-30 19:26:17.801 AFD Error: Could not find codec parameters. file was "/shares/recordings/LiveTV/1356_20080730192421.mpg".
19:28<MrGandalf>And that's a PVR250 channel
19:28<gbee> glad to know I'm not alone
19:31<iamlindoro>gbee: Heh, you'll be glad to know that the Mythtube thing takes advantage of your reflection effect :)
19:33<iamlindoro>Am actually going to pretty it up here a bit
19:42<gbee>can't wait until I get to really play with theming in mythui
20:01-!-Dibblah [] has quit [Read error: 113 (No route to host)]
20:08-!-_gunni_ [] has joined #mythtv
20:24-!-|gunni| [] has quit [Read error: 110 (Connection timed out)]
20:42-!-Shariff [] has joined #mythtv
20:42-!-Shariff [] has left #mythtv []
20:53-!-Winkie [] has quit [Remote closed the connection]
20:53-!-Winkie [] has joined #mythtv
20:53-!-foxbuntu [] has joined #mythtv
21:17-!-cattelan is now known as cattelan_away
21:17-!-cattelan_away is now known as cattelan
21:33-!-thatdood [] has joined #mythtv
21:43-!-huhlig [] has quit [Read error: 104 (Connection reset by peer)]
22:04-!-davilla [] has joined #mythtv
22:43-!-famicom [] has quit ["Leaving"]
23:02<iamlindoro>Can anyone confirm whether redirection works/is supposed to work properly in httpcomms.cpp?
23:17-!-robthebob [] has quit [Read error: 113 (No route to host)]
23:19<iamlindoro>Am looking at this MythTube enhancement, which works fine so long as youtube doesn't do a redirect, but once it does, it seems to munge up the redirect URL:
23:20<iamlindoro>I figure it's either being called incorrectly or something about redirect isn't working quite right. Obviously no pressure helping me on something not actually part of myth, but would like to be able to help him out getting it working all the time-- looks like it's working here until high traffic times and then it more or less stops working because it can't get to the redirects
23:21<iamlindoro>Also figured out how to get it to download the high quality h.264 in MP4 instead of the FLV he had it downloading, but that's sort of moot until I can get it to download... well... anything :)
23:22<Chutt>iamlindoro, you can get a .mp4 file from youtube?
23:22<iamlindoro>yeah, just takes a few extra chars at the end of the request URL
23:22<Chutt>what chars?
23:23<iamlindoro>IIRC &fmt=18 but let me double check
23:24<iamlindoro>yeah, that's it-- needs to be a properly constructed URL, but that's the end bit
23:24<iamlindoro>ie would be the normal FLV, would be the high qual MP4
23:25<iamlindoro>(where available)
23:26<Chutt>and it does produce a normal mp4, i wonder if it's hinted properly
23:27<Chutt>quality's still kinda low
23:27<iamlindoro>I don't think all have high qual versions
23:28<iamlindoro>I found it interesting in testing this earlier that they come down in 16:9 or even 2.35:1 when possible
23:28<Chutt>i just added youtube streaming (using rtsp) to our multimedia stack at work
23:28<Chutt>and that's 176x144 h263 with amr audio
23:28<iamlindoro>When it was working a few hours ago before dinnertime, I pulled in a Dark Night trailer and was shocked to find it in decent quality in correct aspect ratio
23:30<Chutt>it looks like it's properly hinted, even
23:34<iamlindoro>ah! I think I may have missed something silly
23:34<iamlindoro>int redirectCount = 0;
23:34<Chutt>the redirect stuff?
23:34<iamlindoro>that can't be good :)
23:34<iamlindoro>oh n/m
23:34<iamlindoro>that redirectcount, not maxredirects, sigh
23:36<Chutt>did you see this: Redirection to: '
23:36<Chutt>notice the doubled hostname
23:36<iamlindoro>yeah, the URL is all broken
23:38<iamlindoro>Plugin's no good if you can only use it 9-5 :)
23:40<Chutt>do a dump of resp.toString() in the redirection code
23:42<danielk22>Is David Blain still around? I've been Qt4'izing libmythupnp.. Still a few things left before I can even turn off qt3support, but it might be good to talk to him about some of the odd stuff in there.
23:43<Chutt>danielk22, yes, he's in here fairly often
23:43<Chutt>danielk22, CDev
23:44<danielk22>ah, ok, i've seen that nick. i'll keep an eye out -- & start collecting my thoughts :) thx
23:44<Chutt>he's mentioned the qt4 port recently
23:44<Chutt>was planning a more extensive rewrite due to the new networking code
23:45<Chutt>he was the one that started the overall qt4 port, btw =)
23:46<danielk22>What I'm doing is more of a programming in the small exercise, but the socket stuff is the big one remaining. I also need to grok the Q3Membuf_Ptr stuff.. never used that class.
23:46-!-moj0rising [] has joined #mythtv
23:47-!-moj0rising [] has left #mythtv ["Leaving"]
23:47<Chutt>i had started that lib
23:47<Chutt>and just copied over Q3MemBuf
23:47<Chutt>cuz it was tiny
23:47<danielk22>it still requires linking to the compat stuff though
23:47<Chutt>not when you rename it and port it to qt4 it doesn't
23:48<danielk22>i'd need to copy the whole thing and rename.
23:48<Chutt>iirc, it was just a header
23:48<iamlindoro>Got it
23:48<iamlindoro>This was dumb
23:48<Chutt>iamlindoro, what was it?
23:48<iamlindoro>cmdURL = QString("").arg(tArgString).arg(vidString);
23:49<danielk22>can't be just a header, I trying renaming and got linking errors on the ctor/dtor .. but it's probably small..
23:49<iamlindoro>it was trying to correct the to
23:49<iamlindoro>and mucking it up
23:49<iamlindoro>add the www. back in and it follows the redirects properly
23:50<Chutt>oh wait
23:50<Chutt>it's already there
23:51<danielk22>chutt: the copy? yeah -- or are you replying to iamlindoro
23:51<Chutt>the copy
23:51<Chutt>the .cpp is src/qt3support/other/q3membuf.cpp, btw
23:51<Chutt>it's < 200 loc
23:52<danielk22>173.. ok, i guess i'll just copy it for now then :)
23:53<danielk22>that just leaves one Q3Dict and the Q3SocketDevice stuff
23:53<Chutt>the Q3SocketDevice can be switched to MSocketDevice
23:53<Chutt>(from libmythdb)
23:53<Chutt>cuz that's just q3socketdevice renamed
23:54<danielk22>hehe, so I'll finish this up in the morning then! :)
23:54<Chutt>you should've said this earlier, btw
23:54<Chutt>i had that lib 3/4 ported :p
23:55<danielk22>wasn't too bad, I just got started on it because I had already done some of the work when qt4izing mythbackend
23:56<danielk22>i had also done some of it earlier when doing the libmythtv qt4izing, but threw that away at the time.. i just didn't want to do it a third time, so i figured i'd finish it up...
---Logclosed Thu Jul 31 00:00:44 2008