Back to Home / #mythtv / 2008 / 03 / Prev Day | Next Day
#mythtv IRC Logs for 2008-03-06

---Logopened Thu Mar 06 00:00:37 2008
00:03-!-aeha [] has joined #mythtv
00:04-!-carvajal [n=carvajal@] has joined #mythtv
00:13<xris>I'll check that one on my end
00:13<xris>mythvideo does give me a dump when it dies. just not sure how useful it is.
00:19<Anduin>gbee: re 4816 I spent some time looking at it, I'll try to get a patch in, maybe GreyFoxx of CDev can confirm but the ugly manual reference counting stuff in BackendSelect looks wrong, FillListBox does AddRef on the device, hands it to AddItem which creates a new ListBoxDevice (which does an AddRef) then doe a Release, and now back from AddItem FillListBox does another Release. When the ListBoxDevice gets destroyed it does anothe
00:20<Anduin>so another QMutex problem that is not related to QMutex, just related to being inside a deleted object.
00:20<Anduin>(which valgrind does confirm)
00:25<Anduin>This patch seems to fix it for me:
00:27<Anduin>I was also looking at the channel scanner, significantly less luck there.
00:29<Anduin>Hmm, still seems to be off
00:38-!-Agrajag- [] has quit [Read error: 110 (Connection timed out)]
00:59<Chutt>do we do something silly like scan files without a seektable?
01:10<Chutt>GreyFoxx, do you know where the seek to frame 0 comes from?
01:11<Chutt>ah, there it is
01:11<Chutt>GreyFoxx, it's causing startup to take a long time if there's no cutlist.
01:11<Chutt>at least for mpeg2-ts files
01:18-!-gnome42 [] has quit [Remote closed the connection]
01:28-!-PointyPumper [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
01:29-!-PointyPumper [i=Pintlezz@] has joined #mythtv
01:44-!-ToadP [] has quit []
01:58-!-rhpot1991 [] has joined #mythtv
01:59-!-rhpot1991 [] has left #mythtv []
02:15-!-loops [] has quit ["Leaving"]
02:17<Chutt>danielk22, janneg, gbee: Could someone take a look at #3773? I can fix the issue by deleting the 'av_read_frame_flush()' in avformatdecoder's OpenFile(), but I'm not sure if that's valid.
02:18<Chutt>you'll have to revert r16193 to get it to play via 'mythtv', though
02:36-!-Agrajag- [] has joined #mythtv
02:37-!-czth_ [n=dbrobins@nat/microsoft/x-ed76fd6821ebcf0a] has joined #mythtv
02:40-!-Dibblah [] has joined #mythtv
02:53-!-clever[rev] is now known as clever
02:53-!-czth [n=dbrobins@nat/microsoft/x-288c0f8629169846] has quit [Read error: 110 (Connection timed out)]
02:56-!-jhulst|phone [] has joined #mythtv
02:57-!-jhulst|phone [] has quit [Read error: 104 (Connection reset by peer)]
03:26-!-loops [] has joined #mythtv
03:30-!-cesman [n=cecil@] has joined #mythtv
03:59-!-loops [] has quit ["Leaving"]
04:00-!-mykeul [n=mykeul@] has joined #mythtv
04:10-!-loops [] has joined #mythtv
04:24-!-aeha [] has left #mythtv []
05:07-!-cecil [n=cecil@] has joined #mythtv
05:07-!-cesman [n=cecil@pdpc/supporter/sustaining/cesman] has quit [Read error: 104 (Connection reset by peer)]
05:15<janneg>Chutt: I'm not sure if the av_read_frame_flush() is valid
05:20<janneg>it might be needed for the DVD playback
05:21-!-Puhi [] has quit ["leaving"]
05:22<janneg>GreyFoxx: a simple "ringBuffer->Seek(0, SEEK_SET);" in avformatdecoder's OpenFile() does not resolve the not playing from start problem?
05:29<gbee>xris: the 100% cpu bug, are you running trunk or -fixes?
05:48<gbee>Anduin: that release is definately superfluous so I'd tend to agree that it's the cause, I placed a little too much faith in the backtrace where I expected pEntries to be null if the reference counting had cause the object to be deleted
05:54-!-MaverickTech [] has joined #mythtv
05:57<gbee>janneg: the av_read_frame_flush() is valid, at least something like it is required - scanstreams and one other function performed before that point decode a few frames at the beginning of the file moving our initial position forward, we have to flush those frames or seek back to start playback from the beginning of the file
05:58<gbee>the av_read_frame_flush() in OpenFile() predates the changes GreyFoxx recently made, but he did move it forward so that it happens after the ScanStreams call
06:00<gbee>I did experiment with rb->Seek but I can't remember what the results were, once GreyFoxx started making progress on it I left him to it
06:02<gbee>pretty sure Seek alone wasn't enough, that only resets the ringbuffers read pointer but av* maintains it's own sense of position within the file and so the two were out of sync (but I could be remembering wrong)
06:12-!-XChatMav [] has quit [Read error: 113 (No route to host)]
06:35-!-mattwire [] has joined #mythtv
06:40<janneg>gbee: I don't see where ScanStream consumes frames
06:45<gbee>janneg: I took GreyFoxx's word on that bit and moving the flush forward seemed to work for him
06:56-!-mattwire_ [] has joined #mythtv
07:01-!-MrGandalv [] has quit ["work"]
07:10<janneg>the problem is something else though, the timestamps for the video stream are complete off after the av_read_frame_flush()
07:12-!-ToadP [] has joined #mythtv
07:17-!-mattwire [] has quit [Read error: 110 (Connection timed out)]
07:17<gbee>janneg: ok
07:18-!-ahbritto [] has quit [Read error: 110 (Connection timed out)]
07:18-!-TelnetManta [] has quit ["Ex-Chat"]
07:19<gbee>do we just need to reset the dts and starttime as per the the change GreyFoxx made in [16325]?
07:29<janneg>no, the problem is that the dts starts at zero and not at the dts from the stream
07:42-!-dagar [] has quit [Read error: 110 (Connection timed out)]
07:43-!-dagar [] has joined #mythtv
07:45<laga>gbee: i was going to use ccache for bisecting, or is that a no-no when doing that?
07:46<gbee>I've never had an issue with ccache, but some people reckon it can occassionally do the wrong thing
07:46<gbee>I'd use ccache because the benefits out weight the risk
08:11-!-loops [] has quit ["Leaving"]
08:18-!-mattwire_ [] has quit [Remote closed the connection]
08:22-!-mattwire [] has joined #mythtv
08:46-!-gbee [] has left #mythtv ["Gone"]
08:48-!-gbee [] has joined #mythtv
08:48-!-loops [] has joined #mythtv
09:01-!-siXy [i=siXy@] has joined #mythtv
09:20<gbee> << Anyone else seen this? We get stuck and the only way to recover is to restart the frontend
09:20<laga>mythlcdserver just went postal here.
09:21<laga> <- repeat a hundred times. that happened after mythfrontend crashed (i'm bisecting right now)
09:23<laga>hum, it doesn't happen again. maybe it happened because DBHostName wasn't set in mysql.txt
09:25-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv
09:28-!-jgarvey [] has joined #mythtv
09:29-!-skamithi [] has joined #mythtv
09:30<laga>gbee: interesting. mythfrontend will only segfault during scaling when it's running fullscreen. if i use --geometry 800x600, it doesn't segfault during scaling but it'll usually segfault on exit. it just segfaulted on me when i exited a recording, too. so i wonder if the segfault during scaling just happens because scaling takes so long at 1680x1050
09:30<laga>gbee: should i get backtraces for those segfaults i'm seeing with --geometry 800x600, too?
09:31<gbee>it's a race condition I think, certainly explains why I only started seeing it recently
09:31<gbee>laga: no, unless they are different to the one you've already provided (SSDPCache)
09:32<laga>gbee: well, there's no way to know that w/o getting some backtraces :)
09:32<gbee>heh, yeah I just mean no need to attach them to the ticket :)
09:33<laga>before i start running gdb, i should probably remove the libmyth-* packages from my box.
09:33<gbee>laga: did you see the patch that Anduin created? Would be interested to know if it fixes the problem for you
09:34-!-TelnetManta [n=benwilli@] has joined #mythtv
09:34<gbee>try the patch first, could save you a lot of time
09:34<gbee>should have mentioned it earlier but I was distracted, sorry
09:34<laga>no problem.
09:35<laga>i was still using old libs anyways.
09:35-!-loops [] has quit ["Leaving"]
09:43<laga>gbee: ignoring the fact that mythlcdserver is flooding my logs again, anduin's patch seems to have improved things
09:48-!-loops [] has joined #mythtv
09:49<laga>now it's hanging when i exit a recording :/
09:56-!-cattelan [] has quit ["This computer has gone to sleep"]
09:56-!-loops [] has quit ["Leaving"]
09:58<laga>gbee: ok. no more segfaults which is probably good. without the patch, it'd segfault when i watched a recording and went back to the menu. now it's just hanging. how do i debug that? attach gdb and ctrl+c when it's hanging?
09:58<gbee>yeah, attach and break to find where it's stuck
09:59<gbee>ugh, fish is broken in KDE
10:02<laga>gbee: and
10:02<gbee>laga: thanks, I'll take a look in about half an hour, have to make some phone calls
10:03<laga>gotta run, too.
10:03-!-otwin [n=otwin@] has quit [Read error: 60 (Operation timed out)]
10:10<Cardoe>laga: you're on Gentoo no?
10:16-!-gnome42 [] has joined #mythtv
10:17-!-skamithi [] has quit ["WeeChat 0.2.3"]
10:18<laga>Cardoe: no, kubuntu
10:19<Cardoe>oh well nevermind. I was going to give you some tips to tweak your Gentoo configs for better backtraces.
10:22-!-gbee [] has left #mythtv ["Gone"]
10:27-!-skamithi [] has joined #mythtv
10:39-!-siXy [i=siXy@] has quit [Read error: 104 (Connection reset by peer)]
10:39-!-flithm [] has joined #mythtv
10:39-!-flithm [] has left #mythtv []
11:05-!-JoeBorn [] has joined #mythtv
11:06-!-ToadP [] has quit [Read error: 113 (No route to host)]
11:07-!-MrGandalf [] has joined #mythtv
11:09-!-johnp__ [] has joined #mythtv
11:25-!-gbee [] has joined #mythtv
11:31<Anduin>gbee: Even with my path I see a similar crash (sometimes) on exit now, I'll try to track it down later today.
11:31<Anduin>Is anyone actively working on the channel scanner as well?
11:31<gbee>don't think so
11:33-!-johnp__ [] has quit [Read error: 104 (Connection reset by peer)]
11:34<Anduin>Hmm, I'm sure it will get more eyes soon. I don't see the same crash but I do see a crash (looks like one of the icky deleteLater but happening too later ones), still haven't tracked it down though.
11:35<Chutt>apparently ddraw in wince5 is not binary compatible with ddraw in wince6.
11:35-!-johnp__ [] has joined #mythtv
11:35<Chutt>(no mythtv work until i figure that out) =)
11:36<Cardoe>silly Chutt expecting Microsoft not to break itself ;)
11:37<gbee>my brain has turned to goo this week, getting hardly any work done
11:43-!-mykeul [n=mykeul@] has left #mythtv []
11:45<GreyFoxx>gbee: I know the feeling. I've had a cold/fever for days now and my synapses are shutting down :)
11:57<Chutt>yay for really ugly hacks
12:00-!-gbee [] has quit ["Gone"]
12:04-!-gbee [] has joined #mythtv
12:09-!-dekar1 [] has joined #mythtv
12:17-!-mattwire [] has quit [Remote closed the connection]
12:25-!-dekarl [] has quit [Read error: 110 (Connection timed out)]
12:33-!-ahbritto [] has joined #mythtv
12:40-!-otwin [n=otwin@] has joined #mythtv
12:40-!-jwhite [] has quit [Remote closed the connection]
12:41-!-jwhite [] has joined #mythtv
12:41-!-beavis [] has joined #mythtv
12:52-!-mykeul [] has joined #mythtv
12:56-!-xris [] has quit []
13:02-!-xris [] has joined #mythtv
13:19-!-johnp__ [] has quit [Read error: 104 (Connection reset by peer)]
13:35-!-robthebob [] has joined #mythtv
13:46-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
13:55-!-Dibblah [] has left #mythtv ["Kopete 0.12.4 :"]
13:59-!-_gunni_ [] has joined #mythtv
14:00-!-renatofilho^ [n=renato@] has quit ["Ex-Chat"]
14:02-!-JoeBorn [] has joined #mythtv
14:15<gbee>stuarta: freesat still talking about DVB-S2, or are they sticking with DVB-S?
14:18-!-JoeyBorn [] has joined #mythtv
14:23-!-Dibblah [] has joined #mythtv
14:25<laga>gbee: have you had a change to look at the BTs?
14:27<gbee>laga: I did, can you open a ticket though? Mention what OSD themes etc you were using in the ticket
14:30<laga>a new ticket? is it a different issue?
14:31<gbee>looks different, QMutex again but this time it's the error in the log which interests me - "couldn't find base dialog"
14:31<gbee>that shouldn't ever happen
14:31<gbee>unless the theme is broken (which I'm assuming it's not)
14:31<laga>i probably have old themes somewhere.
14:34<gbee>it suggests the basedialog container is missing from the OSD theme, unless those themes are years old then I doubt that is actually true
14:34-!-renatofilho^ [n=renato@] has joined #mythtv
14:34<gbee>the requirement for basedialog goes back to revision [1230]
14:35<laga>hum, i've just ran "make uninstall" which removed /usr/local/share/mythtv
14:35<laga>and i've also removed anything left behind by packages. let's try this again..
14:36<laga>heh. nice.
14:37-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
14:37<laga>i dont have that particular OSD installed.
14:38<laga>gbee: i've reset the OSD to one that's actually installed and it's not hanging anymore
14:38*laga cries
14:42<laga>so it looks like anduin's patch might have been good enough for me.
14:46-!-GlemSom [] has joined #Mythtv
14:50-!-kormoc [n=kormoc@unaffiliated/kormoc] has joined #mythtv
14:51<gbee>laga: ok, cool, just glad the issue is sorted - one less thing to worry about for 0.21
15:10-!-richards [] has joined #mythtv
15:16<xris>anyone here use SD with more than one lineup? it looks like maybe the lineup-info grabber has a bug
15:22<xris>or something else in mythtv-setup
15:23<Anduin>I do
15:26<xris>go into mythtv-setup and look at the lineup source menu... on mine, it acts like one of the sources doesn't have a name.. pressing up to select things will go 1-1-2, and pressing down from there gets 2-2-1
15:26<Chutt>GreyFoxx, hello?
15:26<xris>so I guess it's more something in the display code
15:28-!-robthebob [] has quit [Read error: 104 (Connection reset by peer)]
15:28-!-robthebob [] has joined #mythtv
15:31<Anduin>xris: The Video Source menu? I don't see anything wrong there, or in the editor, even after a Retrieve lineups...
15:32<xris>it has all kinds of weird rendering issues for me
15:32<xris>but maybe it's VNC
15:32<xris>I'll test it out live later when I'm not too lazy to walk into the other room
15:33<Anduin>I have two lineups each of which has a faky subset lineup based on it and another dummy lineup, everything seems fine.
15:33<gbee>Anduin: that area doesn't use a translistbox does it?
15:33<gbee>translistsettingsbox or something like that
15:33<Anduin>gbee: Don't know, I'll check the code
15:34<gbee>there is an open ticket for odd scrolling behaviour with that widget
15:35<xris>that sounds similar to what I'm seeing
15:38<Anduin>gbee: It is a ListBoxSetting (what TransListBoxSetting derives from)
15:39<gbee>hmm, I dismissed the original report as a QListBox bug but I was thinking I'd look at it again because it seems more likely to be Myth's fault
15:47<Chutt>removing the av_read_frame_flush and the DoFastForward(0), i don't see a difference in startup frames
15:47-!-glemsom_ [] has joined #Mythtv
15:48<Chutt>GreyFoxx, i need a sample file =)
15:57-!-GlemSom [] has quit [Read error: 110 (Connection timed out)]
16:03<gbee>danielk22: any preference on a new default keybinding for the Signal monitor? Or should I leave it without a binding?
16:03<danielk22>i dunno, Alt-F7?
16:03<MrGandalf>is the sigmon available for DVB as well?
16:04<danielk22>I don't it's really only useful when your adjusting your antenna's, but that usually occurs before you've figured out how to set keybindings
16:04<danielk22>MrG: Yes
16:04<gbee>Alt-F7 it is then
16:05-!-skamithi [] has quit [Read error: 113 (No route to host)]
16:05<gbee>MrGandalf: but only if you have Interactive TV disabled (since the keybindings currently conflict)
16:05<Chutt>danielk22, you added the av_read_frame_flush to avformatdecoder.c (a long time ago) - happen to remember why?
16:05<MrGandalf>never tried bringing it up after a tune
16:05<danielk22>MrG: I've never tried it with DVB-S, and it doesn't have sound so it's not useful from another room (biggest problem IMHO)
16:06<danielk22>chutt: not off the top of my head, lemmy look at the code real quick..
16:06<MrGandalf>danielk22: no, but it's handy during a storm
16:06<Chutt>gbee, need to delete the SIGNALMON binding
16:07<Chutt>danielk22, to OpenFile()
16:07<gbee>Chutt: I thought about that, but figured that existing users can fix the conflict manually, the problem only affects UK/NZ users who have ITV enabled
16:08<Chutt>ah, ok
16:08<danielk22>about line 915?
16:08<danielk22>I think that was for a memory leak
16:08<Chutt>ScanStreams, AutoSelectTracks
16:09<Chutt>(between those)
16:09<danielk22>dunno what that is for
16:09-!-Dibblah [] has quit [Remote closed the connection]
16:09<Chutt>it's leaving tonight, then =)
16:09<danielk22>there might be a commit message to go along with it..
16:10<Chutt>it was just committed as "doesn't seem to cause any problems"
16:10<Chutt>one of mark spieth's old patches, i think, to do with timestamps
16:10<danielk22>bad commit message..
16:11<gbee>if danielk22 doesn't object I'm happy to delete the existing binding and force users to change their lirc configs, I just wanted to minimise the number of tickets/complaints we might get from users about Signal Monitoring no longer working in 0.21
16:11<danielk22>it must have been something that fixed allowed a particular file to play.. but if it's causing problems should be removed..
16:11<Chutt>it's causing another file to not play
16:11<Chutt>but ffplay doesn't have the flush there
16:12-!-Dibblah [] has joined #mythtv
16:12<Chutt>so i'm just inclined to remove it
16:12<danielk22>gbee: delete the existing binding. it was my mistake I checked in both F7 bindings.. Signal monitor is older, but it doesn't require as prominent a keybinding...
16:14<danielk22>chutt: I'm all for removing it... The original problem will pop-up again, but we can fix it properly when we have the two samples...
16:14<GreyFoxx>Chutt: I can give you a sample. I have 1 that starts almost 12 seconds in
16:14<Chutt>GreyFoxx, great
16:14-!-MavT [] has joined #mythtv
16:15<xris>gbee: I'm wondering if maybe the return of the 100% cpu bug has to do with zero-byte recordings..
16:16<gbee>xris: if it's the same bug, and I guess it is, then no
16:16<sphery>xris: I've never had any 0-byte recordings, but used to have the 100% cpu thing. Haven't had a chance to update since you said it returned, though.
16:16<gbee>I did wonder if it was somehow fixed by the MythEvent changes, but I can't see any connection other than it started working after those were committed and stopped after they were reverted
16:17-!-glemsom_ [] has quit [Remote closed the connection]
16:17<sphery>Is the extradata being used somewhere during that process?
16:18<gbee>sphery: not directly in any parts of the code I looked at, but it's possible that it is used somewhere
16:20<gbee>I've given up on that bug, I just don't know the code well enough or have the time to spend researching, needs a more experienced eye
16:21<gbee>I was hoping CDev might take a look since he wrote the code, but he's busy at the moment
16:21<sphery>gbee: As long as your doing a dbcheck to move the ITV Menu context bindings into TV Playback context, why not throw in a: DELETE FROM keybindings WHERE context = 'TV Playback' AND action = 'SIGNALMON' AND keylist = 'F7';
16:21<danielk22>extradata is only used in maybe a dozen places.. this is why the unthreadsafeness of it wasn't apparent earlier...
16:22<sphery>that way, if the user has fixed it, it continues working as before, but if it's the default, they get the new default...
16:22<gbee>I've attached a patch to that ticket which does fix the bug for me, but it's a hack and I can't take any credit, it was already in the code but commented with a health warning by CDev
16:22<sphery>and, we won't have the onslaught of, "Just press F7 to get the signal monitor," "Didn't work, I had to use Alt-F7" posts
16:24-!-TelnetManta [n=benwilli@] has quit ["Ex-Chat"]
16:25<gbee>sphery: tagging on keylist ='F7' is a good idea, should have thought of that but then I'm not thinking straight atm, I blame the cold/virus I've got atm + lack of sleep
16:26<sphery>gbee: Yeah. I was mainly seeing it from a mess of confusion on the lists standpoint... (I say as Disturbed's remake of Land of Confusion plays on the radio...)
16:28-!-ahbritto [] has quit [Client Quit]
16:28-!-MrGandalf [] has quit ["home"]
16:30-!-JoeyBorn [] has quit [Read error: 110 (Connection timed out)]
16:33-!-MaverickTech [] has quit [Read error: 113 (No route to host)]
16:34<richards>gbee: sphery: Evening, trust you are well. - The compile fail I am suffering occurs in both gcc 3.3.4 & 4.2.3.
16:36<gbee>I've no problems with GCC 4.2.2 or 4.3.*
16:36<janneg>richards: with unmodified sources? gcc 4.1.2 and 4.2.2 working here like a charm
16:36<gbee>or 4.2.1
16:38<richards>I do not suspect the compiler now. Someone else has this issue and the suggestion there is qt or stl (in the reply)
16:39<richards>There was a Trac bug raised which was closed with a comment of "uclic not supported"
16:40*sphery doesn't understand the desire to use uclibc on a box that processes multi-gigabyte video files...
16:40<richards>janneg: unmodified sources, yes
16:41<richards>sphery: neither do I.
16:41<Cardoe>sphery: if you NFS mount...
16:42<Cardoe>you could potentially have a nice embedded machine that is capable of processing the necessary data
16:42<sphery>net boot?
16:42<janneg>I still can't reproduce the backend deadlock during mass preview generation
16:51-!-rn114 [] has joined #mythtv
16:51-!-robthebob [] has quit [Read error: 104 (Connection reset by peer)]
17:00-!-xris [] has quit [Read error: 113 (No route to host)]
17:01<richards>gbee: sphery: janneg: What version of qt are you using?
17:02<janneg>same here
17:03<richards>I'm on 3.3.3
17:03<Anduin>richards: I could only see your problem happening with Qt without STL support (whey QValueListIterator doesn't have iterator_category)
17:03*richards insearch of 3.3.8
17:04-!-xris [] has joined #mythtv
17:05<sphery>richards: I have 3.3.7. If you use 3.3.8, make sure it's a patched version (if you're installing from a package is likely is) or all Myth programs will segfault on exit.
17:05<Cardoe>there's been some crashers fixed between 3.3.3 and 3.3.8
17:05<richards>Anduin: Thank you. Is STL support part of Qt?
17:05<richards>or a separate lib?
17:06<richards>sphery: Ta.
17:06<Anduin>richards: Qt can have STL "support", we do depend on it (QT_NO_STL would be the define to disable it)
17:08<richards>Thanks all, sounds like Qt 3.3.7 with STL enabled is the next move.
17:12<sphery>richards: If you want 3.3.8, you can get a patch. See (courtesy of gbee). Though I don't know how to get the patch from QT's tracker.
17:14<sphery>richards: Oh. A couple posts later and someone gives a link to the patch SUSE is applying ( )
17:14<richards>sphery: Ta. I was going with 3.3.7 to save hunting the patch. You have just changed the plot :-) 3.3.8 + patch it is.
17:14<gbee>patch isn't on their tracker, it was emailed to me direct - still have a copy somewhere
17:16<sphery>Hmmm. Differs from the SUSE one. Wonder if SUSE has some other stuff in theirs.
17:16-!-teprrr [i=tpr@kde/developer/rytilahti] has joined #mythtv
17:16<teprrr>hello, just wanted to note that the configure doesn't check whether Xxf86vm library is installed :p
17:18<richards>gbee: ta. patch captured.
17:19<gbee>sphery: the patch I posted is the official fix, suse one is their own attempt at a fix
17:19<sphery>gbee: "this is what I get for making last minute changes" -> "this is what I get for listening to sphery's suggestion" :)
17:20-!-mykeul [] has left #mythtv []
17:20<gbee>sphery: heh, nah, it was a silly mistake
17:21<sphery>but I was the reason you made the last-minute changes.
17:25<teprrr> here's a patch to check for it
17:25<teprrr>though I don't know if it's always required or not...
17:27<richards>sphery: There is a 3.3.8 & a 3.3.8b, wonder if the patch got rolled in.
17:29<sphery>teprrr: may want to post that patch to Trac ( ) so someone (probably janneg) can take a look at it.
17:30-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
17:30-!-MrGandalf [] has joined #mythtv
17:31<teprrr>sphery, as a ticket or something else?
17:31<teprrr>sphery, ah, sorry, found the guide
17:31<gbee>danielk22: any chance that the quartz and directx rendering options can be hidden by defines in the video profile settings? Prevents a user shooting themselves in the foot
17:33-!-Alowishus [] has joined #mythtv
17:33<danielk22>gbee: That was my initial plan, but it caused it's own problems. It's pretty much impossible to shoot yourself in the foot too badly, some render that works on your hardware will be picker regardless of how crazy you make the configuration.
17:34<danielk22>i do have some new profiles i plan to finish up tonight though, that are simpler than the existing ones.
17:35-!-ToadP [] has joined #mythtv
17:43-!-ToadP [] has quit [Read error: 104 (Connection reset by peer)]
17:44-!-ToadP [] has joined #mythtv
17:45<Cardoe>danielk22: as the saying goes, every time you design something to be idiot proof.. someone designs a better idiot.
17:50-!-Vaelys [i=awong@slammer.cs.Dal.Ca] has quit ["leaving"]
17:52-!-Vaelys [i=awong@slammer.cs.Dal.Ca] has joined #mythtv
17:59*laga wonders if that's a proof for not-so-intelligent design
17:59<MrGandalf>danielk22: since you said you've never tested it, sigmon with DVB(S) does bad things.
18:00<danielk22>MrG :(
18:00<danielk22>what does it do?
18:01-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
18:01<MrGandalf>half the time pauses playback, other times kills it altogether
18:01<MrGandalf>(an error occured while displaying....)
18:01<danielk22>hmmm, maybe we should disable it for DVB-S ...
18:02<MrGandalf>well, nobody is complaining
18:02<gbee>better it stays that way
18:02<MrGandalf>I'd leave it and maybe I'll take a look at it someday
18:02<danielk22>well I guess we can fix it later.. shame though
18:03<MrGandalf>if I get bored this weekend..
18:03<MrGandalf>which I'm sure I will
18:06-!-jgarvey [] has quit ["Leaving"]
18:08<MrGandalf>looks like the backend stops recording
18:09<CDev>gbee: which problem did you want me to look into? (was it the 100% cpu on preview generation?)
18:10<gbee>CDev: yeah, thought you might have an insight?
18:11<CDev>I can take a look, but I have never had the problem and the preview code has changed a lot since I wrote it.
18:11<MrGandalf>heh, all the PID filters get closed
18:13<gbee>it's not the preview code, seems to get stuck in a loop somewhere around BufferedSocketDevice::WaitForMore() and BufferedSocketDeviceRequest::ReadBlock()
18:13<CDev>ok, I'll try to find some time later tonight.
18:14<danielk22>MrG: the video is supposed to be completely stopped, so closing the PID filters would be ok.
18:14<MrGandalf>danielk22: looks like the recorder stops and TuningFrequency() is called for some reason.
18:14<gbee>it's a random problem that only seems to be exposed by preview image requests through mythweb because of the volume/load of those requests
18:14<xris>gbee: it was never volume/load for me. even happened if only a single image needed to be generated
18:14<MrGandalf>danielk22: ok, I guess knowing how it's supposed to work might help :)
18:14<danielk22>the recorder should stop, but I think TuningFrequency should only be called on channel changes..
18:15<danielk22>the video is stopped because some drivers go crazy when you query signal quality..
18:15<MrGandalf>ok, and once the sigmon is started, how is it stopped by the user?
18:15<gbee>xris: yeah it was possible that one single request would cause it, but lots of requests were guarenteed to do it
18:15<danielk22>hit Alt-F7 again
18:16<xris>gbee: ah
18:16<gbee>CDev: uncommenting the block in WaitForMore seemed to fix the issue for me, but caused other problems, like no machine names in autodiscovery
18:16<xris>maybe that's why I see it only sometimes now...
18:16<MrGandalf>ok, looks like it works about half the time (or more) then
18:16<danielk22>heh, still not so great
18:16<MrGandalf>danielk22: about the driver thing, is that still the case?
18:17<danielk22>depends on the driver, some have always been ok
18:17<MrGandalf>I've used 5 different DVB-S cards and none of them have had that issue.
18:17<danielk22>*shrug* I know if affected all the pcHDTV and FusionHDTV cards
18:17<MrGandalf>I see.
18:18<CDev>gbee: I'll try deleting all my preview images to see if I can re-create ir. I didn't like the sleep loop when I wrote it, and I still don't :-(
18:18<MrGandalf>and those you would think it would work on, considering they're OTA.
18:18<gbee>xris: any chance of attaching a backtrace to the ticket which might help cdev? I can't reproduce it so easily now and I don't think I've got any copies of the backtraces
18:18-!-beavis [] has quit [Remote closed the connection]
18:19<xris>I can test it later if you show me how to get a backtrace from something that doesn't die.
18:19<CDev>gbee: It's most likely something simple in the buffering that's wrong (Stole that code from a internal QT class)
18:19<Deek>#2996 -- because it's been transcoded, you dolt.
18:20<gbee>CDev: found one of the backtraces I pastebin'd (discovered they have a useful search feature)
18:20<CDev>gbee: Without looking at the code, did the spawning a new backend to generate a preview go in or get thrown out?
18:20<sphery>CDev: I was able to reproduce it without MythWeb by just doing preview generation requests (I used the script/hack at ).
18:21<sphery>CDev: it's in.
18:21<CDev>sphery: thanks.
18:21<gbee>it went in, but I've since amended the GetPreviewImage code to rescale an existing image if it exists - so in theory it should never trigger the preview generator and will produce faster results
18:22<gbee>if no image exists, i.e. it wasn't created by a backend or frontend then it's business as usual
18:22*CDev goes and eats diner with his family. will be back later.
18:23<gbee>one further change I want to make is that if you request an image bigger than the defaul then it calls the preview generator
18:23*xris waits in anticipation for his first 1080i Good Eats episode to record
18:23<sphery>heh... My hack (above) works, but I didn't change the help text from the original scriptk, so the help is useless...
18:23<gbee>but it was starting to get a little complicated and those changes didn't end up fixing the bugs
18:23<Deek>xris: bastard
18:26*sphery just realized that he can actually fix the comment as easily as telling CDev it's wrong
18:28<gbee>we can probably close #2996?
18:30<sphery>gbee: I can't think of any reason why XvMC would work for LiveTV but not for recordings on 0.20-fixes...
18:31<sphery>I'm guessing it is an invalid ticket (but could be written off as--"a lot has changed since 0.20--including playback profiles--so this probably no longer applies"
18:35<gbee>well I didn't mention playback profiles, but that's pretty much what I've done
18:37<gbee>how big a deal is fixing mythweather? I don't really want to release it in it's current state but I'm not sure how much I'll achieve in the time remaining
18:40<teprrr>hmm, is mysql5 supported? and ouch, why are database schema upgrades hardcoded inside the sourcecode? :)
18:40<sphery>teprrr: With 0.21-fixes or trunk, MySQL 5+ is mandatory
18:41<sphery>teprrr: And, we do /not/ support downgrading, so they work well in the source code.
18:41<teprrr>sphery, yup, I'm running trunk.. just thought if an upgrade error could be because of that debian uses mysql5 per default..
18:42<teprrr>but I'll check the query out and make a patch if necessary
18:42<sphery>which upgrade failed?
18:42*gbee senses impending doom
18:42<teprrr>sphery, 1212->1213
18:42<gbee>ahh feck
18:43<gbee>teprrr: which revision?
18:43-!-dagar [] has quit [Remote closed the connection]
18:44<sphery>Yeah. Make sure you have [16409]+
18:44<teprrr>Last Changed Rev: 16412
18:44<teprrr>hmm, oddly enough those queries pass just fine when doing manually...
18:45<teprrr>Query was: DELETE FROM keybindings WHERE action LIKE 'MENU%' AND context='TV Playback';DELETE FROM keybindings WHERE action='TEXTEXIT' AND context='TV Playback';DELETE FROM keybindings WHERE action='SIGNALMON' AND context='TV Playback' AND keylist='F7';UPDATE keybindings SET context='TV Playback' WHERE context='ITV Menu'; -- and error was: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for ...
18:45<teprrr>... the right syntax to use near ';DELETE FROM keybindings WHERE action='TEXTEXIT' AND context='TV Playback';DELET' at line 1
18:45<sphery>gbee: missing commas in the array...
18:46<teprrr>ahh, indeed :P
18:46*sphery thinks it will soon be "make sure you have at least r16414"
18:47<GreyFoxx>Chutt: Still want the example file that shows the playback not starting until several seconds in?
18:47<gbee>two lots of typos in one schema update, that neatly sums up my week so far
18:48<sphery>gbee: I think getting some rest/getting over your cold is more imporant than getting what you want done for the 0.21 release. :) There's always the liberal definition of a "fix." Besides, the 100% CPU thing may give you some extra time.
18:49<sphery>(that was real concern for you, not a jab at you)
18:50<gbee>heh, well I think I will go to bed
18:50<gbee>thanks for your concern ;)
18:51<sphery>get well
18:52<gbee>teprrr: should be fixed now, sorry for the screwup :)
18:52<gbee>better just install to check, though if it's still broken I don't really want to know ;)
18:54-!-dagar [] has joined #mythtv
18:55<gbee>"Couldn't upgrade database to new schema, exiting."
18:57<gbee>be nice if it explained why it couldn't upgrade
18:58<teprrr>gbee, nah, no problem and thanks :)
19:03<sphery>gbee: That's my code...
19:03-!-|gunni| [] has joined #mythtv
19:03<sphery>which did you run? mythbackend, mythtv-setup or mythfrontend? Looks like that was mythfrontend
19:04<gbee>frontend with -u
19:04<sphery>gbee: mythfrontend is no longer allowed to upgrade the DB
19:04<gbee>not at all?
19:04<sphery>that should work.
19:04<sphery>with -u it is.
19:05<sphery>In theory, if -u is specified, it will only give that message if !UpgradeTVDatabaseSchema()
19:05<gbee>backend will take forever to compile and I've got recordings in progress, could just run mythtv-setup on the frontend I guess
19:05<sphery>Did you get any errors fro UpgradeTVDatabaseSchema()
19:05<gbee>nothing to stderr anyway
19:06<gbee>running mythtv-setup on the frontend worked
19:06<sphery>Hmmm. I'll look into that.
19:07<sphery>It will take a bit to prep for the upgrade, though, so get some sleep. You've proved your code works, now the ball's in my court. :)
19:10<sphery>gbee: If you're still here, did it do the PromptForSchemaUpgrade() thing?
19:11<gbee>yeah, got the prompt, I selected upgrade and it just exited with the "couldn't upgrade" error
19:12-!-_gunni_ [] has quit [Read error: 104 (Connection reset by peer)]
19:13<sphery>ok, thx
19:15<CDev>sphery: how long do you need to let that script run before you see the problem?
19:19<sphery>When I was affected, it would usually run for about 100 previews before it got to 100% CPU (and I'd get no response at that point).
19:19<CDev>ok. I'll let it run for a while.
19:20<sphery>I'm running a rev right now that doesn't have the problem. xris said he saw the problem resurface on his r16394 (I have r16316)
19:21<sphery>To make it more likely, you might want to rm *.png in the recordings directory.
19:21<CDev>I'm running the test on my production box (it has the most recorded shows) and it's running 16407
19:21<sphery>if using MythWeb, also rm *.png in data/cache
19:22<sphery>Yeah. I never could get it to happen on my dev box (not enough recordings). Unfortunate since it's the only one with debug symbols...
19:23<CDev>I just want to see it happen, then I will add some debug code to track it down.
19:23-!-cecil is now known as cesman
19:28<CDev>I don't know, I had 4 processes requesting preview images at the same time after deleting all preview images (saw a total of 4 mythbackend's running at the same time.) and didn't have any problems. once all images were regenerated (282 of them) all was well again.
19:29<sphery>I can update mine to see if it's affected again, but I have recordings for the next 3 1/2 hours...
19:30<sphery>I can let you know what I find after that.
19:39<Deek>does AAC decode actually work in 0.21-fixes?
19:40<GreyFoxx>Deek: works fine for me with libfaac and libfaad support compiled in
19:41<Deek>The configure option was accepted, libfaad was linked into libmyth*, but the internal player says:
19:41<Deek>2008-03-06 19:22:09.341 AFD: codec AAC has 2 channels
19:41<Deek>2008-03-06 19:22:09.342 AFD Error: Could not find decoder for codec (AAC), ignoring.
19:41*CDev thought the preview issue might be a problem when recording, so I'm recording a HD channel from a HDHomeRun (using the same network card), and regenerating all images while requesting 6 recorded Programs pages concurrently, (had a total of 6 mythbackend instances running at once), and still no problem.
19:41<GreyFoxx>Deek: and libfaac as well?
19:43<Deek>making sure now
19:44*GreyFoxx needs to find a 10ft serial cable in his mess of cables
19:46-!-loops [] has joined #mythtv
19:49<sphery>Deek: in MythMusic or MythTV?
19:51<Deek>The internal player via mythnews (easiest source of h264/AAC files)
19:52<sphery>If MythMusic, there's this whole libmp4v2 and libmp4ff issue... But in TV, I don't know of any configure args for it...
19:57-!-rn114 [] has quit [Read error: 113 (No route to host)]
20:04-!-loops [] has quit ["Leaving"]
20:13-!-|gunni| [] has quit ["KVIrc 3.2.4 Anomalies"]
20:18<Chutt>what the hell's up with the SNR on mythtv-users lately
20:18<Chutt>it's way lower than usual
20:29-!-dagar [] has quit [Remote closed the connection]
20:30-!-dagar [] has joined #mythtv
20:34<Deek>GreyFoxx: yes, and libfaac as well.
20:41<Deek>AFD Error: Could not find decoder for codec (AAC), ignoring.
20:43-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Connection timed out]
20:50-!-dagar [] has quit [Remote closed the connection]
20:51-!-dagar [] has joined #mythtv
21:02-!-ahbritto [] has joined #mythtv
21:07-!-reynaldo [] has joined #mythtv
21:39-!-davilla [] has joined #mythtv
21:49<danielk22>Chutt: I think it's just a few threads sucking up all the air..
21:58-!-loops [] has joined #mythtv
22:03<xris>good thing I ignore that stuff most of the time
22:05<GreyFoxx>xris: See my reply earlier about the one I got that worked for me ?
22:06<xris>GreyFoxx: no
22:06<GreyFoxx>Good Eats, season 11 episode 14
22:06<GreyFoxx>played fine on the 3 frontends I tried it on
22:08<xris>will check it in a bit
22:11<xris>totally fails for me
22:11<GreyFoxx>I tried it here on 2 frontends and on my work one
22:11<GreyFoxx>all running svn as of yesterday
22:12<GreyFoxx>release branch of course
22:15<xris>I'm running trunk
22:15<xris>but that shouldn't matter
22:16<xris>ok, that's f'd up.
22:16<xris>it works if I use -v all
22:17<xris>or not
22:18<GreyFoxx>calling mythtv directly or mythfrontend?
22:19<xris>this works: mythfrontend -v all -O ThemePainter=Qt 2>&1 > h244_bt.txt
22:19<xris>this fails: mythfrontend -v all -O ThemePainter=Qt 2>&1
22:19<xris>I am extremely confused
22:20<xris>let me go try it in front of the actual box, rather than over vnc...
22:23<xris>ok, it always fails when I go through the frontend itself... but the 2>&1 doesn't work, presumably because mythvideo is a separate thread
22:25<GreyFoxx>try using mythtv /path/to/file ?
22:27-!-Alowishus [] has left #mythtv []
22:29<xris>just using mythtv call directly, on local file (to make sure there wasn't an issue with whitespace in the filename):
22:29-!-Thunder [] has quit ["Leaving"]
22:29<xris>must be something to do with the qt version?
22:38<xris>ok, managing channel icons is just way too much work.
22:38<xris>would be really nice if SD could allow use of the raw tms data for looking up icons.
22:38<xris>I suspect that it's something we'll have to add to the next contract
22:38<xris>GreyFoxx: so does that give you any helpful info?
22:43-!-kdub [n=kyle@] has joined #mythtv
23:00<kdub>hey i broke apache somehow, now http://localhost/mythweb is trying too access /mythweb.php/mythweb/
23:00<kdub>i think i need to reset the apache root directory or something
23:00<kdub>any idea on like, how
23:01<kdub>what happened to my message that i accidently joined the dev channel?
23:02<kormoc>kdub, you get that once every x months or so iirc
23:02<kdub>i see
23:02<kormoc>least, the automated one
23:02<kdub>thats the one i meant
23:03<xris>kdub: wrong channel
23:03<kdub>oh really?
23:04<xris>sorry, hadn't read far enough.
23:05-!-HeatHawk[LI] [] has joined #mythtv
23:07-!-HeatHawk[LI] [] has left #mythtv ["Ex-Chat"]
23:26-!-mzuverink [n=mzuverin@unaffiliated/mzuverink] has joined #mythtv
23:26-!-XChatMav [] has joined #mythtv
23:36-!-reynaldo_ [] has joined #mythtv
23:39-!-cattelan [] has joined #mythtv
23:41<dagar>will mythtv be involved in summer of code this year?
23:41-!-reynaldo1 [] has joined #mythtv
23:44<Chutt>dagar, not likely
23:44-!-MavT [] has quit [Read error: 110 (Connection timed out)]
23:48*xris keeps forgetting to write something up for mythweb...
23:48<xris>then again, I probably don't have enough time to monitor/mentor a student this summer
23:49<xris>esp. if SD moves forward with self-hosting plans.
23:50<dagar>ahh too bad
23:50<dagar>maybe I'll find something to do for free anyway
23:50<xris>dagar: we had a pretty bad time of it in 2006
23:50<dagar>how so?
23:50<xris>unproductive students
23:51<xris>no one really finished anything, none stuck around after the program ended
23:51<dagar>that's unfortunate
23:51<xris>I have 2 projects for mythweb that could go in, but it's a matter of having time to type them up, and then time to mentor a student through the process
23:53<dagar>I was thinking of starting to get involved this summer stipend or not
23:53-!-reynaldo [] has quit [Read error: 110 (Connection timed out)]
23:53<dagar>what are your mythweb projects?
23:54<dagar>I've been wanting to get some aspect ratio detection for the flash video player working
23:54<xris>rewriting the music portion from scratch, and enhancing the FLEX-based flv player for more speed/control, as well as music playback
23:54<xris>aspect ratio detection will go into the backend.. then it'll work fine
23:54<dagar>I thought the mp3act integration was a big improvement
23:55<xris>it was, but it's a pain in the neck to deal with because it's still basically a separate app
23:56<dagar>so I take it your 2006 students didn't get there stipends?
23:57-!-reynaldo_ [] has quit [Read error: 110 (Connection timed out)]
23:57<Chutt>one or two did
23:57<Chutt>out of 10
23:58<Chutt>shouldn't have had to babysit college kids.
23:59<dagar>is this a common experience with gsoc?
23:59<xris>no clue
23:59<xris>one of mine got paid (the other dropped out).. but I was too easy on him.
23:59<dagar>I know a few people that tried it last year but dropped it half way through
---Logclosed Fri Mar 07 00:00:14 2008