Back to Home / #mythtv / 2004 / 09 / Prev Day | Next Day
#mythtv IRC Logs for 2004-09-23

---Logopened Thu Sep 23 00:00:49 2004
00:07--- <<-- cmorgan [] has quit (Remote closed the connection)
00:10--- ---> uierh4 [] has joined #mythtv
00:10--- <--- uierh4 [] has left #mythtv ()
00:15Captain_Murdo| Chutt, fyi, I've got the sage comskip.c modified program compiled and running under Linux. Made a Makefile for it and setup a few #ifdef's for a few windows-specific things in the code.
00:15--- <<-- mdz [] has quit ("Client exiting")
00:32--- ---> uierh4 [] has joined #mythtv
00:32--- <--- uierh4 [] has left #mythtv ()
00:42--- ---> Icks [] has joined #mythtv
01:07--- ---> k-man_ [~jason@k-man.user] has joined #mythtv
01:07k-man_ hello
01:07--- <--- k-man_ [~jason@k-man.user] has left #mythtv ("Client exiting")
01:13Captain_Murdo| now I just need to make a wrapper script that will run the comskip executable against a bunch of the videos that have been flagged by the current code in Myth to see how much the two agree/disagree.
01:14ObsidianX hey Captain_Murdoch, you're a dev, so it would seem, and so may have some expertice in my problem
01:14Captain_Murdo| support or dev question?
01:15Captain_Murdo| not sure if you're going to ask about what you were talking about on the -users channel earlier.
01:15ObsidianX ah, you saw the strange error,
01:16Captain_Murdo| yeah. did you search the -users mailing list to see what others had done if/when they got that error?
01:17ObsidianX could you point me in the direction of the mailing list? :)
01:18Captain_Murdo| at the bottom is a link to the -users list and there's also a searchable archive at
01:20ObsidianX thanks :)
01:37--- ---> Netslayer [] has joined #mythtv
01:37--- <<-- Icks [] has quit (Read error: 104 (Connection reset by peer))
01:43Netslayer Chutt, I figured out which thread is causing all my problems I believe, or at least becoming a zombie. It's the commercialflag thread for sure. I'll look into it tomorrow when I have more time and post a reply on bugzilla
01:43Captain_Murdo| Chutt, that logo detection did seem to do pretty good on the show I just tested with. I haven't looked at his code yet but did run the program against a few recordings.
01:44Captain_Murdo| Netslayer, it's leaving zombies in gdb or just when running normally?
01:46Netslayer in GDB right when it starts, however I've been running the backend normally now (with commercialflag thread commented out) for 24hrs which is a record for .16 for me
01:46Netslayer in 15min when it finishes CSI NY I'll try to connect a few times with frontends, mythweb and know for sure ... i bet i'm having more then one problem.. who knows
01:47Captain_Murdo| is it the main thread from the backend or a thread that is fired off when a commercial flag job starts?
01:47Netslayer backend commercialflag.cpp
01:47ObsidianX hey does Myth support the Conexant cx88 chipset?
01:48Netslayer ObsidianX, I believe so, might even be what i have, check the users list
01:49Captain_Murdo| that thread isn't started detached. :( are you running current CVS?
01:50Netslayer Captain_Murdoch, what do you mean, I put a print out in commercialflag.cpp and watching it come alive as the backend started
01:50Netslayer then it died
01:50Netslayer i'm running CVS as of a day after .16 was released
01:50Captain_Murdo| yeah, it's supposed to die, but it's supposed to run detached so it doesn't go zombie when it finishes.
01:51Captain_Murdo| I didn't create it detached.
01:51Captain_Murdo| want to try a 4-line patch?
01:51Netslayer sure np
01:52Netslayer it fixed so many problems though when i uncommented it, for example i could never do a killall -9 mythbackend in .16, now with it not compiled in I can.. i'm awaiting to see if my connection problems are gone as well ""
01:53Netslayer I also have trouble running gdb mythcommflag :-P so I have no idea where the problem is
01:54Captain_Murdo| what exactly did you comment out?
01:54Captain_Murdo| here's a patch that will start that thread detatched.
01:54Netslayer //QSqlDatabase *flagdb = QSqlDatabase::database("FLAGDB");
01:54Netslayer //commflag = new CommercialFlagger(ismaster, flagdb);
01:55Netslayer line 610ish in main.cpp mythtv/programs/mythbackend/main.cpp
01:55Captain_Murdo| so you didn't comment out the thread, you commented out the whole flagger, listener and all. the thread thats going zombie is started inside the constructor in CommercialFlagger.
01:56Netslayer ok ic
01:56Netslayer how do I apply diff patches again?
01:57Captain_Murdo| that whole CommercialFlagger object is about to be replaced. I've got the replacement sitting in my CVS tree and have been running the new code for about 3 weeks now. a new common queue that will handle commercial flag jobs and user jobs and eventually transcoding.
01:57Netslayer i can do this manually, just a few lines
01:57Captain_Murdo| patch < commercialflag.diff
01:57Captain_Murdo| yeah, just add 3 and change 1
01:57Netslayer what's wrong with the current commercial skipping setup? haven't been following
01:58Captain_Murdo| the thread is started detatched so there's no need to join it later.
01:59Beirdo wonder if that's related to my devel box stopping commercial marking for no apparent reason...
01:59Beirdo Hmmm
01:59Beirdo it's running CVS updated today
01:59Captain_Murdo| it's not the skip/detect code, it's the way jobs are fired off. the new way is a whole lot more flexible. you'll be able to specify which jobs can run on which servers (ie, comm flag on maybe 2 our of 3 and transcoding on the 3rd).
01:59Netslayer Beirdo, that's what is happening to me
01:59Beirdo Oh, so it's just not getting kicked?
02:00Captain_Murdo| Beirdo, this isn't in cvs yet. I have been emailing Geoffrey (the mythtranscode guy) a little discussing a few things about it before I commit.
02:00Beirdo fair enough
02:00Beirdo I can keep mythcommflag manually for now. My production PVR seems to not have that problem (0.16)
02:00Beirdo at least not yet :)
02:02Captain_Murdo| with the code I have now, you can run any script you want after a recording finishes. there are 4 user-configurable jobs so you could for instance run nuvexport automatically on all CSI episodes if you wanted.
02:02--- ---> uierh4 [] has joined #mythtv
02:02--- ---> BlackHussar [] has joined #mythtv
02:03pigeon Captain_Murdoch: sounds neat
02:03Netslayer Captain_Murdoch, alright restarting mythbackend
02:03--- <--- uierh4 [] has left #mythtv ()
02:03Beirdo nice
02:04Beirdo I'm sure nuv2avi will be in there soon enough on some peoples' boxes
02:04--- <<-- Octane [] has quit (Read error: 110 (Connection timed out))
02:04Beirdo once I finish ironing out the stupidities
02:04Captain_Murdo| on one of the mailing lists, I described a simple perl "archive" script which moves recordings off to different directories and then links them back in. I have that working as well and will drop that in the contrib directory when I commit the other stuff.
02:06Netslayer Captain_Murdoch, i just tried a killall mythbackend, and -9 on it and it won't die :-/ (w/ commercial skip patch)
02:06Netslayer I have 4 backend threads left just sitting here
02:07Captain_Murdo| from the restart or from prior?
02:07Netslayer i killed the one before fine, started it manually (with your patch)
02:07Netslayer it's definetly the commercial skip thread doing this
02:08Netslayer it was started, I have a a printf in there showing me in my logs
02:08Captain_Murdo| are they sitting there or is it actually flagging shows? in the last gdb log you posted in that bug report, it is in FlagCommercials so it's trying to flag.
02:08Netslayer chris 4117 0.0 2.9 93512 15412 pts/1 RN 23:04 0:00 /usr/local/bin/mythbackend -v
02:08Netslayer not doing anything, and unkillable, only option is to reboot the machine
02:09Netslayer each has the same exact info, except pid #'s
02:09Captain_Murdo| all threads have the same info
02:09Netslayer chris 4110 0.0 2.9 93512 15412 pts/1 RN 23:04 0:00 /usr/local/bin/mythbackend -v
02:09Netslayer chris 4112 0.0 2.9 93512 15412 pts/1 RN 23:04 0:00 /usr/local/bin/mythbackend -v
02:09Netslayer chris 4114 0.0 2.9 93512 15412 pts/1 RN 23:04 0:00 /usr/local/bin/mythbackend -v
02:09Netslayer ...
02:10Captain_Murdo| did it print any messages in the mythbackend log about trying to flag shows?
02:10Netslayer nope
02:10Captain_Murdo| something like "Started Commercial Flagging for"
02:10Netslayer nope
02:10Netslayer i'll reboot the machine and use gdb, i'm sure it will zombie out
02:11Netslayer mythcommflag also just hangs when i run it with no output, there is something very odd going on
02:12Captain_Murdo| could be mmx problems since the commflag uses the decoder which uses mmx/sse/whatever.
02:12Captain_Murdo| is this a backend/frontend or just backend?
02:12Netslayer master backend and frontend
02:13Netslayer chris 2928 0.0 0.0 0 0 pts/0 Z 23:13 0:00 [mythbackend] <defunct>
02:13Netslayer gdb .. there's my zombie again
02:14Netslayer and the backend froze
02:14Captain_Murdo| the only way it could be firing off 4 threads is if it was told to flag 4 shows by receiving events either from the "restart unfinished jobs" thread or from another process. there's a lock around the code that starts the jobs so they can't all start at the same time they're started sequentially.
02:14Captain_Murdo| gdb will leave zombies whenever a thread exits.
02:14Netslayer there are about 4 shows that i remember having it say starting a few days ago that it hasn't work on yet
02:15Netslayer i'm running 2.4.23
02:16Netslayer it's worked fine up and till .16
02:17Captain_Murdo| 0.16 fired off jobs differently than 0.15.1. the main commercialflagger thread handled it rather than having the recorder threads firing off jobs.
02:17Netslayer so you think this is my kernel doing this?
02:17Netslayer hell i'll just reboot it in 2.6 and see
02:17Captain_Murdo| no, I'm just saying that when you run a threaded program under gdb, expect zombies.
02:18Netslayer i have a semi working 2.6 kernel
02:18Netslayer Captain_Murdoch, but ic a zombie process without GDB when the backend freezes by itself
02:18Netslayer not at startup normally however, after i start playing with it using mythweb
02:19Netslayer what do you recommend I do?
02:20Captain_Murdo| you could try commenting out that whole "if' statement that you added lines to in the constructor, that would disable any jobs being restarted when you start the backend. then the only jobs being fired off are the automatic ones if you have auto-flagging on or if you tell Myth to flag a show from the watch recordings screen.
02:21Netslayer ok sounds like a good idea one sec
02:22Netslayer so this whole if statement "if ((gContext->GetNumSetting("AutoCommercialFlag", 1)) &&
02:22Netslayer (isMaster))
02:22Netslayer {"
02:22Captain_Murdo| yeah. everything in there after the addListener line.
02:23Netslayer no more zombies!
02:23--- <<-- Chutt [] has quit (Read error: 110 (Connection timed out))
02:24Captain_Murdo| is this under gdb or just regular?
02:24Captain_Murdo| and are you running with -v
02:25Netslayer gdb and yes -v all
02:25Netslayer ekk i just asked it to flag a program, and i have 4 zombies
02:26Netslayer and a killall -9 leaves "chris 2924 0.0 0.0 0 0 pts/0 Z 23:23 0:00 [mythbackend] <defunct> chris 2952 0.0 3.5 119932 18432 pts/0 RN 23:24 0:00 /usr/local/bin/mythbackend -v all"
02:29Captain_Murdo| can you go around line 55 or so (look for "if (message.left(14) == "LOCAL_COMMFLAG")" and add a line above the "if" statement like this: cerr << "Message: '" << message << "'\n";
02:29Captain_Murdo| sounds like you're getting duplicate jobs fired off, but I'm curious what the messages being received look like.
02:29--- ---> Chutt [] has joined #mythtv
02:30Netslayer ok one sec
02:31Netslayer run in gdb?
02:31Captain_Murdo| Chutt, I have the sage comskip program compiled and running under Linux. the logo detection did pretty good at picking up the a transparent logo on one show I tested earlier tonight.
02:31Captain_Murdo| sure
02:32Netslayer Captain_Murdoch, should I manually ask it to comm flag?
02:33Captain_Murdo| yeah, that should only cause 1 Message line to be printed, but it sounds like it will print 4
02:33Netslayer Message: 'GLOBAL_COMMFLAG START 1002 2004-09-22T22:00:00 master'
02:33Netslayer Message: 'LOCAL_COMMFLAG START 1002 2004-09-22T22:00:00 master'
02:34Netslayer [New Thread 180236 (LWP 2964)]
02:34Netslayer no zombies yet
02:34Netslayer and no activity on the cpu at all
02:34Captain_Murdo| do you have a niced process running?
02:35Netslayer it just started recording the tonight show, i'm seeing 4 new threads, and 2004-09-22 23:35:02 Commercial Detection initialized: width = 720, height = 480, fps = 29.97, method = 1
02:35Netslayer i'll cancel the recording now
02:36Netslayer cancelled recording 4 Zombies appeared
02:37Captain_Murdo| that's normal since you're running under gdb. those recorder threads finished.
02:37Captain_Murdo| do you see any mythbackend threads that are niced?
02:37Netslayer i'm up to 6 zombies now
02:38Netslayer actually yes it's running
02:39Captain_Murdo| want to try stopping that and uncommenting the "if" statement you commented out earlier? I want to see what events are received there if possible.
02:39Netslayer just do a killall ?
02:40Captain_Murdo| does CTRL-C work?
02:40Netslayer nm ctrl c worked in mythbackend this time :-P
02:40Netslayer i mean gdb
02:40--- ---> marc_in_lux [] has joined #mythtv
02:41Netslayer Captain_Murdoch, leave your patch in or out?
02:41Captain_Murdo| in definitely. :)
02:42Captain_Murdo| without the patch it will probably always create 1 zombie thread even when not under gdb since i wasn't joining the thread once it finished.
02:43Netslayer ok 1 zombie, ekk i'm going to post all of this in a pastebin one sec
02:43Captain_Murdo| ok.
02:44Captain_Murdo| the zombie should be the thread that got fired off to check for unfinished flagging jobs.
02:45Netslayer it's currently comm flagging 5 things
02:46Captain_Murdo| yeah, looking at it now. can see the events that started them and the message from the flagging threads after they were started. with the new job queue I mentioned, it will allow you to specify a max number of jobs to run at one time on a per-backend basis.
02:47Netslayer cool, where do you set that? .. and why is this working now heh
02:47--- <<-- BlackHussar [] has quit (Client Quit)
02:48Captain_Murdo| it will be set in the backend setup program along with specifying how much cpu to use (so you can run jobs at low cpu on one backend and high cpu on another). hopefully I'll get that code in by Saturday, but if not it should be there Saturday afternoon when I have some free time.
02:48Netslayer wonder if I should let this finish comm flagging or hammer it with mythweb and frontend connections to see if my bug is still there
02:48Captain_Murdo| nothing to lose by hitting it.
02:49--- ---> DonLKSAB2 [~don@] has joined #mythtv
02:50Netslayer waiting on a recorded programs click "2004-09-22 23:50:05 NULL in string in list in WriteStringList"
02:51Netslayer only got back a few shows on all recordings
02:52Captain_Murdo| was that mythweb and can you still use the frontend?
02:52Netslayer :-P
02:52Netslayer i killed it
02:52Netslayer 2004-09-22 23:52:35 waiting for a thread..
02:52Netslayer clicked recorded programs .. then clicked backend status, went to connect with mythfrontend
02:53Netslayer I have results from "thread apply all bt full"
02:54Captain_Murdo| can you put that on the same (or another) paste?
02:54Netslayer interesting i think the output is repeating is that normal?
02:54Captain_Murdo| is your webserver running on the same machine?
02:55Netslayer yup
02:55Captain_Murdo| could look like it's repeating but it's really a bunch of similar threads since you were running 5 flagging jobs.
02:56Netslayer crap your right, let me post the full one
02:56Captain_Murdo| can you restart the backend, then make it break, then before you stop it, do a "netstat -an | grep 6543" to see how many TCP connections you have to the backend.
02:57Netslayer heh sure one sec
02:59Netslayer heh
03:00Netslayer i have 4 established and 2 time waits
03:00--- <<-- ObsidianX [] has quit (Read error: 104 (Connection reset by peer))
03:00Captain_Murdo| ok, nothing unusual. each mythfrontend will make 2 and I think mythweb does the same when it connects.
03:01Captain_Murdo| I was wondering if there were dozens.
03:02Netslayer what are the chances there is some weird value in my db that's causing this when it's cycling through each recording producing an output?
03:02Captain_Murdo| not sure how Chutt has all the socket thread pooling setup but was wondering if mythweb had too many connections still open.
03:02Netslayer so where do i go from here?
03:04Captain_Murdo| it's possible, I'm not sure about bad data messing this up. how many recordings do you have?
03:05Netslayer ~100
03:05Netslayer i also have a ton of bad sym links in my recording dir .. should del those
03:05Netslayer hmm would that cause a problem?
03:06Captain_Murdo| shouldn't, myth doesn't read the whole dir, it only looks at files it knows about.
03:06Netslayer actually i wouldn't mind deleting these i'll do that now, flashing red is annoying
03:07Captain_Murdo| you sure you have current cvs of mythweb right?
03:08Captain_Murdo| does the recorded programs page come up correct for the recordings it does show or are the ones it shows messed up?
03:09Netslayer Captain_Murdoch, i'm pretty sure, i'll check the dir date one sec, sometimes it will error out right at the beginning and only show the template, other times it just stops after a few rows, sometimes it shows all of them
03:10Netslayer sept 10th
03:11Captain_Murdo| can you break it only by playing with mythweb or can mythfrontend alone break it?
03:12Netslayer mythfrontends were doing the same thing before yes .. it's harder though i'm not sure what causes it. it was insane the last few days, sometimes it will just lock up, other times it will do the "Waiting for a thread..." .. could be related to the commercial skip though
03:14Captain_Murdo| I'm going to have to sleep on it and maybe I'll have some better suggestions after that. 3am here, nightly backups are starting and my system is getting as slow as my mind right now. :( I also should have that new replacement code in in the next few days so if it is somehow related to commercialflag.cpp, then that might cure it since that code is being replaced.
03:14Netslayer Captain_Murdoch, thanks so much for the help, i'll play with it in the next few days
03:14Captain_Murdo| ok. I'll keep an eye on that bug report in case I miss any conversations in here or emails on the list.
03:15Captain_Murdo| thanks for trying those things.
03:15Netslayer scary "lrwxrwxrwx 1 root root 10 Feb 14 2004 * -> /mnt/pvr/*
03:15Netslayer crap how do i delete that
03:16Netslayer rm \* ?
03:16Captain_Murdo| ls -al \* (then when that works and you're confident, use \* with rm)
03:16Captain_Murdo| yeah
03:17Captain_Murdo| you could also put it in single quotes usually.
03:17Netslayer cool got it
03:18--- ---> uierh4 [] has joined #mythtv
03:19Netslayer i'm posting the backtraces to the bug now
03:20--- ---> Icks [] has joined #mythtv
03:20Captain_Murdo| ok. not sure if they'll show anything but something might stand out to Chutt.
03:20--- <--- uierh4 [] has left #mythtv ()
03:23--- ---> phranck [] has joined #mythtv
03:26Netslayer hah just caused a broken pipe by clicking refresh a few times on the backend status
03:26--- <<-- phranck [] has quit (Client Quit)
03:27Captain_Murdo| almost sounds like you have too many things going on to be caused by something in Myth's code.
03:27Netslayer it's a dedicated machine, idle all the time, :-P dunno
03:28Captain_Murdo| s/going on/going wrong/
03:29Captain_Murdo| have you done a make clean recently?
03:30Netslayer nope
03:31Netslayer hrm it's breezing through recorded programs now :-P
03:31* Netslayer places bets on the * -> /mnt/pvr.. bad symlink
03:32--- ---> mdz [] has joined #mythtv
03:33Netslayer gnight, i'll hammer the crap out of this tomorrow and let you guys know
03:33Captain_Murdo| g'night.
03:46--- ---> Octane [] has joined #mythtv
03:55--- <<-- mdz [] has quit (Remote closed the connection)
03:56--- ---> mdz [] has joined #mythtv
03:58--- ---> KillerBunny [] has joined #mythtv
04:12--- ---> choenig [] has joined #mythtv
04:19--- <<-- KillerBunny [] has quit ("Leaving")
04:37--- <--- woodwizzle [] has left #mythtv ("Leaving")
04:48--- ---> KillerBunny [] has joined #mythtv
04:52--- ---> KillerBun [] has joined #mythtv
04:57--- <<-- KillerBun [] has quit (Client Quit)
05:00--- <<-- marc_in_lux [] has quit ("Leaving")
05:01--- ---> KillerBun [] has joined #mythtv
05:05--- <<-- KillerBun [] has quit (Client Quit)
05:05--- ---> KillerBun [] has joined #mythtv
05:09--- <<-- KillerBun [] has quit (Client Quit)
05:09--- ---> KillerBun [] has joined #mythtv
05:19--- <<-- KillerBunny [] has quit (Read error: 110 (Connection timed out))
05:57--- <<-- KillerBun [] has quit (Read error: 54 (Connection reset by peer))
05:57--- <<-- Icks [] has quit (Read error: 104 (Connection reset by peer))
06:01--- <<-- linuxwhore [] has quit (Connection timed out)
06:03--- <<-- choenig [] has quit (Remote closed the connection)
06:04--- ---> Geck0Fiend [] has joined #mythtv
06:16--- ---> linuxwhore [] has joined #mythtv
06:21--- <<-- GeckoFiend [] has quit (Read error: 110 (Connection timed out))
07:21--- ---> Icks [] has joined #mythtv
07:23--- <<-- jeffpc [] has quit ("Client exiting")
07:50--- ---> KillerBunny [] has joined #mythtv
07:50--- <<-- Icks [] has quit (Read error: 104 (Connection reset by peer))
07:56--- <<-- KillerBunny [] has quit ("Leaving")
08:32--- User: *** splAt1 is now known as splat1
08:41--- ---> dhofstra [] has joined #mythtv
09:07--- User: *** GreyFoxx_ is now known as GreyFoxx
09:27--- ---> Icks [] has joined #mythtv
09:43--- <<-- nrh1703 [] has quit (".<UPP>.")
09:52--- ---> streamtrade [] has joined #MythTV
10:10--- ---> o_cee [] has joined #mythtv
10:39--- <<-- DonLKSAB2 [~don@] has quit ()
10:43--- ---> mecraw_ [~lmarlow@] has joined #mythtv
11:07--- ---> adante_ [] has joined #mythtv
11:08--- <<-- adante [] has quit (Read error: 104 (Connection reset by peer))
11:08--- User: *** adante_ is now known as adante
11:09--- ---> schultmc [] has joined #mythtv
11:10--- ---> nrh1703 [] has joined #mythtv
11:17--- ---> choenig [] has joined #mythtv
11:32--- ---> Captain_Murdoch_ [] has joined #mythtv
11:46--- ---> lmatter [] has joined #mythtv
11:47--- ---> lmatter_ [] has joined #mythtv
11:47--- <<-- lmatter [] has quit (Remote closed the connection)
11:47--- <<-- lmatter_ [] has quit (Remote closed the connection)
11:56--- ---> dc_ [] has joined #mythtv
11:58--- <<-- streamtrade [] has quit ()
11:58--- ---> streamtrade [] has joined #MythTV
12:02--- ---> ajlill [] has joined #mythtv
12:04--- <<-- o_cee [] has quit (Read error: 110 (Connection timed out))
12:18--- ---> Baylink [] has joined #mythtv
12:19--- ---> KillerBunny [] has joined #mythtv
12:21ajlill is the "record new programs only" option working in CVS yet?
12:28--- ---> KillerBun [] has joined #mythtv
12:29--- <<-- KillerBun [] has quit (Read error: 104 (Connection reset by peer))
12:37--- <<-- mdz [] has quit (Read error: 60 (Operation timed out))
13:01--- <<-- Icks [] has quit (Read error: 104 (Connection reset by peer))
13:26--- ---> mdz [] has joined #mythtv
14:00--- <<-- dhofstra [] has quit ("ChatZilla 0.9.61 [Mozilla rv:1.7/20040616]")
14:09--- ---> jeffpc [] has joined #mythtv
14:38--- ---> Icks [] has joined #mythtv
14:43--- ---> jeffpc__ [] has joined #mythtv
14:45--- <<-- jeffpc [] has quit (Read error: 110 (Connection timed out))
14:57--- ---> dja_ [] has joined #mythtv
14:57--- <<-- dja [] has quit (Read error: 104 (Connection reset by peer))
15:09--- ---> Beirdo2 [~gjhurlbu@beirdo.user] has joined #mythtv
15:09--- <<-- Beirdo [~gjhurlbu@beirdo.user] has quit (Read error: 54 (Connection reset by peer))
15:17--- <<-- Baylink [] has quit ("User disconnected")
15:19--- ---> Baylink [] has joined #mythtv
15:23--- ---> Urgo [] has joined #mythtv
15:35--- ---> DonLKSAB2 [] has joined #mythtv
15:43--- <<-- DonLKSAB [] has quit (Read error: 60 (Operation timed out))
15:46--- ---> DonLKSAB [] has joined #mythtv
15:59--- <<-- streamtrade [] has quit (Read error: 110 (Connection timed out))
16:03--- <<-- DonLKSAB2 [] has quit (Read error: 110 (Connection timed out))
16:09--- <<-- mdz [] has quit ("Client exiting")
16:09--- ---> mdz [] has joined #mythtv
16:10--- <<-- BuzzBud [] has quit (Read error: 60 (Operation timed out))
16:23--- ---> BuzzBud [] has joined #mythtv
16:26--- <<-- dc_ [] has quit (Connection timed out)
16:33--- <<-- choenig [] has quit (Remote closed the connection)
16:34--- <<-- KillerBunny [] has quit ("Leaving")
16:45--- ---> CanadaBoy [] has joined #mythtv
17:16--- <<-- mdz [] has quit (Remote closed the connection)
17:17--- ---> choenig [] has joined #mythtv
17:20--- ---> mdz [] has joined #mythtv
17:29--- ---> tobo [] has joined #mythtv
17:30tobo anyone here using televinkens scripts to xmltv listings?
17:34Captain_Murdo| try #mythtv-users
17:37tobo tnx
17:44--- <<-- schultmc [] has quit ("Client exiting")
17:45--- <<-- choenig [] has quit (Remote closed the connection)
17:46--- ---> marc_in_lux [] has joined #mythtv
17:52--- <--- marc_in_lux [] has left #mythtv ("Leaving")
17:52--- ---> marc_in_lux [] has joined #mythtv
17:52--- ---> Roots^ [] has joined #mythtv
17:54splat1 nn all
17:54--- User: *** splat1 is now known as splAt1
17:56--- <<-- tobo [] has quit ("Leaving")
17:56--- User: *** Beirdo2 is now known as Beirdo
18:03--- ---> _GeckoFiend [] has joined #mythtv
18:10Netslayer hey Captain_Murdoch what's up? I have a feeling that removing that bad sym link and your commercial skip patch fixed 99% of my problems
18:13Rince_ *g*
18:13--- ---> snowblind [] has joined #mythtv
18:20--- <<-- marc_in_lux [] has quit (Read error: 104 (Connection reset by peer))
18:20--- ---> marc_in_lux [] has joined #mythtv
18:20--- <--- marc_in_lux [] has left #mythtv ("Leaving")
18:23--- <<-- Geck0Fiend [] has quit (Read error: 110 (Connection timed out))
18:24Captain_Murdo| Netslayer: the patch is in CVS now, so you might get errors when you try your next "cvs update" unless you typed it exactly like I had it.
18:24--- User: *** Netslayer is now known as tuxnewb2
18:24--- User: *** tuxnewb2 is now known as Netslayer
18:24--- User: *** _GeckoFiend is now known as GeckoFiend
18:24Netslayer Captain_Murdoch, ok i'll remember that
18:25Rince_ ;-)
18:25snowblind you aply your big patch yet Captain_Murdoch
18:26Captain_Murdo| snowblind: not yet. thinking about doing it tonight. My wife is out so I have time to commit and fix any bugs if anyone finds any right away. be nice to get that out of my local tree.
18:27Beirdo Ah, the wife's away, so the hubby will play.
18:29Captain_Murdo| yeah, getting ready to head home so I'll give it another once-over and get it in a little later tonight.
18:29Netslayer is that the logo detection stuff your talking about?
18:29Captain_Murdo| nice to be able to go to the status screen(s) and see what's flagging, what's done, etc..
18:30Netslayer wha .. where's that going to be? sounds cool
18:30Captain_Murdo| no, this is the rework of the way jobs are fired off. I made a common queue that can/will be used for comm flagging, transcoding (eventually), and user-jobs.
18:30snowblind Captain_Murdoch: mythweb need that staus info too, heh i dont even run a frontend atm
18:31Captain_Murdo| on the existing status screen in mythfrontend, I added an item at the bottom to allow you to scroll through all the pending and running jobs and see their info and status. the other place that info is visible is on the mythbackend status screen which is already accessible via mythweb so you can see it in mythweb also now.
18:31snowblind oh kickass
18:32Netslayer nice
18:32Beirdo sweeet
18:32Captain_Murdo| awsome!
18:32snowblind Captain_Murdoch: are you gonna do anything else after that?
18:33Captain_Murdo| to the queue or comm flagging?
18:33snowblind not that this aint enough just wondering
18:33--- ---> cmorgan [] has joined #mythtv
18:33snowblind oh user jobs isnt ready yet?
18:34Captain_Murdo| I want to add a way to pause/stop/resume jobs from the mythfrontend status screen and also want to add a way to make jobs prefer running on backends that aren't recording to help spread the load out more.
18:34snowblind Captain_Murdoch: you even gonna let us play with the commflaging settings? with out a recompile
18:34snowblind even = ever
18:34Captain_Murdo| user jobs is working, have 4 configurable jobs that can be setup to run automatically after new recordings. no way to fire them off otherwise right now though.
18:35snowblind hmm so i can put my script there?
18:35Captain_Murdo| commflag settings, haven't given it much thought. that could be on the backend setup screen possibly so it's not confusing to normal users.
18:36Beirdo isn't more of a profile thing, not a backend thing, thouh?
18:37Beirdo We need an "Advanced Settings" page maybe :)
18:37Beirdo heh
18:37snowblind wish it could be per channel settings
18:37Captain_Murdo| some guys that use SageTV took my commercial detect code and ported it so they could use it under windows and have been trying to tweak it. I got their code back-ported and running under Linux now. one guy rewrote a lot of the logo detection code and I was looking at that a little last night to see if I could get that back into Myth if it works better.
18:38snowblind maybe i could make commflag work better for some of the channels it messes up on
18:38Beirdo nice :)
18:38Captain_Murdo| yeah, but I don't want to end up with 3 pages of comm flag settings and it could easily be that many. havent' thought about it that much really.
18:38Beirdo well, something for you to mull over in your copious spare time :)
18:38Captain_Murdo| yeah, while I'm driving to/from work.
18:39Beirdo as long as the road rage comments stay in the code :) hehe
18:39Captain_Murdo| which I need to do the from part now otherwise I'll be here all night.
18:39Captain_Murdo| bbl
18:39--- <<-- Captain_Murdoch_ [] has quit ()
18:39Beirdo Seeya
18:39snowblind later
18:44snowblind hell i dont even care if you add GUI stuff to set the commflag stuff, i could put it in the DB myself.. but that alot of work for you and not enough poeple will beable to use it that way
18:57--- ---> m0j0__ [] has joined #mythtv
19:05--- <<-- m0j0 [] has quit (Read error: 60 (Operation timed out))
19:22Captain_Murdo| snowblind: I thought about adding some "undocumented" settings that way, but I don't know about per-channel stuff like that. I can see adding settings for things like "black level", "max commercial break length", etc..
19:24Captain_Murdo| re: the max length, I plan on adding code to set a max commercial break length anyway, but was thinking of making it tunable with a setting as described above.
19:24Netslayer crap the frontend just causing a "Waiting for a thread..." error and gdb didn't have any usefull output
19:27Captain_Murdo| try going into mainserver.cpp and upping the total number of threads to 10 or so. it's currently 5 and in a for loop around line 105 in mainserver.cpp
19:27Captain_Murdo| see if it still happens and just takes longer or if that solves it.
19:30Netslayer Captain_Murdoch, me?
19:30Netslayer it could have been caused by me polling mythweb a few times a few hours ago
19:32Beirdo funny, I pound the crap outta my mythweb, I don't think I've ever seen the issues you have
19:32Captain_Murdo| Netslayer: yeah, you. I wonder if something is holding them open forever or just holding them open a little too long.
19:41--- <<-- Captain_Murdoch [] has quit ("X-Chat!")
19:50Baylink Oh; it's *Murdoch* that's working on that job status stuff. Ok; that's a reason to update to CVS. :-)
19:50--- ---> Captain_Murdoch [] has joined #mythtv
19:51Baylink Sure, Cap; join back up *5 seconds* after I gurgle about your patches. :-)
19:52Captain_Murdo| had to reboot. I played that test HDTV clip that Jarod posted the other day and it hosed up XV on one of my displays. restarting X didn't work so I had to reboot.
19:54Netslayer i've never seen hd on my comp yet .. that would be cool to see
19:55Netslayer aren't there any HDTV capture cards that have dvi or component in?
19:55Captain_Murdo| it was nice.
19:55Captain_Murdo| don't know. haven't looked at hdtv much.
19:58Netslayer dang 330 bucks
19:58mikegrb lalala
19:58--- <<-- jeffpc_ [] has quit (Remote closed the connection)
20:03--- <<-- gnuyen [] has quit (Remote closed the connection)
20:08Netslayer what's with this "ran out of free AUDIO buffers :-(" message now, it's screwing up a few recordings every few days :-/
20:09Netslayer i swear i have every problem that can be imagined :-P for the last 6 months i've had an unstable box because of hardware problems i finally got under control
20:09Beirdo that sounds surprisingly like a user support issue :)
20:09Netslayer ya your right
20:10Netslayer but it could be interpretted as in what did you guys change in the ALSA native code that did this
20:15Netslayer great, just rebooted and the backend locked up again, couldn't kill it
20:21Netslayer took out the commercialflag instantiation again Captain_Murdoch and it works perfectly now :-/ i think at around midnight i'll upgrade to the latest CVS and go from there
20:24Captain_Murdo| ok.
20:27--- <<-- mecraw_ [~lmarlow@] has quit ("Trillian (")
20:30snowblind Captain_Murdoch: does it access the recordedmarkup table in that part?
20:30snowblind wonder if his DB is foobared
20:31Captain_Murdo| it's accesses it after flagging shows when it saves the markers I think he's having problems before that.
20:41--- <<-- snowblind [] has quit ()
20:56--- ---> wiz [] has joined #mythtv
20:56wiz Beirdo: I got myth working...sorta :)
20:57wiz it won't change channels, I hit down or up and it just flashes back to the same channel
20:57Beirdo you're on the wrong channel :)
20:57wiz hm?
20:57wiz ah mythusers sorry bud
20:57Beirdo :)
20:58--- <--- wiz [] has left #mythtv ("Leaving")
21:21--- <<-- ajlill [] has quit ("[x]chat")
21:29--- <<-- vagrant [] has quit (Remote closed the connection)
21:39--- ---> vagrant [] has joined #mythtv
22:02--- ---> CyberKnet [] has joined #mythtv
22:10--- <<-- jeffpc__ [] has quit (Read error: 110 (Connection timed out))
22:12--- ---> jeffpc__ [] has joined #mythtv
22:28--- <<-- Baylink [] has quit (Remote closed the connection)
22:45* Captain_Murd wonders if that commit log was a bit too long... :)
22:51Beirdo no such thing :)
22:51Captain_Murdo| have you seen it? :) I've been appending to it for weeks.
22:51Beirdo not yet, but I'll see it soon
22:51Beirdo my commits digest hasn't arrived yet
22:53mikegrb hehe
22:53Captain_Murdo| email just arrived here but I don't use digest mode.
22:53Captain_Murdo| might take 2 digests for this one.
22:54Beirdo hehe
22:57Captain_Murdo| be back in a few.... (gives people time to recompile) :)
22:58--- ---> cwall_64 [~cwall64@] has joined #mythtv
22:58--- <<-- cwall64 [~cwall64@] has quit (Read error: 104 (Connection reset by peer))
22:58--- User: *** cwall_64 is now known as cwall64
22:58Beirdo once I read the commit, I might be doing just that
23:08CyberKnet BTW... thanks for such a great program. =)
23:27--- <<-- Roots^ [] has quit (Read error: 104 (Connection reset by peer))
---Logclosed Fri Sep 24 00:00:11 2004