Back to Home / #mythtv / 2003 / 04 / Prev Day | Next Day
#mythtv IRC Logs for 2003-04-11

00:00<inman>i thought that memory freed was not immediately returned to the os.
00:00<Chutt>it generally gets reused
00:01<Chutt>but the backend shouldn't be continually newing up ram and stuff
00:01<inman>right, but i would not expect them to see spikey allocation.
00:01<inman>i'd just expect to see it reach a threshold after which it does not grow.
00:01<inman>doesn't malloc scale allocs ahead of the actual alloc that gets returned to the app?
00:02<Chutt>yup, but not massively so
00:02<inman>it's been awhile since i did the whole memory analysis thing with gc, etc.
00:02<inman>i thought it was a toned-down exponential function of some sort.
00:03<inman>i don't understand how people can expect to provoke massive architecture changes in mythtv.
00:03<FryGuy>you mean the linux virtual machine doesn't automatically garbage collect your unused memory? :)
00:03-!-BSDToxic is now known as zZzToxic
00:06<Chutt>'even when the backend dies'
00:06<Chutt>when does the backend die?
00:06<thor>probably flaky/hot hardware
00:07<inman>you mean it's not supposed to die? ;-)
00:07<Chutt>it occurred to me that they never said that mythbackend was using ram
00:07<Chutt>they could be just referring to the system usage
00:07<Chutt>which would of course be including file system buffers, etc
00:08<inman>maybe the backend is dying due to a failed malloc().
00:10-!-hfb [] has joined #mythtv
00:10<Chutt>i could see that if it were really leaking
00:10<Chutt>but i don't see how it could be leaking
00:10<Chutt>12 bytes per recording
00:11<Chutt>that's going to take down a system in like 40 years
00:11<inman>are you using valgrind or something?
00:11<inman>i mean, anything other than gdb?
00:11<Chutt>no, i just went through everything and matched news to deletes
00:11<Chutt>all the recording stuff, at least
00:11<Chutt>most stuff gets deleted in the destructor
00:11<Justin_>i never realized mythtv was so slow because of qt itself..
00:11<Justin_>i just turned off all the eye candy options, and it flies:)
00:12<Chutt>and there's no way that the destructor doesn't get called
00:12<thor>I've had a post 0.8 CVS running for at least a couple of weeks
00:12<Chutt>thor, mdz just said his 0.8 has been running for 3 weeks, but just died =)
00:12<thor>memory looks fine
00:13-!-Ripp [] has quit ["Client exiting"]
00:13<inman>this may be the first complaint about memory that i've seen in the six weeks i've been following myth.
00:13<Chutt>inman, which is why i'm thinking he's just a newbie and looking at what 'free' tells him
00:13<inman>right. i would ignore him until he gave me a reason to worry. :-)
00:14<Justin_>free is good, as long as you look at the right line, which no one does:)
00:14<Chutt>buffers/cache? what's that?
00:14<Chutt>i don't understand?
00:14<inman>i dunno, free ain't that great imo.
00:14<Chutt>it's fairly accurate, as long as you look at the second line =)
00:15<inman>i find ps gives a better picture of where my memory is going.
00:16<inman>Chutt: have you ever used flex++?
00:16<inman>i can't believe i'm going to have to drop down to flex.
00:16<thor>SGI/IRIX used to have this SUPERB little gui tool that would let you drill into memory usage. You always knew what was happening in about 4 clicks
00:17<Chutt>whacha doin with it?
00:17<inman>thor: are you talking about the opengl process manager?
00:17<inman>Chutt: parsing, of course, but the C++ interface is really confusing me.
00:18<Chutt>can someone apply ben brown's patch he just sent?
00:18<Chutt>i should just get a clean cvs tree, hrm
00:18<thor>No, wasn't GL based, just a vertical coloured widget that you could keep clicking on for further detail
00:18<inman>thor: have you seen ClickNet?
00:19<inman>thor: it's like that. uses SNMP to build a hierarchial model of your network, like openview or something, and you can dig into networks->nodes->resources.
00:19<inman>openview is nice, but pricey.
00:19<thor>HP has made buckets on that.
00:20<inman>yeah but they are slipping.
00:20<Chutt>actual use for the cardtype field in capturecard
00:20<inman>product is going downhill and many shops just don't want to run NT just for openview anymore. :-D
00:21<inman>isn't colombo on the ivtv list?
00:21<thor>the OSD patch ... I'm on it if no-one else is already doing it.
00:22<Chutt>thor, thanks
00:22<Chutt>ah, one minor change
00:22<Chutt>the sprintf he added needs to be lld, not ld
00:24-!-PeteCool [] has quit ["Client exiting"]
00:25<inman>omg qmake is renaming my lexer classes
00:26<inman>no wonder...
00:29<Chutt>ok, he was just clueless
00:31<thor>Chut ls to ldd in both sprintf's?
00:31<thor>ld to ldd
00:32<Chutt>just the framesPlayed one
00:32<Chutt>it's a long long, needs to be formatted appropriately =)
00:33<Chutt>the encoding options stuff in nvr is a mess
00:42<thor>Chutt, any thoughts on this stuttery mythmusic interface problem?
00:42<Chutt>i have absolutely no idea
00:55-!-gora [] has quit ["Client exiting"]
01:00<Chutt>someone else had the exact same problem, though
01:00<Chutt>month or two back
01:02<Chutt>thor, sorry for the l/1 confusion
01:03<thor>man sprintf
01:03<thor>it's been a while
01:03<thor>I'd be more confortable with QString's in any case
01:04<Chutt>take a look at the bt that brian foddy just posted
01:04<Chutt>dieing in qstring =)
01:05<inman>the l/1 error is oldschool!
01:08<-- thor(~thor@ has left #mythtv
01:12<inman>did you learn to type on a typewriter before the 1 key was added to the keyboard?
01:14-!-thor [~thor@] has joined #mythtv
01:17-!-knight_ [] has joined #mythtv
01:52<inman>sup knight
03:09-!-choenig [] has joined #mythtv
03:28<Captain_Murdoch>Chutt: you mean use pthread_attr_setdetachstate before the pthread_create?
03:49<Captain_Murdoch>looks like it based upon other times I see 'detach' in the source, so... done...
04:10-!-nziarek_ [] has quit [Read error: 54 (Connection reset by peer)]
04:30<poptix>stop killing random processes.
04:30<inman>i agree, though.
04:33-!-inman [] has quit ["Client exiting"]
04:55-!-hurdel [] has quit []
04:56-!-hurdel [] has joined #mythtv
05:16-!-inman [] has joined #mythtv
06:09-!-lydickaw [] has joined #mythtv
06:09<lydickaw>hmmm. anyone know which X server/driver I need for the shuttle SK41G?
06:36-!-Ndiin [] has quit [Read error: 60 (Operation timed out)]
06:54-!-lydickaw [] has quit [Read error: 113 (No route to host)]
08:01-!-Ndiin [] has joined #mythtv
09:53-!-nziarek_ [] has joined #MythTV
10:00-!-Tack_ [] has joined #mythtv
10:01-!-hfb [] has quit [Read error: 110 (Connection timed out)]
10:14-!-PhracturedBlue [] has joined #mythtv
10:14-!-PBlue_away [] has quit [Read error: 111 (Connection refused)]
10:22<-- hurdel( has left #mythtv
10:50-!-hfb [] has joined #mythtv
11:06-!-Captain_Murdoch [] has quit ["X-Chat!"]
11:07-!-Captain_Murdoch [] has joined #mythtv
11:12-!-Chang [] has joined #mythtv
11:18<rkulagow>nziarek, are you here?
12:21<Chutt>if people are using cvs, please let me know if my two checkins this morning broke anything..
12:21<Chutt>everything seems fine here, but ya never know
12:22-!-StarHeart [] has joined #mythtv
12:24<rkulagow>chutt, did you ever get any cycles to look at the mythmusic config stuff i sent you a few weeks ago?
12:24<Chutt>i just sent it to the list, though =)
12:24<Chutt>i've had time, but i've spent it on other stuff :(
12:25<rkulagow>i'm pretty far behind on the list, so i didn't know you'd just sent it.
12:39<thor>Sheesh ... and will your at it, could you do my taxes for me
12:40<rkulagow>if that's directed to me, i've got my own taxes to do. last minute is the most productive!
12:40<thor>No, to the dev list
12:40<rkulagow>ah, still catching up.
12:49-!-lydickaw [] has joined #mythtv
12:53<rkulagow>chutt: still here?
12:53<rkulagow>two things, easy one first. can you please sync the docs to CVS?
12:53<Chutt>of course
12:54<rkulagow>thanks. next, i've been trying to catch mythbackend in a segfault, so i'm running it in gdb.
12:54<rkulagow>i didn't have a "handle sigpipe", so it went back to gdb when i went to "watch TV" on a remote frontend.
12:55<rkulagow>i had to walk to the main server, set the "handle sigpipe" so that it'd continue, then typed "continue"
12:55<Chutt>shouldn't be getting broken pipes
12:55<Chutt>in the first place
12:55<rkulagow>the backend is saying
12:56<rkulagow>"unknown state transition: watching livetv to watchinglive tv"
12:56<rkulagow>frontend is just sitting at the "connecting to backend server:6543" line.
12:56<Chutt>is that when it crashes, or now that you've added the sigpipe handle?
12:56<Chutt>might be best to restart the backend and frontend
12:57<Chutt>add the sigpipe handler
12:57<Chutt>before running
12:57<rkulagow>i was going to do that next to get everything back in sync, but the backend has gotten confused.
12:57<rkulagow>right, i'll make the changes, then restart
12:58<Chutt>docs update's done
13:00<rkulagow>nziarek: are you here?
13:00<rkulagow>chutt: it's running now (of course); i'll see if i can't catch the segfault. last time it happened i think the backend was kicking off two recordings at the same time, but i'll see if i can get the specifics.
13:01<Chutt>are you using current cvs?
13:01<Chutt>just getting info is all =)
13:01<rkulagow>current as of last night, not after this morning's changes.
13:01<Chutt>ok, that shouldn't make a difference
13:01<rkulagow>i think i need to redownload iulius; it's now crapping out because it can't find ui.xml
13:03<nziarek>rkulagow: i am now...although i don't know what _I_ can help anyone with =)
13:04<Captain_Murdoch>Chutt: your message about the flag commercials thread just meant I needed to use pthread_attr_setdetachstate before starting the thread right?
13:05<Chutt>you're never joining the threads, so it'd leak resources
13:05<Captain_Murdoch>ok, wanted to make sure what I put in cvs was right.
13:05<Captain_Murdoch>ok. thanks for catching that. so if you detach you don't need to join?
13:05<Chutt>it'll clean up for itself once the thread has exited
13:06<Chutt>but it's usually cleaner to join things
13:06<yebyen>i should install 0.8
13:06<Captain_Murdoch>never did much thread programming although that could help in another program I wrote that does use threads a little. :)
13:06<yebyen>or update to more current cvs
13:06<rkulagow>something i noticed last night, not sure if was intended. the status information like "paused, xxx minutes behind" is now in the upper right hand corner. i had paused something, then went away for about 30 minutes. when i looked at it, it appeared that there was font corruption; the first few digits had what i thought was a weird tearing effect. it wasn't until i came back 4 hours later (still paused) that i realized that the "fill ba
13:07<rkulagow>nziarek: can you send me a new pdf file please?
13:07<Chutt>rkulagow, you got cut off at 'fill ba...'
13:07<nziarek>sure; gimme ten minutes, i have to boot in to windows
13:07<rkulagow>something i noticed last night, not sure if was intended. the status information like "paused, xxx minutes behind" is now in the upper right hand corner. i had paused something, then went away for about 30 minutes.
13:08<rkulagow>when i looked at it, it appeared that there was font corruption; the first few digits had what i thought was a weird tearing effect. it wasn't until i came back 4 hours later (still paused)
13:08<rkulagow>that i realized that the "fill bar" had also moved to the upper right hand box, but wasn't on its own line, and so was causing the fonts to look corrupted because it was filling its way across the % full line.
13:08<Chutt>which osd theme?
13:08<rkulagow>i think it's blue. hold on please.
13:08<yebyen>Chutt: what's currently going on in cvs? anything besides the stuff in the top article?
13:09<Chutt>nope, not really
13:09<Chutt>the ui changes moegreen's working on are getting copied to other sections
13:09<Chutt>i'm working on the winttv pvr stuff
13:09<yebyen>moegreen = john danner?
13:09<rkulagow>liquid / blueosd / fresans
13:09<yebyen>cool :)
13:09<yebyen>i haven't played with the new UI stuff, but the screnshots look awesome
13:09<yebyen>"I think I'll watch voyager."
13:10<yebyen>*BAM* which episode? :)
13:10<yebyen>any new work on the commercial skipping?
13:10<Chutt>some, i think
13:10<Chutt>ask captain_murdoch 'bout that
13:11<yebyen>Captain_Murdoch: ehlo
13:11<yebyen>in the cvs version I've got (pre-0.8), it's still only moderately usable
13:11<yebyen>is it still that way in current?
13:12<nziarek>i just used it on futurama the other day and it worked fine
13:12<yebyen>do you skip, then go into the editing thing?
13:13<yebyen>in the version I run, it tags the commercials (or, what it thinks is a commercial) automatically.
13:14<Captain_Murdoch>yebyen: I've been tweaking it a little and should put in a few more tweaks over the weekend sometime. I've been recording shows from different networks and running them through to try to find where it fails. I went off on a rabbit trail detecting the ratings logo but have stopped messing with that for a little while to get the blank frame stuff working better.
13:14<nziarek>i have it set to tag as they record; for some reason I think it runs better that way
13:14<yebyen>Captain_Murdoch: cool
13:14<Captain_Murdoch>yebyen: cvs version requires you to hit 'Z' to convert the commercial break list into a cutlist while in edit mode.
13:14<yebyen>Captain_Murdoch: oh, ok..
13:14<Captain_Murdoch>then 'C' will clear all cutpoints if you want to
13:14<yebyen>does Z do anything else?
13:15<Captain_Murdoch>Z skips forward while watching a recording and Q skips backwards
13:15<yebyen>oh, ok
13:15<yebyen>got it
13:15<yebyen>(it's the "B" key on my remote, heh)
13:15<yebyen>Q skips backwards, hmmm
13:15<Captain_Murdoch>I used "next" and "prev" on my remote
13:15-!-lydickaw [] has quit [Read error: 113 (No route to host)]
13:16<Captain_Murdoch>so ffwd is the 30-second jump, rew is the -5 sec jump, and prev/next jump to the commercial skip marks.
13:16<yebyen>i've gotta try it, heh
13:16<Captain_Murdoch>Chutt: do you think it would make sense to move the cutlist and bookmark into the recordedmarkup table at some point?
13:17<yebyen>any database changes since 20030312? :)
13:17<Captain_Murdoch>I probably already asked that but it just came to mind.
13:19<Chutt>if you want to do it, go right ahead
13:20* Captain_Murdochneeds to get his wife a better video card so she can watch "Designer's Challenge" on her computer instead of taking up the TV. :)
13:20<yebyen>better question, were there any db changes in the 4 days preceeding 0.8 releae :)
13:20<moegreen>heh, i accidentally blew away my ui.xml with my weather changes :(
13:21<Chutt>well, hopefully shouldn't be too hard to reproduce :(
13:21<moegreen>yeah...well I was planning on changing it around some anyways - so any of the weather data can be displayed on any page
13:21<moegreen>I already got radar map back!
13:22<Captain_Murdoch>Chutt: ok, I'll take a look at that sometime. should be easy to merge into there. Also, the framesPlayed display in the edit mode UI, is it ok to have that displayed in the default and blueosd xml files?
13:22<rkulagow>i think i'm going to put some strong wording in the HOWTO to not use KDE as a WM; i switched the frontend to blackbox and i haven't had any problems with window focus and lirc.
13:24<Chutt>i'd kind of rather not
13:24<rkulagow>chutt: was that directed to me?
13:24<yebyen>what's this about MC/programs/convert
13:25<Captain_Murdoch>ok. figured i'd ask first. I was using it to debug and didn't take that line out before commiting. that's why ben submitted his patch I guess.
13:25<-- nziarek( has left #mythtv
13:25<Captain_Murdoch>I'll just add it to my xml file when I need it, or make a copy of a theme so it won't get trashed when I do a make install.
13:28<yebyen>Chutt: what's MC/programs/convert... is it old and dead, or new and nonexistant-yet?
13:33<rkulagow>chutt: any thoughts on the fill bar issue? this is with the liquid theme and the blueosd
13:34<Chutt>not really, other than maybe the osd theme isn't quite right
13:34<rkulagow>ok, thanks.
13:35-!-nziarek [] has joined #MythTV
13:35<rkulagow>is the fillbar supposed to be on a seperate line and isn't so this is a bug, or is this something else?
13:35<nziarek>so, i havn't checked my email in around 4 days, since I've been in Linux and didn't set it up and I am now at 400 new MythTV emails.
13:40<nziarek>rkulagow: pdf is on its way, oh and I am at 600 emails now :P
13:40<rkulagow>nziarek: thanks.
13:42<yebyen>i think i'm going to get a zaurus
13:42<yebyen>maybe not immediately, but when i'm on campus with wireless everywhere, I'm not going to want to carry my laptop everywhere
13:48-!-_shad [] has quit []
13:48-!-paperclip [] has quit []
13:48-!-mdz [] has quit []
13:48-!-_shad [] has joined #mythtv
13:48-!-paperclip [] has joined #mythtv
13:48-!-mdz [] has joined #mythtv
14:04<Chang>Is there a way to clear the cutlist for a recording. I accidentally did a "cut all after" at the very beginning of a show :-(
14:04<Chutt>can't hit e real quick when it goes to playback?
14:05<Chang>It's on a fast box so I can't
14:05<Chang>I was going to nuke the DB and import my shows (only have about 20)
14:05<Chang>But not sure if the cuts are in the DB or in the file
14:05<Chutt>you can just modify the entry in the db
14:05<Chutt>don't need to clear everything
14:05<Chang>Ok, I'll get in there and check it out. Thanks
14:09<rkulagow>chutt: were you going to add the show information as a text type to the .nuv files? if you recall, i had mentioned it as a way around the
14:09<Chutt>i've been wanting to, yes
14:09<rkulagow>"what do i do if i whack the database"?
14:09<Chutt>just haven't had a chance to yet
14:09<rkulagow>ok; not a feature request. :)
14:09<rkulagow>just wondering if it's something within my limited ability to tackle.
14:09<Chutt>might be
14:09<rkulagow>i could probably look at how the CC data is added to the .nuv file
14:12-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
14:13<Chutt>you'd want to do something different
14:14<rkulagow>isn't the CC being added as a "T" type?
14:14<Chutt>yeah, but it's in a special format and stuff
14:14<rkulagow>let me poke around.
14:14<Chutt>easiest way would be to add a new frametype
14:14<Chutt>then store a bit of xml or whatnot
14:14<rkulagow>ah, in that case i'll look around some more.
14:14<Chutt>check out the X header, for instance
14:14<Chutt>the extended stuff
14:14<Chutt>it'd be similar to that, i'd expect
14:15<rkulagow>right; i don't suppose that a full copy of the XML data we have for the show is going to add significantly to the filesize. :)
14:17-!-StarHeart [] has quit ["Client exiting"]
14:20<moegreen>setup is coming along:
14:24<moegreen>Chutt: should I just make a weather-ui.xml file for both blue and liquid, then have my make file install it into those directories?
14:24<Chutt>i'm thinking of only distributing one theme with the main tarball
14:25<Chutt>and splitting the rest into separate downloads
14:25<moegreen>ok, will that be iulius or blue?
14:25<Chutt>haven't decided =)
14:26<moegreen>heh, ok...well I'll probably make a weather-ui.xml for blue and liquid, then we'll at least have them for when the themes get taken out
14:28<yebyen>conflict resolution gone?
14:29<yebyen>or is it inside something now
14:31<moegreen>yebyen: it's in the schedule area
14:32<yebyen>it's an area now, heh
14:33<yebyen>have any themes been updated for the new playback thingy?
14:33<yebyen>besides blue
14:33<moegreen>I put some stuff in there for liquid, though it's a lot like blue's. Iulius has something, but again - it's a lot like blue's right now.
14:35<Chutt>hey look, another dumbass crossposting between both lists
14:36<yebyen>i'm fine with stuff being a lot like it, I just don't really like blue :)
14:36<yebyen>compared to iuluis
14:36<yebyen>it seems faster than a month ago too
14:36<yebyen>(the interface)
14:38<Chutt>it sure better be
14:38<Chutt>moegreen, next, animated weather maps!
14:38<moegreen>yebyen: what was in there a month ago faintly resembles what is in there now
14:39<yebyen>excellent :)
14:39<yebyen>the commercial skipping does seem improved too
14:39<moegreen>Chutt: I though about that...and I think it is possible, it would just need a couple images from the past
14:39<yebyen>even if it's just because you can skip back now
14:39<Chutt>how do they do it on
14:39<Chutt>flash or something?
14:40<yebyen>Can't open: /usr/local/stow/MC-2003041100/share/mythtv/themes/iulius/ui.xml
14:40<moegreen>using the weathercomms class it would be easy to setup a little daemon that downloaded the map every hour or so
14:40<yebyen>Failed to get selector object.
14:40<Chutt>i seem to remember seeing flash on their pages, i dunno =)
14:40<moegreen>hmm...not sure, i'll check
14:40<yebyen>so you've written the stuff for iulius, but it's not in the main site's version of the theme?
14:40<Chutt>yebyen, grab iulius from
14:40<yebyen>ok :)
14:41<moegreen>yebyen: i've just be putting stuff in there so the theme is still usable when I make a change to a screen
14:41<yebyen>april 10 release?
14:42<yebyen>oh, probably :)
14:42* yebyenlooks at today's date
14:42-!-StarHeart [] has joined #mythtv
14:42<moegreen>Chutt: it's a javascript app, imagenames = new Array( ''+thisMap[thisNum]+'1L.jpg'+uniquei,''+thisMap[thisNum]+'2L.jpg'+uniquei,''+thisMap[thisNum]+'
14:42<moegreen>there are 5 images
14:43<Chutt>i was just joking about the animation, really
14:43<Chutt>it'd be neat, but not all that useful =)
14:43<moegreen>well, I suppose I should leave something for the TODO list :)
14:44<yebyen>ooh, sexy
14:44<rkulagow>huh, just got a lame error when i activated PiP on remote frontend and the frontend exited. i think i saw something in the ML a few days ago.
14:44<yebyen>hm, small font though
14:44<Chutt>same error, -3?
14:45<yebyen>moegreen: have you looked at the delete dialog in iulius.3?
14:45<moegreen>yebyen: no
14:45<rkulagow>"lame decode error, exiting" is all that's on the frontend. let me take a look at the backend.
14:46<Chutt>rkulagow, ah, that's not current cvs
14:46<yebyen>ok :)
14:46<yebyen>you should
14:46<Chutt>err, wait
14:46<yebyen>i've gotta go, seeing george carlin
14:47<yebyen>erm, if I leave current cvs running, are there any problems that will cause me to miss star trek?
14:47<yebyen>known problems
14:47<Chutt>not that i'm aware of
14:47<yebyen>in that case, carlin time
14:48<Chutt>rkulagow, can you update the frontend to current cvs?
14:48<Chutt>i made that error message a tad better
14:48<rkulagow>sure, just did that. compiling.
14:49<Chutt>well, current cvs as of when i asked you just a couple seconds ago?
14:49<rkulagow>nuppeldecoder was the only update
14:50<Chutt>that's good, then
14:50<Chutt>thanks =)
14:51<rkulagow>ick. backend is ugly:
14:51<rkulagow>[New Thread 163851 (LWP 7028)]
14:51<rkulagow>adding: pvr as a player 0
14:51<rkulagow>adding: pvr as a remote ringbuffer
14:51<rkulagow>[New Thread 180236 (LWP 7029)]
14:51<rkulagow>Changing from None to WatchingLiveTV
14:51<rkulagow>[New Thread 196621 (LWP 7030)]
14:51<rkulagow>[New Thread 213006 (LWP 7031)]
14:51<rkulagow>[New Thread 229391 (LWP 7032)]
14:51<rkulagow>[New Thread 245776 (LWP 7033)]
14:51<rkulagow>adding: pvr as a player 0
14:51<rkulagow>[New Thread 262161 (LWP 7034)]
14:51<rkulagow>only read adding: pvr as a player 0
14:51<rkulagow>2048 from VIDIOCSYNC0: Interrupted system call
14:51<rkulagow>adding: pvr as a remote ringbuffer
14:51<rkulagow>8192 bytes from '/dev/dsp'
14:51<rkulagow>read audio: Success
14:51<rkulagow>[New Thread 278546 (LWP 7035)]
14:51<rkulagow>only read 2048 from 8192 bytes from '/dev/dsp'
14:51<rkulagow>read audio: Success
14:51<rkulagow>VIDIOCSYNC0: Interrupted system call
14:51<rkulagow>Changing from None to WatchingLiveTV
14:51<rkulagow>[New Thread 294931 (LWP 7036)]
14:51<rkulagow>only read 6144 from 8192 bytes from '/dev/dsp'
14:51<rkulagow>read audio: Success
14:51<rkulagow>VIDIOCSYNC1: Interrupted system call
14:52<rkulagow>[New Thread 311316 (LWP 7037)]
14:52<rkulagow>only read 6144 from 8192 bytes from '/dev/dsp'
14:52<rkulagow>read audio: Success
14:52<rkulagow>[New Thread 327701 (LWP 7038)]
14:52<rkulagow>VIDIOCSYNC0: Interrupted system call
14:52<rkulagow>[New Thread 344086 (LWP 7039)]
14:52<rkulagow>VIDIOCSYNC1: Interrupted system call
14:52<rkulagow>only read 6144 from 8192 bytes from '/dev/dsp'
14:52<rkulagow>read audio: Success
14:52<rkulagow>Changing from WatchingLiveTV to None
14:52<rkulagow>Changing from WatchingLiveTV to None
14:52<rkulagow>baby time. brb
14:52<yebyen>i don't think mine does that
14:52<Chutt>that's just because you're running in gdb
14:52<Chutt>normal, for the most part
14:53<moegreen>heh, canada's weather map does suck :) Too bad.
14:54<inman>moegreen: your latest weather sshot is phenomenal.
14:54<moegreen>inman: the one of europe?
14:54<yebyen>whoa, apparently I don't hae that font installed yet
14:54<yebyen>it's sweet :)
14:54<inman>moegreen: frankly, i wasn't looking at the weather. :-)
14:55<yebyen>ooh and there's an iulius osd :D
14:55<inman>moegreen: i just like the rounded corners, the "mounted" effect. looks much better than the last radar shot.
14:55<yebyen>inman: where is this?
14:56<inman><moegreen> setup is coming along:
14:56<yebyen>sweet :D
14:56<yebyen>ok, i really am leaving
14:57<inman>thar's precipitation in the forecast.
15:11<thor>anyone out there with a big music collection
15:12<Tack_>thor: define big?
15:12<thor>I dunno, 5,000 tracks? Just want someone to try the multithreaded patch to see if it improves things
15:12* Tack_doesn't qualify. :)
15:13<Tack_>I have a humble 1900 tracks.
15:13<Tack_>I don't find startup too terribly bad.
15:14<Chutt>thor, it looked mostly correct to me
15:14<Tack_>At least of mythmusic. I'm more interested in improving startup of mythtv. :)
15:14<Chutt>might be wise to put up a message in the database box that it's still loading
15:15<Chutt>and do something about the playback box too
15:15<Tack_>When will mythmusic look like ? :)
15:15<rkulagow>hrmm: that was unexpected behavior (two tuner system): watching comedy central (chan 42). go to the standard epg, get down to foodtv, hit "R" on the show in the current timeslot. "R" shows up in the EPG, then i hear the audio change. hit ESC, and livetv has changed to foodtv. backend shows the change from watchinglivetv to watching recording. not responding to keypresses, but livetv is still going on food channel. i'll see what hap
15:15<thor>Haven't touched playback ... just wanted to see if it improves things sufficiently to make it worth my while to do it reasonably well
15:16<Chutt>it should improve startup time
15:16<Chutt>at least, time it takes for the main menu to appear
15:16<thor>still can't use it till load is done, but sometimes style *is* more important than substance
15:17<Chutt>rkulagow, the scheduler doesn't consider live-tv being in use =)
15:17<Chutt>and it'll only ask if you're before the start time of the show
15:17<Chutt>i think
15:18<Chutt>not responding to keypresses at all, or just channel changes?
15:18<Chutt>you should still be able to exit out of it
15:18<rkulagow>not getting anything. let me check.
15:18<thor>Tack, as soon as I can figure out how the new UI stuff works
15:18<Tack_>thor: Fair enough. :)
15:19<Chutt>i was messing around with various coverart grabbers
15:19<Chutt>all of em suck
15:19<Chutt>like < 50% success rates
15:19<rkulagow>volume and pause work
15:19<thor>should be a themed option ... I guess
15:20<Chutt>rkulagow, yeah, so, it's in watching the recording mode =)
15:21<inman>Chutt: that's why i think a way to trade metadata (coverart, album/artist/track info) would be useful.
15:21<Chutt>coverart is tricky
15:21<rkulagow>remote is dead; bet you window focus is getting eaten again, even though this is blackbox
15:21<rkulagow>and not kde
15:21<Chutt>track information is facts, so not copyrightable
15:21<Chutt>but coverart is
15:22<inman>but using freenet or something would make it hard to prosecute. ;-)
15:22<Chutt>i'd rather not even get into that :p
15:22* inmanrejoins reality.
15:22<inman>um, yeah, it's a tough problem.
15:23<thor>Chutt, you weren't thinking about some myth-builtin routine to automatically grab coverart, were you?
15:23<Chutt>why yes, yes i was
15:24<thor>oh ...
15:24<thor>I was thinking theme/ui to show it if it happens to exist
15:25<Chutt>but how's it going to happen to exist? =)
15:25<thor>geeky type user who is seriously into music downloads it /scans it/ whatever
15:26<inman>someday when i get some myth time i'm going to move images into the db.
15:26<thor>regular user never bothers
15:26<inman>been on my todo for ages.
15:26<Chutt>thor, it'd be nice if it could snag it when it ripped something
15:26<inman>if the tool to snag coverart isn't included in myth, would you worry about ©?
15:27<Chutt>snagging it yourself is fine
15:27<inman>so there's the solution. :-)
15:27<Chutt>you're not copying it =)
15:27<Chutt>redistributing it to other people is copying it
15:27<inman>i'm not talking about redistributing the images.
15:28<Chutt>then what does 'trade metadata' mean?
15:28<inman>does that mean you don't have a problem including a tool that slurps images off the network?
15:28<Chutt>what's the network?
15:28<inman>Chutt: individual users trading metadata with each other, p2p.
15:28<inman>Chutt: inet.
15:28<Chutt>yes, that's other people redistributing things
15:29<inman>Chutt: but they are the ones breaking the law.
15:29<inman>Chutt: this is the whole p2p argument, you realize?
15:29<Chutt>i know
15:29<inman>and you don't want to go there?
15:29<inman>fair enough. :-)
15:29<Chutt>grabbing coverart from amazon or walmart or whatnot is fine
15:30<inman>the irony. :-)
15:30<thor>You know ... a frontend streaming from a backend on another box is ... well ... redistributing content :-)
15:30<Chutt>thor, yup
15:30<inman>thor: tivo does that.
15:30<Chutt>but, bandwidth requirements are kind of steep
15:30<Chutt>can't exactly go out on the net
15:31<thor>A Pragmatist, how rare
15:32-!-zZzToxic is now known as NonToxic
15:33<Tack_>I don't suppose there exists RHL RPMs for Myth and its deps?
15:33<Chutt>i think so
15:33<Chutt>mythtv rpm @ google :p
15:35<Tack_>Nightly RPM builds of myth would be cool. :)
15:35<Chutt>no they wouldn't
15:35<Chutt>thou shalt not distribute binaries of cvs
15:36<Chutt>else public cvs access goes away =)
15:36<Tack_>Well, there goes that idea. :)
15:36<thor>Chutt .... you know that as soon as the PVR stuff is in a release that someone is gonna write MythGnutella.
15:36<Chutt>i want people to know they're using cvs, have compiled it themselves so they can switch over to debug mode
15:37<Tack_>Chutt: Fair enough.
15:46<thor>for new theme-ifying mythmusic (starting with playback box), I need to write my own widget (in libmythtv?) that inherits from MythDialog and then keep track of the "namespace" between ui.xml and what my constructor code is doing. Does that sound fundamentally wrong to anyone?
15:46<Chutt>it'd live in mythmusic, but, yeah
15:47<Chutt>the playbackbox in the frontend should be similar
15:47<Chutt>or the viewscheduled box, though that lives in the libmythtv lib
15:47-!-lydickaw [] has joined #mythtv
15:48<moegreen>thor: what type of widget do you need?
15:49<thor>just starting to think about this ...
15:49<moegreen>I made the UIListType to be versitile, but that means you have to do some of the legwork on your end
15:50<thor>for playback .... just a list of tracks
15:50<thor>I'll look at UIListType ...
15:50<thor>when you're writing a theme, the buttons are just text and png for on and off, yes?
15:51<thor>never mind .. found it
15:51<moegreen>thor: the listtype doesn't hold all the data, it just displays a segment of it and takes care of the selection bar and the layout
15:51<Chutt>i think there'll be a 3rd button image as well eventually
15:51<thor>for the other half of caputo marital bliss
15:54<thor>holy crap there's a lot of work that's been done in here
15:56<moegreen>thor: i'm going to commit some changes uitypes and xmlparse in sec
15:58<moegreen>Chutt: if I broke something (doubt I did), but if I did, could you fix it? I'll be travelling for day or so - thanks!
16:00<Chutt>have fun
16:00<lydickaw>any debian "testing" users care to share which 4.3 backport they are using?
16:14-!-PeteCool [] has joined #mythtv
16:30-!-Ndiin [] has quit []
16:30-!-inman [] has quit []
16:31-!-bigguy [] has quit []
16:31-!-StarHeart [] has quit []
16:31-!-Chang [] has quit []
16:31-!-Captain_Murdoch [] has quit []
16:31-!-hfb [] has quit []
16:31-!-nziarek_ [] has quit []
16:31-!-rkulagow [] has quit []
16:31-!-FryGuy [] has quit []
16:31-!-sc00p [] has quit []
16:31-!-knight_ [] has quit []
16:31-!-Justin_ [] has quit []
16:31-!-Chutt [] has quit []
16:31-!-StarHeart [] has joined #mythtv
16:31-!-Chang [] has joined #mythtv
16:31-!-Captain_Murdoch [] has joined #mythtv
16:31-!-hfb [] has joined #mythtv
16:31-!-nziarek_ [] has joined #mythtv
16:31-!-knight_ [] has joined #mythtv
16:31-!-rkulagow [] has joined #mythtv
16:31-!-FryGuy [] has joined #mythtv
16:31-!-sc00p [~oldendic@] has joined #mythtv
16:31-!-Chutt [] has joined #mythtv
16:31-!-Justin_ [~justin@H233-225.STATE.RESNET.ALBANY.EDU] has joined #mythtv
16:31-!-Ndiin [] has joined #mythtv
16:31-!-inman [] has joined #mythtv
16:31-!-bigguy [] has joined #mythtv
16:32-!-lydickaw [] has quit [Read error: 113 (No route to host)]
16:37-!-sc00p [] has quit []
16:37-!-rkulagow [] has quit []
16:37-!-knight_ [] has quit []
16:37-!-nziarek_ [] has quit []
16:37-!-Captain_Murdoch [] has quit []
16:37-!-Chang [] has quit []
16:37-!-Justin_ [] has quit []
16:37-!-FryGuy [] has quit []
16:37-!-Chutt [] has quit []
16:37-!-StarHeart [] has quit []
16:37-!-hfb [] has quit []
16:37-!-inman [] has quit []
16:37-!-Ndiin [] has quit []
16:37-!-bigguy [] has quit []
16:38-!-Ndiin [] has joined #mythtv
16:38-!-inman [] has joined #mythtv
16:38-!-bigguy [] has joined #mythtv
16:38-!-StarHeart [] has joined #mythtv
16:38-!-Chang [] has joined #mythtv
16:38-!-Captain_Murdoch [] has joined #mythtv
16:38-!-hfb [] has joined #mythtv
16:38-!-nziarek_ [] has joined #mythtv
16:38-!-knight_ [] has joined #mythtv
16:38-!-rkulagow [] has joined #mythtv
16:38-!-FryGuy [] has joined #mythtv
16:38-!-sc00p [~oldendic@] has joined #mythtv
16:38-!-Chutt [] has joined #mythtv
16:38-!-Justin_ [~justin@H233-225.STATE.RESNET.ALBANY.EDU] has joined #mythtv
16:38<-- Chang( has left #mythtv
16:38-!-nziarek [] has joined #MythTV
16:39<thor>that was weird
16:41-!-Tack [] has joined #mythtv
16:43-!-nziarek [] has quit [Read error: 54 (Connection reset by peer)]
16:52<inman>nice debugging by gerald britton to find the thread leak.
16:57<Chutt>what, saying 'it doesn't shut down threads on image transfers'?
16:59-!-lydickaw [] has joined #mythtv
17:01<inman>yeah, noone else grabbed it. :-)
17:03<Chutt>i imagine not many people run remote frontends
17:04<Chutt>the repeat/previous shown date isn't in the tv_grab_na data
17:06<inman>i know, that's why i asked. :-)
17:06<inman>i think it's uk due to 'MyTelly'.
17:28<Chutt>the teletext code looks all crappy
17:28<Chutt>static buffer sizes, not checking lengths
17:37-!-PeteCool_ [] has joined #mythtv
17:37-!-PeteCool [] has quit [Read error: 104 (Connection reset by peer)]
17:44<PeteCool_>hmm, do the G400 X drivers do Xv? I see lots of stuff about using it in the framebuffer... but that would mean slow video in my slow system
17:44<PeteCool_>I don't have much cycles to spare
17:46-!-lydickaw [] has quit [Read error: 113 (No route to host)]
17:47<choenig>PeteCool_: they do at least on the first head
17:58-!-TheAsp [] has joined #mythtv
18:08<PeteCool_>what permissions should the video device have? 777? it's at 555 now, and I can't use the contrast/etc settings
18:09<choenig>I have 555, too, but everything works fine here
18:10<PeteCool_>lr-xr-xr-x 1 root root 10 Apr 8 17:11 /dev/video0 -> v4l/video0
18:10-!-PeteCool_ is now known as PeteCool
18:10<choenig>elara ~ # dir /dev/video*
18:10<choenig>lr-xr-xr-x 1 root root 10 Apr 11 11:19 /dev/video0 -> v4l/video0
18:10<choenig>lr-xr-xr-x 1 root root 10 Apr 11 11:19 /dev/video1 -> v4l/video1
18:10<choenig>lr-xr-xr-x 1 root root 10 Apr 11 11:19 /dev/video2 -> v4l/video2
18:11<TheAsp>crw-rw---- 1 root video 81, 0 1969-12-31 20:00 /dev/v4l/video0
18:12<choenig>same here
18:12<TheAsp>PeteCool: thats just the link, not the device
18:12<PeteCool>TheAsp: yeah, just saw that
18:12<PeteCool>crwxrwxrwx 1 root video 81, 0 Dec 31 1969 /dev/v4l/video0
18:17<PeteCool>what's wrong with that?
18:19<TheAsp>my screen got messed
18:19<TheAsp>though it doesn't need to be +x
18:24<choenig>PeteCool: I wish you much luck with you g400 ;-) ... see you tomorrow
18:24-!-choenig [] has quit ["wuuuusch..."]
18:26-!-hfb [] has quit ["Huh! Wasn't me!"]
18:47<Tack>I'm going to email the dev list with probably 30k worth of text (backtraces and system info and such). Is this kosher?
18:48<PeteCool>Tack: that's big
18:48<Tack>I know.
18:49<PeteCool>Tack: do you have an url to post it to?
18:49<Tack>Yes, but that sucks for archives. I really, really loathe it when people do that when I'm trying to debug something. :)
18:49<thor>bzip2 -9
18:50<Tack>Not very convenient, but okay.
18:50<Tack>Whatever Chutt wants. :)
18:50<Tack>(Convenient for the viewer, that is.0
18:50<thor>When I had my huge mythmusic patches before the commits, he suggested bzip2 -9
18:51* Tacknods.
19:00-!-hfb [] has joined #mythtv
19:12<Tack>Okay, the total text is only about 17k. Should I still bzip?
19:12<Tack>Ah, I see emails bigger than that in -dev. :)
19:37-!-StarHeart [] has quit ["Client exiting"]
19:44<inman>Tack: if you were going to include all that data, you should have include the output of `lsmod`. :-)
19:46<Tack>Yeah, it's funny all the relevant things you forget. :)
19:46<inman>i don't think it matters here, though.
19:46<inman>this looks like an i/o problem.
19:47<Tack>You might be right. It occurs to me that I was running a debug build. I'm rebuilding MythTV for release and see if the problem goes away.
19:48<Tack>The irony is that I built the debug version to troubleshoot another problem, and then found all sorts of new problems, and I forgot what the original thing I wanted to troubleshoot was. :)
19:48<inman>normal playback works fine, right?
19:49<inman>is the backend unpausing and the frontend staying paused, maybe?
19:49<Tack>So it does seem like a CPU or I/O thing, which is what I thought, and then it occured to me I should build a non-debug version to see if it's related to lack of optimizations.
19:49<inman>definitely feels like a sync issue, but i can't take a look at it right now.
19:50<inman>anyway, i'm sure someone else more familiar with the code will figure it out pretty quickly.
19:50<Tack>Well, it could be that the back-end is unpausing. How do I tell?
19:50<inman>put some debugging statements into tv_rec.cpp
19:51<inman>just cerr << "i'm unpausing" << endl (or whatever).
19:51* Tacknods.
19:51<Tack>If the non-debug build has the same problem, I shall.
19:51<inman>how current is your cvs?
19:52<Tack>A few minutes. :)
19:57-!-gora [] has joined #mythtv
19:57<gora>is mythtv installable "easily" with gentoo? or should i expect bumbs along the installation road the whole way through?
19:58<Tack>Live tv is pretty fubared. Same problem with the non-debug build; and audio is horribly out of sync, and about 50% of the time I start watching tv, after 5 seconds or so I get this horrible scratchy/hissy loud noise, as if someone did cat /dev/urandom > /dev/dsp
19:59<Tack>It _was_ working, anyway. I'll go back to 0.8 and see if that works. :)
20:03<gora>is there a way to install mythtv but not mysql if i am going to use the mysql database on another machine ?
20:05<thor>gora, you still need qt with mysql support and then just client side mysql stuff
20:05<gora>thor: hmmm, okay. might as well just install mysql anyways, not a big deal at all.
20:11-!-hfb [] has quit ["Huh! Wasn't me!"]
20:12-!-nziarek_ [] has quit [Read error: 60 (Operation timed out)]
20:14<Chutt>tack, should have that hang you wrote in about fixed soonish
20:14<Chutt>but i'm goin out now, so, not yet =)
20:16-!-lydickaw [] has joined #mythtv
20:16<Tack>Chutt: I followed up again ...
20:17<Tack>Another very alarming problem, which might be related, is the hiss/scratch/noise (cat /dev/urandom>/dev/dsp) I hear half of the time.
20:17<Tack>But, thanks. :)
20:18<Tack>Actually it doesn't sound quite like that. There's more high frequencies.
20:28-!-Ripp [~ripp@] has joined #mythtv
20:28<PeteCool>gora: there were a few little problems in gentoo with mythtv, but nothing really important, mostly hardware-related problems
20:28<PeteCool>gora: lke the /dev/lirc entry not being recreated at boot
20:29<PeteCool>gora: or certain kernel patchsets crapping out with the nvidia drivers
20:30<gora>PeteCool: hmmm. is there anyway to not have mysql need to be emerged? i would like to use the mysql database running on another machine on my LAN
20:30<PeteCool>gora: no, because qt needs mysql parts to access that db
20:31<PeteCool>gora: also, make sure you have the mysql use flag set when building qt
20:31<gora>hmmm damnit. so i must install mysql on the local machine that mythtv will be on, but i can then post-install have it use a mysql database on my lan. correct?
20:31<PeteCool>gora: exactly.
20:31<PeteCool>gora: and you don't need to actually start mysql on the myth box
20:32<gora>got cha
20:32<gora>just trying to save a LONG emerge time ;)
20:32<Tack>Does anyone know where xmltv gets its data from?
20:33<PeteCool>gora: if you have multiple gentoo boxes (maybe other distros too) you can distribute the emerges with distcc
20:33<gora>PeteCool: already doing that with 2 machines
20:33<gora>one step ahead of ya
20:33<PeteCool>gora: my celeron500 emerges 2x faster now, grafting on a celeron1100 helps a lot
20:34<gora>i have a 1,4 athlon on my desk that overheats A LOT. so i share its compiles with the p3/933 at my feet.
20:34<gora>if i get mythtv working on my desk just as a tester i'll scrounge up antoher machine to dedicate for mythtv.
20:34<PeteCool>gora: you can undervolt the 1.4 without problems... my 1700+ is specced at 1.75, but I run it at 1.70... temps dropped >5C
20:34<Chutt>tack, totally unrelated problems
20:35<gora>PeteCool: what setting would i look for in the bios to do such a thing? i'd rather not kill my machine either ;)
20:35<Tack>Chutt: do you have an idea of what the other one could be?
20:35<Chutt>sounds like a driver problem, honestly
20:35<PeteCool>gora: I don't think undervolting can kill it, it'd just hang/reboot
20:35<PeteCool>gora: bios option depends on your board model and maker
20:36<Chutt>or, how much free cpu do you have when it's in livetv mode?
20:36<gora>PeteCool: i believe this will be my one and only athlon. i'm a poor college student, i'd like to build an smp box next.
20:36<Tack>Chutt: I've used both ALSA and OSS, same problem. MythTV 0.8 doesn't do it.
20:36<Chutt>there hasn't been a change that could cause that
20:36<Tack>Chutt: Not much, but some. Between 5-20% usually.
20:36<PeteCool>gora: well, athlons can make pretty good smp boxes... you simply picked the "worst" (most heat-emitting) athlon ever
20:37<Chutt>in the debug build, too?
20:37<Tack>I haven't tried a debug build of 0.8, but it happens with both debug and release builds of CVS.
20:37<gora>PeteCool: thats what i've heard from EVERYONE. i happen to buy the hottest one out. i got a new monster fan/heatsink on it, and sometimes i have to pop open the case and point a small turbo fan at the cpu for big emerges.
20:37<PeteCool>gora: I don't see what inexpensive smp is available at the moment, can you tell what you're planning?
20:37<Chutt>try not compressing the audio
20:37<Chutt>see if that changes anything
20:37<Tack>Chutt: I am. I'm using uncompressed audio at 44100hz.
20:37<Tack>So, I can try compressing. :)
20:38<Chutt>ok, drop the res down and try compressing it
20:38<gora>PeteCool: i was waiting for the intels new HT architecture to become more mature and "cheaper." i've seen it in use on campus in big computational uses and its just flat out amazing
20:38<PeteCool>Tack: what about the other sampling rates?
20:38* Tackgoes to try that out. (Myth box in another room.)
20:39<PeteCool>gora: that's not really smp, it pushes the instructions to the cpu in a different way... I don't think HT is going to get college-student-inexpensive for quite a while
20:39<PeteCool>gora: though it would force amd to get off their butts some more
20:40<gora>PeteCool: i'm 20 minutes from a couple huge AMD R&D campuses. our campus gets inundated with donations from amd when new chips are released. as with intel.
20:48-!-Tack3 [] has joined #mythtv
20:48<lydickaw>anyone have a an XF86Config-4 for a ProSavage3? (onboard on a SK41G shuttle system)
20:48<Chutt>tack, fixed your hang, it was a typo i introduced earlier today
20:48<Tack3>Chutt: Goodie. :)
20:49<Chutt>not in cvs yet, going to fix the other guy's similar issue
20:50<PeteCool>gora: hmm, care to send me some if you have leftovers? =)
20:51<Tack3>Something is awfully volatile here. I was playing with the mixer, and suddenly I'm no longer getting sound in live tv. Line-in is set for record, and the volume is fine for output (playing prerecorded shows works). I feel awfully silly here. What am I missing?
20:52<PeteCool>Tack3: what sound chip are you using?
20:52<Tack3>And stuff _is_ coming in on line-in. If I unmute it, works fine.
20:52<Tack3>PeteCool: emu10k1
20:52<Tack3>Guess I can try unloading/loading the sound modules.
20:53<Tack3>I think the frontend could handle things more gracefully if the back end goes away. :)
20:53<Chutt>it doesn't make any attempt to yet
20:53<PeteCool>Tack3: maybe there's a capture-level setting somewhere
20:54<PeteCool>Tack3: raising/lowering the Capture line has an effect on the cmi8738
20:55<Tack3>No affect here.
20:55<Chutt>since when did it allow you to ff/rew while paused?
20:55<Tack3>I don't know, but I rather like that. :)
20:55<Chutt>it's assuming you can't do that
20:55<Chutt>i wonder how..
20:56<PeteCool>Chutt: what do you mean exactly?
20:57<PeteCool>Chutt: I do that almost every myth "session"
20:58<Chutt>it didn't use to be allowed
20:58<Chutt>used, whatever
20:59<PeteCool>Chutt: hmm, it has always done that for me since jan ~25
20:59<Tack3>It's a very useful feature. :)
20:59<Tack3>But if I pause, and fast forward to the end of stream and unpause, it seems to either hang or crash. :)
20:59<Tack3>(Perhaps that's what you're working on.)
20:59<Chutt>well, yes
21:00<Chutt>since it's assuming you can't ff while paused
21:00<PeteCool>Chutt: though sometimes it crashes, when I move too much in small steps in the stream
21:00<Tack3>Please don't remove that "feature." :)
21:00<PeteCool>pgup/down seems ok
21:00<Chutt>i'm just going to make it unpause when you hit ff or rew
21:01<Tack3>That's probably okay.
21:01<Chutt>of course, edit mode is different, but..
21:02<PeteCool>are you sure it's not possible to keep it working like it is now while keeping the code sane?
21:02<Chutt>i dunno
21:02<Tack3>If I'm not paused and I seek to the end (i.e. press and hold the right key for a while), myth will hang.
21:02<Tack3>But maybe you know that too.
21:03<Tack3>Oh it didn't hang forever, it snapped out of its deep thought after about 15 seconds.
21:03<PeteCool>I'm pretty sure that could be fixed with a small test
21:03<Tack3>It seems to me that it'd be easier to make seeking work while paused than while playing. :)
21:03<Chutt>not hardly
21:04<Tack3>Well what do I know. :)
21:04<Tack3>Does anyone here know if a separate VGA->NTSC converter will produce better quality than the svideo out on a video card?
21:07<Chutt>really depends on the quality of the converter and the quality of the video card
21:07<PeteCool>Chutt: what problem(s) does the unplanned seeking create?
21:07<PeteCool>unplanned = unaccounted for
21:08<Tack3>Chutt: True, but assuming prosumer range in both cases. I just have no experience at all with VGA to NTSC converters so I wouldn't know what to expect.
21:08<Chutt>well, that hang ben bucksch reported on the list, for one :p
21:08<Tack3>The reason why I ask is I need either a new video card or a VGA->NTSC converter to complete my setup. :)
21:09<Tack3>(And a hole saw, but that's another issue.)
21:12<Chutt>your hang is fixed
21:12<Chutt>so's the other guy's
21:12<PeteCool>Chutt: Seems like he's going past the eof, could the unpaused eof ff test be used for paused ff ?
21:12<Tack3>Chutt: Awesome. :)
21:12<Chutt>it shouldn't be letting you seek that close to the end
21:13<Chutt>tack, does audio pause when you pause live tv?
21:13<Tack3>Chutt: Yes.
21:14<Chutt>just checking
21:14<Tack3>I unloaded/reloaded snd-* modules. It started working again. *shrug* Could have been a bug in the drivers. Also could have been a user error. :)
21:14<Tack3>Testing out cvs ...
21:14-!-lydickaw [] has quit ["food."]
21:15<Tack3>It sure scares the shit out of me when the 100% volume noise starts spewing.
21:15<Tack3>ran out of free AUDIO buffers :-(
21:15<Tack3>audio behind 245 313
21:15<Tack3>audio behind
21:15<Tack3>Getting a number of those kinds of messages from mythbackend.
21:16<Tack3>"ran out of free AUDIO buffers :-(
21:16<Tack3>audio behind 245 313
21:16<Tack3>audio behind" from mythfrontend.
21:16<Tack3>Also mythtv took about 10-15 seconds to start. I bet if I go back in, it'll work fine.
21:16<PeteCool>Tack3: many people said on the list the oss driver was better for recording on the sblive
21:16<PeteCool>(better than the alsa drivers)
21:16<Tack3>Yeah, go back in again and it's fine.
21:16<Tack3>PeteCool: That could be true, except I'm seeing the same thing with OSS.
21:17<Chutt>really almost sounds like a driver issue
21:17<PeteCool>Tack3: got any other sound cards to test with?
21:17<Tack3>It does. I'll drop the sample rate and go back in a few more times and see what happens.
21:18<Tack3>PeteCool: I have a SB Live Value with an emu10k1x chipset (emu10k1-wannabe that Dell sells) that doesn't work with Linux. :)
21:19<Tack3>Will MythTV support ogg in the future?
21:20<PeteCool>Tack3: not until they "fix" the library
21:20<Chutt>i don't really think it's worth it, honestly
21:20<PeteCool>Tack3: Chutt told me it takes lots of time to get back the encoded audio
21:20<Chutt>and, there's that
21:20<Chutt>no low-delay encoding mode
21:20<Tack3>I see.
21:21<Tack3>Same results with 32khz uncompressed. I'll try mp3.
21:21<PeteCool>Chutt: hmm, their int-based lib is decode only right?
21:21<PeteCool>Chutt: found it
21:23<Tack3>I can usually tell when it's going to mess up because mythtv might take very long to start (10-15 seconds) or when it does start, it starts extremely jerky and out of sync with audio. That lasts 5-10 seconds then SCREEEEETCCHH.
21:24<PeteCool>Tack3: lsmod output?
21:24<PeteCool>Tack3: I mean can you post it?
21:26<Tack3>Hmm, well mythtv started massively out-of-sync again, but this time it didn't mess up. Using mp3.
21:26<Tack3>I don't want to use mp3 thought.
21:27<Tack3>Actually, it seems that with mp3, it consistently starts out-of-sync and takes about 5-15 seconds to sync up. Is this normal? (i.e. do other people see this?)
21:27<Tack3>My CPU isn't maxed.
21:27<PeteCool>I don't think you need tvaudio - I'm not using it, I don't know anybody who does, and it could be messing with the tuner's state
21:28<PeteCool>Tack3: maybe your disk can't keep up... it does that to me too, and my disk is slow
21:28-!-Chang [] has joined #mythtv
21:29<Tack3>Yeah, I don't have the fastest disk.
21:29<Tack3>PeteCool: I might not need tvaudio, but it's not letting me rmmod it.
21:30<Tack3>And I don't know what device I can lsof to see if a process has it grabbed.
21:30<Tack3>PeteCool: What does hdparm -tT show for you?
21:32<Tack3>Chutt: Have you played much with your PVR-250?
21:32<PeteCool>Tack3: it's giving me negative numbers... but I think the kernel patchset I'm using might be messing with that
21:32<PeteCool>Tack3: it gave me about 20~25 for the slowest of the two numbers with previous kernels
21:33<PeteCool>Timing buffered disk reads: -1387 MB ;)
21:33<Tack3>Yeah, I get about that. I get something like 45 at work. I should invest in a fast disk. :)
21:33<Tack3>Well, that's impressively slow. :)
21:34<PeteCool>it can't keep up with >10 mins of live rtjpeg-170
21:34<PeteCool>but mpeg4 is fine, even with HQ encoding
21:35<PeteCool>even if it's only at 320x480, it's still a celeron 1.1ghz encoding
21:37<Tack3>I won't be happy until I can do 640x480 high quality mpeg4 with uncompressed 44.1k audio. :)
21:40<Tack3>I'm still able to confuse mythtv by holding down the right arrow during live tv.
21:41<Tack3>It's not a critical problem. It's just a matter of robustness.
21:42<Tack3>It freezes, and takes about 20-30 seconds to snap out of it.
21:42<Tack3>But the freezing problem I reported in CVS appears fixed.
21:45<PeteCool>Chutt: any ideas of why I might not be able to change contrast/brightness/etc? I think my kernel is using v4l2, but IIRC it's compatible with v4l1 for those kind of things, and bttv still is a 0.7.x version
21:45<Tack>That might be more of an Xv issue?
21:47<PeteCool>Tack: you can tweak those settings directly at the tv tuner level with myth/videox4linux... that's what's not working
21:48<Tack3>Ah, I see.
21:49<PeteCool>VIDIOCSPICT: : Invalid argument is the exact error
21:54<PeteCool>as if 32768 was an out-of-range value... hmm
21:56<Tack3>What are MythTV's default values for bitrate, and max/min quality for mpeg4?
21:57<PeteCool>2200, 2 (or 3?), 15
22:03-!-PeteCool [] has quit ["Client exiting"]
22:11<Tack3>Chutt: When one is viewing at the end of the ring buffer (i.e. seeking forward does nothing), maybe mythtv should display "Live"? My friend's Bell Expressvu PVR works like that.
22:29<thor>Damn non-thread safe Qt classes!!
22:30<Chutt>which one?
22:31<thor>have to lock around whole damn iteration loop
22:31<thor>(four days of playing with threads and I'm already complaining)
22:33<thor>I'm cleaning up that Caputo patch .. not sure it does much good since you get to the Select screen real fast, but then everything is greyed out until all your metadata and playlist arrive
22:36<thor>better for playing though ... Enter -> Down -> Enter (visual feedback, walk away)
22:39<Chutt>live-tv mode
22:39<Chutt>with pausing, still no seeking =)
22:40<thor>YYYAAAAHHHHHH^n (where n > 2)
22:40<thor>(I'm assuming this Hauppauge PVR)
22:40<Chutt>no channel changing, either
22:40<Tack>Very cool.
22:40<Chutt>but that should be relatively easy
22:40<Chutt>just have to update things to use v4l2 to switch channels
22:41<Tack>Noticeable performance improvement with the PVR?
22:42<Chutt>only thing using cpu is the playback stuff, so, yeah
22:42<Tack>So that uses mpeg2 for compression? Presumably the quality is quite good.
22:42<Chutt>huge files
22:43<Tack>Harddisks are cheap. :)
22:43<Chutt>quality's about the same as mpeg4 at the same rate
22:43<Chutt>maybe a tiny bit better
22:43<thor>Huge vis-a-vis RTJPEG?
22:43<Chutt>thor, well, i dunno, 4 gig an hour for very good quality
22:43<Chutt>so less than rtjpeg
22:44<Chutt>and probably looks better than rtjpeg, i'd say
22:44<thor>and easier for other tools to mess with afterwards
22:44-!-Ripp [] has quit [Read error: 113 (No route to host)]
22:44<Chutt>yeah, probably
22:44<Chutt>the .mpg format sucks, btw
22:44<thor>yes, but ubiquitous
22:45<Tack>When things get down to 1G/hour that'll be nice. On par with ripping DVDs. :)
23:07<TheAsp>hey chutt, where do you open the files?
23:07<TheAsp>for playing
23:07<TheAsp>well, and recording
23:07<TheAsp>(nupplevideo files)
23:10<Chutt>the ringbuffer
23:10-!-PeteCool [] has joined #mythtv
23:10<TheAsp>not libavformat?
23:11<Chutt>that's only used to read mpg files
23:11<TheAsp>oh, ok
23:12-!-Edgan [] has quit ["Client exiting"]
23:13-!-NonToxic is now known as zZzToxic
23:17<Chutt>i think it's funny that the ivtv driver says this card has 5 composite inputs, 4 svideo inputs, and 1 tuner
23:17-!-Edgan [] has joined #mythtv
23:21-!-Ripp [~ripp@] has joined #mythtv
23:21<TheAsp>going into view recordings does nothing here