Back to Home / #mythtv / 2003 / 06 / Prev Day | Next Day
#mythtv IRC Logs for 2003-06-15

00:03-!-agar [] has quit ["Client exiting"]
00:05<yebyen>dude, x * x = 2x
00:05<yebyen>didn't you know?
00:44-!-Ripp is now known as wondars_printer
00:46-!-wondars_printer is now known as Ripp
01:12-!-bline [] has joined #mythtv
01:23-!-Timon [] has joined #mythtv
01:52<ahbritto>The backend has died a few times on me, is this common?
01:53<ahbritto>It seems to be related to seeking, any suggestions, or should I try to gdb it?
01:54<Chutt>follow the directions in the howto
01:54<ahbritto>You mean to gdb it?
01:54<extremis>whos been writing the dvb code for mythtv?
01:55<extremis>I just ordered a dvb card
01:55<extremis>it should be here next week
01:55<Chutt>oh great, you're going to want help, aren't you
01:56<Chutt>you do know that it's not going to work, right?
01:58<extremis>Chutt: whatcha mean?
01:59<extremis>from what I hear dvb works great with VDR
01:59<Chutt>in europe.
01:59<extremis>in the US too
01:59<extremis>dishnet is DVB in the US
02:00<Chutt>for free to air channels
02:00<extremis>you can get a CAM interface or use softcam in the US
02:01<extremis>so, chutt whyd o you believe it only works in EU?
02:02<Chutt>because maybe more people would be using it if they could?
02:03<extremis>I've read alot of postings saying that it works with the dish network in the US if you have softcam
02:03<extremis>I've seen mention of CAM interfaces that swupport nagravision
02:04<extremis>does ben hang out here?
02:04<Chutt>he doesn't have a dvb card
02:04<Chutt>he hasn't worked on the code in well over a month
02:05<extremis>is he in the US?
02:05<extremis>hrm, ok
02:06<extremis>what is his email?
02:06<Chutt>subscribe to the mailing list
02:06<extremis>doing so now...
02:15-!-yebyen is now known as YEBYEN
02:16-!-YEBYEN is now known as yebyen
02:20-!-Ripp [] has quit [""]
02:31<Timon>The new plugin style makes running the various modules MUCH MUCH faster!
02:32<Chutt>unless you're running mythgallery
02:32<Chutt>you're not seeing it
02:32<Timon>Thats the one I'm running :-)
02:32<Chutt>or mythweather as of like a minute ago
02:32<Timon>I really dig it. good job!
02:32-!-Drikus [] has joined #mythtv
02:33<Chutt>oh, btw -- QTimer's get auto-deleted, so you don't really have to delete them specifically
02:33<Chutt>that was your patch, right?
02:33<Timon>Ahhh, ok
02:33<Chutt>doesn't hurt anything to delete em manually
02:33<Chutt>but it's not necessary =)
02:34<Timon>I figured after I submitted it a second time that it somehow got automagicly cleaned up.
02:35<Timon>I like the new single window style as well, fixes focus problems for me!
02:37<Timon>In mythgallery, is there a way to rotate the images? I saw the angle column in the gallery.sql file; but didn't see anything in the README how what key to press to make that happen. Or is this a stub for functionally that never made it in?
02:38<Chutt>in the single-image view
02:38<Chutt>just hit [ or ] to rotate em
02:38<Chutt>should get saved to the db
02:39<Timon>Should that make it into the readme?
02:40<Chutt>probably should at some point
02:40<Timon>Ok :-)
02:40<ahbritto>This is so errie. The backend is recording and I can't even tell.... :)
02:41<Timon>Thats sweet, it even remembers the rotation when in multi-image view! Good show!
02:46<Timon>Got a segfault in mythfrontend - setup - General - goto the sliders for PCM and Master volume, change them, then press "["
02:48<Timon>Actually, it happens on any widget thats not a text input widget.
02:49<Chutt>ah yes
02:50<Timon>I would generate a bt, but no I always get the warning in gdb that no symbols exist (Even though I have compiled with debug, and verified that it contains symbols using nm).
02:50<Timon>You found the problem?
02:50<Chutt>naw, i know what it is
02:52-!-Drikus__ [] has joined #mythtv
02:54<Chutt>mind updating and trying it now?
02:55<Timon>will do
03:00<Timon>Fixed! Thanks!
03:01<Chutt>just fallout from the one window stuff
03:01<Timon>Ahhh, ok
03:03<Chutt>probably a couple other things similar to that
03:04<Timon>Don't worry, I'll find some way to break it :-)
03:05-!-Drikus [] has quit [Read error: 110 (Connection timed out)]
03:09<Timon>One thing that would be nice is for the slider widget to visibly change color when it has focus
03:15-!-Drikus_ [] has joined #mythtv
03:15-!-rkulagow__ [] has quit [Read error: 104 (Connection reset by peer)]
03:16-!-rkulagow__ [] has joined #mythtv
03:18-!-Drikus__ [] has quit [Read error: 110 (Connection timed out)]
03:30-!-Drikus___ [] has joined #mythtv
03:45-!-Viddy [] has quit [Remote closed the connection]
03:48-!-Drikus_ [] has quit [Read error: 110 (Connection timed out)]
03:48-!-Timon [] has quit [Remote closed the connection]
04:16-!-hfb [] has joined #mythtv
04:49-!-Drikus___ [] has quit ["toedeledoki"]
04:55-!-hfb [] has quit ["Client exiting"]
05:42-!-kiwi_uk [] has joined #mythtv
05:54<ahbritto>When is the default profile used as opposed to the transcode profile?
06:32-!-knight [] has joined #mythtv
06:33<knight>hey anyone awake that knows the keyboard keys for mythtv cvs?
06:34<knight>erm mythmusic cvs
06:34<knight>i cant figure out how to queue
06:35<knight>that is, add songs to the current playlist from the full list in the currently playing sectiojn
07:58<PeteCool>mdz: Add a Sync() to the ringbuffer, which syncs only the file being written,
07:58<PeteCool>and use that instead of a general sync(): does that make mythbackend write less often to disk?
08:01<PeteCool>Whoa Chutt you went on a bug-killing spree last night ;)
08:03<PeteCool>Chutt: I'd like to know, for my personal knowledge (If you have some time to explain), what were/are the technical advantages to move all the modules to plugins?
08:03<bline>I think the purpose was to move everything to one window.
08:14-!-bline [] has quit ["Client exiting"]
09:05-!-dopez [] has joined #mythtv
09:27-!-Ripp [~ripp@] has joined #mythtv
09:27-!-rkulagow__ [] has quit [Read error: 104 (Connection reset by peer)]
09:28-!-rkulagow__ [] has joined #mythtv
10:26<rkulagow>$ su
10:26<rkulagow># perl -MCPAN -e shell
10:26<rkulagow>cpan> install XML::Twig
10:26<rkulagow>cpan> install Date::Manip
10:26<rkulagow>$ su
10:26<rkulagow># perl -MCPAN -e shell
10:26<rkulagow>cpan> install XML::Twig
10:26<rkulagow>cpan> install Date::Manip
10:30<rkulagow>whoops, sorry about that last bit. pasted when i wasn't looking. =)
10:49<Chutt>can you load up an osd.xml file and tell me if it's not really obvious where the time format is set?
10:50<Chutt>i think that's kinda funny
12:01<grib>are there some problems with today's cvs, or are the makefile dependencies just incomplete? make after cvs up craps out... full rebuild takes so long I'm hesitant to make clean
12:45<extremis>I followed the gentoo instructions on the site for emerging mythtv but it still says :
12:45<extremis>ovitton local # emerge mythtv
12:45<extremis>Calculating dependencies
12:45<extremis>emerge: there are no masked or unmasked ebuilds to satisfy "mythtv".
12:45<extremis>!!! Error calculating dependencies. Please correct.
12:45<extremis>I already extracted it into /usr/local
12:45<extremis>now there is /usr/local/portage_local
13:05-!-dopez [] has quit [Remote closed the connection]
13:31<-- rkulagow__( has left #mythtv
13:34-!-Timon [] has joined #mythtv
13:56-!-paq_ [paq@] has joined #mythtv
14:17<-- ribo( has left #mythtv
14:25-!-Timon [] has quit [Remote closed the connection]
14:29-!-grib [] has quit [Remote closed the connection]
14:36<-- sc00p( has left #mythtv ("Client exiting")
14:43-!-Drikus___ [] has joined #mythtv
14:49-!-Ripp [] has quit ["Client exiting"]
14:52-!-kiweee [] has joined #mythtv
14:53-!-kiwi_uk [] has quit [Read error: 104 (Connection reset by peer)]
14:59-!-kiweee [] has quit [Remote closed the connection]
14:59-!-kiwi_uk [] has joined #mythtv
15:11-!-dopez [] has joined #mythtv
15:23<-- Sg4has quit ()
15:27-!-paq_ [] has quit [Read error: 113 (No route to host)]
15:30-!-dopez [] has quit [Remote closed the connection]
15:35-!-dopez [] has joined #mythtv
15:40-!-dopez [] has quit [Read error: 104 (Connection reset by peer)]
15:47<extremis>what kernel module does lirc install itself as
15:47<extremis>I followed the docs but I don't see any modules getting installed
15:55-!-extremis [] has quit []
15:55-!-dilinger [] has quit []
15:55-!-yebo [] has quit []
15:55-!-Chutt [] has quit []
16:02-!-extremis [] has joined #mythtv
16:02-!-dilinger [] has joined #mythtv
16:02-!-Chutt [] has joined #mythtv
16:02-!-yebo [] has joined #mythtv
16:04-!-dopez [] has joined #mythtv
16:05-!-Chutt [] has quit []
16:05-!-yebo [] has quit []
16:05-!-dilinger [] has quit []
16:05-!-extremis [] has quit []
16:08-!-extremis [] has joined #mythtv
16:08-!-dilinger [] has joined #mythtv
16:08-!-Chutt [] has joined #mythtv
16:08-!-yebo [] has joined #mythtv
16:08-!-Chutt [] has quit []
16:08-!-yebo [] has quit []
16:08-!-dilinger [] has quit []
16:08-!-extremis [] has quit []
16:08-!-dilinger [] has joined #mythtv
16:09-!-yebo [] has joined #mythtv
16:17-!-Viddy [] has joined #mythtv
17:04-!-bLuETeK [] has joined #mythtv
17:04-!-Drikus___ [] has quit ["toedeledoki"]
17:04-!-kiwi_uk [] has quit ["Peace and Protection 4.22"]
17:10-!-sc00p [] has joined #mythtv
17:12<bigguy>lo sc00p
17:26-!-bLuETeK [] has quit ["bye"]
17:28-!-activelow [] has joined #mythtv
17:37-!-activelow [] has quit [Remote closed the connection]
17:41-!-scanline [] has joined #mythtv
17:47<scanline>I'm using CVS, and it looks like the first recording mythbackend makes works fine, but every recording after that until I restart mythbackend comes out as an unplayable file beginning with a bunch of NUL bytes
17:50<sc00p>hey bigguy
17:51<bigguy>sc00p: you the sc00p of fm fame?
17:51<bigguy>or just an impersonator ;)
17:54<ahbritto>Do recording profiles work with WinTV PVR 250s?
17:56<bigguy>ahbritto: onlt the width/height thing afaik
17:57<ahbritto>bigguy: Thanks. Trial and error testing sucks.
18:08-!-extremis [] has joined #mythtv
18:08<extremis>Anyone alive?
18:09<extremis>I can't seem to get the lirc module to create the /dev/lirc file
18:09<extremis>I do have the correct entry in my modules.conf:alias char-major-61 lirc_serial
18:09<extremis>but I still get: irrecord: could not open /dev/lirc
18:10<extremis>ovitton root # lsmod | grep lirc
18:10<extremis>lirc_serial 7776 0 (unused)
18:10<extremis>could someone help a brothah out?
18:10<bigguy>sorry I haven't had the need for lirc
18:11<scanline>extremis: the lirc module doesn't create /dev/lirc unless you're using devfs
18:11<extremis>scanline: I am
18:11<scanline>does the file not exist, or are the permissions just wrong?
18:11<extremis>it doesn't exist
18:12<ahbritto>Is mpeg-4 encoding a real-time thing?
18:12<extremis>scanline: shoudl I see /dev mounted?
18:12<scanline>if you have a fast enough CPU
18:12-!-captain_proton [] has joined #mythtv
18:13<extremis>none on /dev type devfs (rw)
18:13<extremis>ahh, ther eit is
18:13<bigguy>ahbritto: you have a non-PVR card aswell?
18:13<scanline>extremis: yes, that's correct
18:13<extremis>so, scanline any ideas?
18:13<scanline>not really... when I loaded lirc_serial, /dev/lirc popped up without any trouble
18:13<extremis>what is in your modules.conf?
18:13<ahbritto>I only have a PVR-250 but I am trying to clean up the HOWTO and better understand MythTV.
18:13<scanline>ahbritto: I'm using an Athlon XP 2000+, and it takes around 50% of the CPU to encode MPEG-4 at 512x384, with the high quality options on
18:14<ahbritto>The question is, is recording buffered unencoded, and then encoded?
18:14<scanline>I'm sure there's a buffer in memory for that, but not on dis
18:14<scanline>er, disk
18:15<scanline>it's encoded in real-time
18:15<bigguy>ahbritto: encoding on the no pvr cards is realtime
18:15<bigguy>lag sucks
18:15-!-activelow [] has joined #mythtv
18:15<ahbritto>Thanks, there are so many basic things missing from the HOWTO. I am now adding a few sentances on what the program mythtv is...
18:16<extremis>scanline: grep -i lirc /etc/modules.conf for me please
18:16<scanline>extremis: I didn't put lirc in modules.conf, I put it in modules.autoload
18:17<activelow>anyone using mythtv to record on one box and viewing the mythfrontend on another to watch shows? I get extension "XFree86-DRI" missing on display but it's loaded in the Xfree logs, I've tried 16/24 depth resolution too..
18:17<extremis>so you didn't put: alias char-major-61 lirc_serial in your modules.conf?
18:17<scanline>extremis: correct
18:17<scanline>activelow: I've tried it... works alright, though Xinerama confuses the frontend a bit
18:18<ahbritto>How does mythtv choose the quality for pvr-250? Does it reset the card or use the card as is?
18:18<activelow>scanline: I just get a black screen when I try to watch a show or tv, starting mythfrontend (ssh'ed into the myth box and exporting a display) gives me Xlib: extension "XFree86-DRI" missing on display ":0.0, am I doing something wrong?
18:19<bigguy>ahbritto: use it as is unless you use the ivtv program that changes settings or hack the ivtv code
18:19<scanline>activelow: I'm not sure why it's looking for DRI, but that means it needs hardware acceleration on the local display
18:20<ahbritto>bigguy: thanks again!
18:20<extremis>scanline: it says in the howto that you have to add it
18:20<extremis>f the lirc configure program / compile did not mention anything about a kernel module, then you are finished. If it did mention a kernel module, you must edit the /etc/modules.conf file. Add this line as the first thing in the file. It must come first, or it may not work.
18:20<extremis>alias char-major-61 XXX
18:20<extremis>replace XXX with the name which you determined earlier, which in this example was "lirc_serial"
18:20* scanlinesighs
18:20<scanline>extremis: that's just one way of doing it
18:21<scanline>that will make lirc_serial load whenever a device with major number 61 is accessed
18:21<scanline>since I know I'll always need it, I just load lirc_serial on boot by putting it in modules.autoload
18:21<extremis>will that prevent the creation of /dev/lirc?
18:21<bigguy>ahbritto: no problem
18:21<activelow>scanline: that's what I thought, and I know DRI is working locally, I'd like to watch tv on another computer since I dont have a tv connected to the mythbox...
18:21<scanline>(modules.autoload is a gentoo thing... other distros have similar but differently named files)
18:21<captain_proton>/dev/lirc is created when the module is loaded
18:21<extremis>because I load lirc_serial in modules.autoload
18:21<ahbritto>To be clear, there is no way to stop a recording in progress execpt to delete it?
18:21<captain_proton>because thats how devfs works :P
18:21<captain_proton>without devfs it requires an mknod c 61 1
18:22<extremis>captain_proton: its not working that way for me
18:22<captain_proton>or 0 :)
18:23<scanline>Anyone in here that has access to the CVS repository interested in hooking it up to the CIA Bot project?
18:23<activelow>my only other issue is I'm not that sharp with mencoder, I patched it so it plays fine, anyone got a link/pointer to a test mencoder script that reencodes into say mpeg4/avi format?
18:29<extremis>damnit! lirc_serial is loaded but I have no /dev/lirc
18:30<captain_proton>extremis: why don't you just create the char device and be happy?
18:30<extremis>because I'm using devfs
18:30<captain_proton>it could be you build lirc against the wrong headers or config
18:30<extremis>I just emerged lirc again
18:30<extremis>it should have picked it up
18:31<captain_proton>if it works with a handmade device nod, then you can investigate that possibility
18:31<extremis>are there any requirements that I compile into the kernel for lirc to compile?
18:31<extremis>other than module support
18:31<extremis>and serial support
18:32<ahbritto>bigguy: Care to review my doc suggestions before I post them?
18:33<captain_proton>shouldn't lirc_serial -not- use the kernel serial support?
18:33<extremis>hrm, I compiled serial support as a module
18:33<captain_proton>extremis: have you released the uart?
18:33<extremis>serial support was compiled as a module and it is not loaded
18:33<extremis> x x <M> Standard/generic (8250/16550 and compatible UARTs) serial support x x
18:33<extremis>fro my make menuconfig
18:34<captain_proton>ls -a /dev/tts
18:35<extremis>ovitton linux # ls -al /dev/tts/1
18:35<extremis>crw------- 1 root root 4, 65 Dec 31 1969 /dev/tts/1
18:36<captain_proton>setserial -g /dev/tts/1
18:36<extremis>ovitton linux # setserial -g /dev/tts/1
18:36-!-hfb [] has joined #mythtv
18:36<extremis> /dev/tts/1, UART: 16550A, Port: 0x02f8, IRQ: 3
18:37<captain_proton>is that the serial port that your dongle is plugged into?
18:37<extremis> /dev/lirc doesn't get created if its not connected?
18:38<extremis>I can swap ports to see
18:38<extremis>I don't know what port is which
18:38<extremis>I just know I have 2 :)
18:38<captain_proton>for lirc_serial you need to have released the uart
18:38<captain_proton>which is quite clearly stated in the lirc documentation :P
18:38<extremis>how do I release the uart?
18:39<captain_proton>setserial <device> uart none
18:40<captain_proton>if the kernel serial driver has claimed the port, lirc_serial won't
18:41<extremis>well the kernel serial driver didn't load until I did setserial <device> uart none
18:41<extremis>so I rmmod'd it and then modprobe'd lirc_serial
18:41<extremis>still no /dev/lirc
18:43<ahbritto>Is this true: Profiles are used in this order: recording host name, Default, Live TV
18:45<ahbritto>The How to seems to have an error for encoding/decoding stats. Encoding is more CPU intensive?
18:45<scanline>captain_proton: the other option is to have the serial driver compiled as a module and not load it :)
18:45<extremis>I thought I was doing it right
18:45<scanline>ahbritto: encoding is way more CPU intensive
18:46<scanline>ahbritto: on my machine it takes maybe 2% of the CPU to decode, 50 or 60% to encode
18:46<ahbritto>scanline: thanks. The HOWTO has it reversed...
18:47<bigguy>reversed is relative
18:47<bigguy>it might be right if talking about the 250
18:47<ahbritto>bigguy: you're right, I miss read it! thanks!
18:54<extremis>Your hardware does not require a kernel module.
18:54<extremis>hrm, I'm getting that on compile of lirc
18:55<captain_proton>extremis: do you have a mode2 dongle or one of the pctv ones?
18:57<extremis>I have 2, one is the one that came with my pvr250 and I also have a logitech one that I'm trying to setup as a blaster for my echostar reciever
18:57<extremis>the logitech ir is serial
19:02<ahbritto>scanline & bigguy: many HOWTO improvement suggestions e-mailed. Thank you for your help.
19:02<scanline>extremis: the lirc_serial module is only for the homemade-style serial transmitters and receivers
19:06<extremis>what about lirc_sir?
19:08<extremis>hrm, ok well I'll working lirc_i2c for my hauppauge
19:08<extremis>seems I have unresolved symbols with this one
19:08<extremis> /lib/modules/2.4.20-gentoo-r5/misc/lirc_i2c.o: unresolved symbol i2c_attach_client_R2e31ca93
19:09<bigguy>ahbritto: emailed to whom?
19:10<bigguy>he is often here
19:10<bigguy>epic4 1.1.11 (Vacuous) + hienoa 0.51
19:29<ahbritto>If I wanted to add bit rate support for PVR-250 which source files should I look at?
19:36<bigguy>chutt doesn't want it added yet
19:36<bigguy>as the mechanism for changing it is still not set in stone
19:37<ahbritto>bigguy: you use pvr-250?
19:38<bigguy>I don't have the $$
19:38<bigguy>I just keep up with the lists
19:38<ahbritto> :(
19:38<ahbritto>bigguy: I got mine 1/3rd off... :)
19:39-!-bline [] has joined #mythtv
19:39<bigguy>I couldn't afford it if it was $50 atm
19:39<ahbritto> :(
19:40<ahbritto>bline: Could you try a mythtv test for me? I want to confirm if a play back problem is just me?
19:40<bline>Not really.
19:41<ahbritto>ok. thanks anyway.
19:41<bline>One, I'm not at home. Two, Mythtv's sound is busted for me atm.
19:41<ahbritto>Ah, I notice that low bit rate file stutter audio wise with mythtv but not xine or mplayer...
19:42<bline>I don't have a stutter, sound cuts out every 5 to 10 seconds and doesn't come back unless I rewind for a second.
19:43<bline>I'm having other strange problems with that system so I think it may be hardware related. Possibly cpu.
19:44<ahbritto>try play back with xine
19:44<bline>Yeah, xine and mplayer work fine, but it may be just the _way_ they do things does not introduce my hardware bug.
19:45-!-hfb [] has quit ["Client exiting"]
19:45<ahbritto>Try high bit rate recordings, the problem I am seeing seems to be bit rate related.
19:46<bline>I'm using the same bitrate Chutt is.
19:46<bline>But I will play with it when I get home, just in case.
20:03-!-Chutt [] has joined #mythtv
20:03<bigguy>lo Chutt
20:30<extremis>hrm, I can't get audio, yet alsa is loaded
20:30<extremis>I increased the volume on everything
20:30<extremis>yet still can't hear wav's or mp3's
20:30<extremis>I'm not familiar with alsa
20:30<extremis>other than its mixer (now)
20:40-!-bigguy [] has quit [Read error: 113 (No route to host)]
20:44<extremis>I guess no one knows
20:44<ahbritto>extremis: maybe you should find an alsa channel
20:47-!-bigguy [] has joined #mythtv
20:48-!-PeteCool [] has quit ["Client exiting"]
20:59-!-rkulagow__ [] has joined #mythtv
20:59<rkulagow>$ su
20:59<rkulagow># perl -MCPAN -e shell
20:59<rkulagow>cpan> install XML::Twig
20:59<rkulagow>cpan> install Date::Manip
20:59<rkulagow__>disregard that. (again.) i need to clear out my clipboard.
21:00<rkulagow__>chutt: are you here?
21:01<ahbritto>rkulagow: what client are you using that lets you paste so insecurely?
21:02<ahbritto>rkulagow__: ...
21:16<rkulagow__>sorry, watching tv with the family now, but i got a segfault in cvs mythfrontend "will not delete top most widget, error" or something to that effect. not running in gdb (backend was, nothing happened there). re-running in gdb now, i'll see if i can recreate. happened during a delete from the "you are done watching" window.
21:16<rkulagow__>bbl, hopefully with a decent error and bt.
21:32<Chutt>ah, ok
21:39* scanlinewaves to Chutt
22:12-!-Ripp [~ripp@] has joined #mythtv
22:30<bline>yay, the alsa patch is in
22:30<Chutt>there's currently no way to enable it
22:31<bline>ok, then I await that patch as well :)
22:42<mdz_>this via sound driver sucks
22:42-!-dopez [] has quit [Read error: 104 (Connection reset by peer)]
22:43<mdz_>running a second frontend on my desktop with the built-in audio, I get audio buffer overflows
22:43<mdz_>just doing playback
22:47<mdz_>yeah, but I'm using the ALSA driver
22:47<mdz_>it was worse with OSS
22:47<Chutt>i'm not having problems with that
22:47<Chutt>at least with a current driver
22:47<mdz_>this recording seems to be messed up
22:47<Chutt>on my epia-m
22:48<mdz_>at one point, the sync goes bad and there are lots of drops
22:48<mdz_>maybe a big sync() :-)
22:48<Chutt>quite possibly
22:49<mdz_>usually when that happens it clears up
22:49<Chutt>this doesn't?
22:49<Chutt>maybe cron kicked in? =)
22:52<mdz_>could very well be
22:52<rkulagow__>chutt: caught that segfault.
22:52<mdz_>movie started at 5:10am
22:52<Chutt>rkulagow, if it's indeed that error, the segfault's not going to help
22:52<Chutt>how do you reproduce it, though?
22:52<mdz_>seek table seems to be missing some entries
22:52<mdz_>it gets pretty confused over that
22:53<mdz_>I don't know why they would be missing, though
22:53<Chutt>dropped frames, perhaps
22:53<scanline>Chutt: I'm having a bit of a problem with MPEG-4 encoding on MythTV.. semi-randomly it seems that recordings fail- the beginning of the file becomes a bunch of NUL bytes. Any ideas?
22:53<scanline>Chutt: I thought it could be that the mount it's recording to is on NFS, I'm testing that theory now
22:53<Chutt>never seen a recording fail like that.
22:53<rkulagow__>really? that sucks. watched a recording, then press ESC. get the save position, no position, keep watching, delete. do delete, return to the watch recording with the delete popup. hit yes, delete. HD light flashes, then get the segfault.
22:54<rkulagow__>i'll paste in the BT to see if it means anything to you:
22:54<rkulagow__>(gdb) thread apply all bt full
22:54<rkulagow__>Thread 2 (Thread 32769 (LWP 3779)):
22:54<rkulagow__>#0 0x40b5a127 in poll () from /lib/i686/
22:54<rkulagow__>No symbol table info available.
22:54<rkulagow__>#1 0x4095593e in __pthread_manager () from /lib/i686/
22:54<rkulagow__>No symbol table info available.
22:54<rkulagow__>#2 0x40955b8b in __pthread_manager_event () from /lib/i686/
22:54<rkulagow__>No symbol table info available.
22:54<rkulagow__>Thread 1 (Thread 16384 (LWP 3611)):
22:54<rkulagow__>#0 0x08480de8 in ?? ()
22:54<rkulagow__>No symbol table info available.
22:54<rkulagow__>#1 0x4010e7a6 in ~MythDialog (this=0x8480ee0) at mythdialogs.cpp:121
22:55<rkulagow__>No locals.
22:55<rkulagow__>#2 0x40112fb4 in ~MythPopupBox (this=0x8480ee0) at iostream:79
22:55<rkulagow__>No locals.
22:55<rkulagow__>#3 0x0807f6a0 in PlaybackBox::doDelete() (this=0xbfffe980)
22:55<rkulagow__> at playbackbox.cpp:1514
22:55<rkulagow__>No locals.
22:55<rkulagow__>#4 0x080b9f38 in PlaybackBox::qt_invoke(int, QUObject*) (this=0xbfffe980,
22:55<rkulagow__> _id=61, _o=0xbfffe260) at moc_playbackbox.cpp:138
22:55<rkulagow__>No locals.
22:55<rkulagow__>#5 0x4053bd7b in QObject::activate_signal(QConnectionList*, QUObject*) ()
22:55<rkulagow__>---Type <return> to continue, or q <return> to quit---
22:55<rkulagow__> from /usr/lib/qt3/lib/
22:55<rkulagow__>No symbol table info available.
22:55<rkulagow__>[New Thread 81926 (LWP 3784)]
22:55<rkulagow__>Changing from WatchingPreRecorded to None
22:55<rkulagow__>Changing from None to None
22:55<mdz_>this works so much better on the mailing list
22:55<rkulagow__>Not removing top most widget, error
22:55<rkulagow__>Program received signal SIGSEGV, Segmentation fault.
22:55<rkulagow__>[Switching to Thread 16384 (LWP 3611)]
22:55<rkulagow__>0x08480de8 in ?? ()
22:55<rkulagow__>Current language: auto; currently c
22:55<rkulagow__>(gdb) thread apply all bt full
22:55<rkulagow__>mdz: do you mean the bt i just posted, or something else?
22:55<Chutt>which thread did it die in?
22:56<mdz_>rkulagow__: yeah :-(
22:56<rkulagow__>Changing from WatchingPreRecorded to None
22:56<rkulagow__>Changing from None to None
22:56<rkulagow__>Not removing top most widget, error
22:56<rkulagow__>Program received signal SIGSEGV, Segmentation fault.
22:56<rkulagow__>[Switching to Thread 16384 (LWP 3611)]
22:56<rkulagow__>0x08480de8 in ?? ()
22:56<rkulagow__>Current language: auto; currently c
22:56<mdz_>it's easier to read, and then the info is in the list archive so it's easy to look up
22:56<Chutt>i believed you, didn't need to paste it :p
22:57<rkulagow__>(just wanted to make sure that i pasted the relevant bits).
22:57<mdz_>I think I may have seen that crash
22:57<Chutt>it's brand new
22:57<Chutt>as of yesterday
22:57<rkulagow__>it's happened to me now 3 for 3 when i try to delete.
22:57<Chutt>i'll get it fixed in a moment
22:57<mdz_>a crash was reported to me when deleting a recording
22:57<mdz_>after the popup
22:57<rkulagow__>chutt: ok, thanks.
22:57<mdz_>not consistent, though
22:58<bigguy>bruce just pasted the same thing
22:58<rkulagow__>chutt: is there anything else you need from this debug session?
22:58<Chutt>rkulagow, nope =)
22:59<rkulagow__>ok, i'll whack it. hope it helped.
22:59<Chutt>gimme 5 minutes
22:59<rkulagow__>chutt: NP. thanks.
23:01<ahbritto>rkulagow__: I e-mailed you some suggestions for the HOWTO, reply if you want more of the same...
23:01<scanline>Chutt: what would you think about adding MythTV to the CIA Bot project? It's an IRC bot that reports commits in real-time from a bunch of projects
23:02<Chutt>rather not
23:02<bigguy>heh the cia bot got banned from #e cause they got tired of it
23:04<Chutt>rkulagow, can you update and try to reproduce?
23:05<captain_proton>seems silly to ban it when they could just tell it to part the channel
23:07<Chutt>if that doesn't get it, i can change something else
23:08<bline>raster is a silly person..
23:13<extremis>Does anyone have an XF86Config for a geforce3 ti 200?
23:14<extremis>what is the max resulution for svideo btw?
23:15<Chutt>800x600 is generally the max you'll want to do, but really, anything more than 640x480 isn't doing anything
23:19<mdz_>I wonder why this cache flush thing still sucks
23:19<mdz_>you'd think it'd have been fixed with all the people doing capture on linux
23:20<mdz_> says to run a shell script with a sleep 1/sync loop :-)
23:21<ahbritto>mdz_: Doesn't journaling for periodic buffer writes?
23:21<mdz_>ahbritto: parse error
23:21<mdz_>oh, did you mean force periodic buffer writes?
23:22<mdz_>journal writes just bypass the cache, I believe
23:22<poptix>i don't see how people can be having disk IO issues
23:23<poptix>i'm using a P233MMX over NFS for storage, with no problems
23:23<ahbritto>mdz_: What kernel are you using?
23:23<poptix>and that's with a PVR-250, which has considerably higher bandwidth requirements
23:23<mdz_>ahbritto: 2.4.19
23:23-!-Captain_Murdoch [] has quit [Read error: 110 (Connection timed out)]
23:24<mdz_>poptix: which version of myth?
23:24<ahbritto>mdz_: I am using gentoo-sources-2.4.20-r5 with zero problems.
23:24<poptix>cvs as of 2 weeks ago
23:24<mdz_>ahbritto: I'm using NFS
23:24<mdz_>poptix: that code is probably doing a sync() every second or so
23:25<mdz_>which is why you don't see the problem
23:25<poptix>mdz_: i thought you were referring to a kernel issue
23:25<mdz_>poptix: I am. mythtv has that ugly hack to work around it.
23:25<ahbritto>I'm not using NFS.
23:25<Chutt>i need to de-leakify the other modules and make em plugins
23:25<mdz_>without that sync, Linux will basically fill all of main memory with buffer cache
23:25<mdz_>and then try to flush it all at once
23:26<mdz_>which seems like the worst possible strategy for any kind of real-time processing
23:27<poptix>i haven't experienced that kind of behavior with other video capture
23:27<mdz_>I have
23:28<poptix>maybe try 2.4.21 =)
23:28<poptix>i've been running the bitkeeper version +ac
23:29<mdz_>this is a long-standing problem and is not fixed in those kernels
23:29<mdz_>comment out that sync, record something long enough that it will fill your RAM,and watch vmstat
23:29<mdz_>you'll see it write hardly anything, and then write a lot all at once
23:30<mdz_>which usually causes frame drops with software encoding; dunno how much it would affect pvr-250 stuff
23:30<poptix>mdz_: using everything from 'vcr' to avifile for capture, i see a steady stream of data to the hard drive
23:30<Chutt>the mpeg recorder code doesn't have that sync(), actually
23:30<mdz_>poptix: ok, you're right. I"m making it up.
23:31<poptix>mdz_: i'm not saying you're wrong
23:31<poptix>just saying I haven't experienced it
23:31<poptix>how much ram do you have in your system?
23:31<mdz_>I've seen this on every Linux system I've ever had
23:32<mdz_>run dd if=/dev/zero of=somefile
23:32<ahbritto>mdz_: I used to blame over aggressive swaping for delays.
23:32<ahbritto>So, I got rid of my cache, and still had problems.
23:32<ahbritto>Turned out it was the nautilus desktop.
23:33<Chutt>those ram #s i quoted on the mailing list were for a debug build
23:33<Chutt>release is half those
23:33<Chutt>oh well
23:34<mdz_>what ram numbers?
23:34<mdz_>for runtime memory usage?
23:34<extremis>hrm, where can I find info on how I should cvonfigure my X settings for svideo?
23:34<poptix>mdz_: I almost immediately hear the hard drive seeking and writing
23:34<Chutt>some guy was bitching that 256 wasn't enough
23:34<poptix>this is with 768MB of ram
23:35<poptix>i think i'll go try this on my girlfriend's system
23:35<mdz_>poptix: that will fill your write cache very quickly :-)
23:35<poptix>it's running a stock RH kernel
23:35<mdz_>dd if=/dev/zero bs=1M count=<half the megs of RAM in your system> of=somefile
23:35<Chutt>oh, blah
23:35<mdz_>of course, it'll flush the cache when it closes the file
23:36<mdz_>so I guess that won't demonstrate the issue too well
23:36<Chutt>i was going to merge in ben bucksch's movies patch
23:36<Chutt>guess i'll do that later
23:36-!-Ripp [] has quit ["Client exiting"]
23:37<extremis>so, seriously, my tv flips out when I startx and I have it set for 640x480
23:37<Chutt>with the binary nvidia driver, just boot with the tv out the only thing connected
23:38<Chutt>everything should just work
23:40<scanline>the nvtv utility is really helpful though
23:40<Chutt>but the easiest way to get it to work is just boot with only it connected
23:43<poptix>mdz_: alright, on her system with a redhat kernel, it'll buffer up to almost the full amount of free memory before it starts writing
23:44<poptix>mdz_: but, according to gkrellm's disk monitor, it starts writing a continious stream
23:44<mdz_>poptix: which redhat kernel?
23:44<poptix>as in, a fifo
23:44<bigguy>Chutt: that guy with the memory probs had a frontend/backend system on a duron 850
23:44<Chutt>bigguy, i know =)
23:44-!-exs [] has joined #mythtv
23:45<bigguy>Chutt: mistake numero uno
23:45<exs>ok, whatcha mean by 'nvidia binary' ?
23:45-!-extremis [] has quit ["My damn controlling terminal disappeared!"]
23:45<exs>I'm using teh one that gentoo installed
23:45<Chutt>exs, the official nvidia driver
23:45<exs>is that your hint that I should download the one from nvidia?
23:45<Chutt>unless gentoo did that for you
23:45<mdz_>poptix: hmm? how does gkrellm show you the order of the write requests?
23:46<poptix>mdz_: no, only the throughput
23:46<bline>gentoo had the nvidia drivers in portage
23:46<bline>media-video/nvidia-glx and media-video/nvidia-kernel
23:46<poptix>mdz_: the throughput is constant, which would mean it's buffering to 100% then writing in a continious stream
23:47<poptix>mdz_: as in, it's not buffer -> block while dumping to disk -> buffer -> block ...
23:47<mdz_>poptix: no, it's buffer -> write everything to disk causing I/O starvation
23:47<poptix>well yes, i saw that =)
23:47<poptix>49% system cpu usage
23:47<mdz_>the writing process will block when there is no buffer for it
23:47<poptix>i definitely agree there's an issue with that
23:48<mdz_>are you saying you don't see this on a different kernel?
23:48<poptix>mdz_: i don't see it on my 2.4.21-ac3 system, but, it's got a lot of memory in use
23:48<poptix>er, not ac3
23:49* poptixlooks
23:49<poptix>whatever the latest 2.4.21 pre+ac kernel was 2 weeks ago
23:49<poptix>when i recompiled
23:49<Chutt>i'm not getting any email
23:49<Chutt>silly email servers
23:49<-- Snow-Manhas quit (
23:49-!-hachi [] has quit []
23:49-!-yebyen [] has quit []
23:49-!-moegreen [] has quit []
23:49-!-Cloak [] has quit []
23:50<mdz_>there are some tunables in sys/vm/bdflush which might help
23:50<mdz_>I haven't had time to mess with it much
23:50<bigguy>2.4.21 had some fixes for heavy disk io I thought
23:50--> Snow-Man( has joined #mythtv
23:50-!-hachi [] has joined #mythtv
23:50-!-yebyen [] has joined #mythtv
23:50-!-Cloak [] has joined #mythtv
23:50-!-moegreen [] has joined #mythtv
23:50<poptix>mdz_: it could be because this system does have a lot of memory in use, so it may not be buffering nearly as much
23:50<mdz_>that's possible
23:52<poptix>Dave Kleikamp <>:
23:52<poptix> o JFS: Fix hang while flushing outstanding transactions under heavy load
23:52<poptix> o JFS: Avoid deadlock when all tblocks are allocated
23:52<poptix>that's all i see in the changelog, shrug
23:52<bigguy>from .20-.21?
23:53<bigguy>or from a pre
23:53<mdz_>the 2.4.21 patch fixes the documentation for the bdflush sysctl
23:53<mdz_>looks like it's wrong in .20
23:54<ahbritto>Starvation is a known issue and being actively worked on for 2.5.X. While some things may have been back ported, I was under the impression, much more work was planned.
23:54<bigguy>Andrew Morton <>:
23:54<bigguy> o Fix IO stalls and deadlocks
23:54<poptix>looks relevant
23:55<mdz_>ahbritto: I certainly hope this is better in 2.5.x
23:55<ahbritto>It looks like it will be solved or near solved when 2.6 is out.
23:55* blinetravels 20 years in the future to see 2.6 released.
23:57<bigguy>bline: that's E 0.17
23:57<bigguy>you only gotta go 5 years for 2.6
23:57<bline>around the same time perl6 is released.
23:58<ahbritto>Well, 2.6.0 will probably be out sooner than expected, but very unstable. If only to get people to bang on it.
23:58<poptix>i'm waiting a while to start playing with 2.5
23:58<poptix>mostly because of my hatred towards ALSA
23:59<ahbritto>What is better than ALSA?
23:59<poptix>the best tool for the job is better.
23:59<bigguy><raster> I can't stand this hackish coding. I will not have it in e17.(paraphrashed) <joeblow> raster finally learning how to code?
23:59<bigguy>I like alsa
23:59<poptix>my emu10k1 OSS driver works great