Back to Home / #mythtv / 2003 / 03 / Prev Day | Next Day
#mythtv IRC Logs for 2003-03-23

00:02-!-paperclip [] has quit [Connection timed out]
00:05-!-paperclip_ [] has quit [Read error: 60 (Operation timed out)]
00:16-!-paperclip_ [] has joined #mythtv
00:25-!-paperclip_ [] has quit [Read error: 60 (Operation timed out)]
00:39-!-paperclip [] has joined #mythtv
00:54-!-Captain_Murdoch [] has joined #mythtv
00:57-!-paperclip [] has quit [Connection timed out]
01:42<Captain_Murdoch>Chutt, you around?
02:08* bigguyis [away -={ S.L.E.E.P }=- ]
02:26-!-PeteCool [] has quit ["Client exiting"]
03:18-!-Soopaman [] has joined #mythtv
06:55-!-Ndiin [] has quit [Read error: 110 (Connection timed out)]
07:50-!-choenig [] has joined #mythtv
07:56-!-choenig [] has quit [Read error: 54 (Connection reset by peer)]
07:57-!-choenig [] has joined #mythtv
08:09-!-Captain_Murdoch [] has quit [Read error: 104 (Connection reset by peer)]
08:19-!-Ndiin [] has joined #mythtv
08:37-!-bigguy [] has quit [No route to host]
10:52-!-Edgan [] has quit ["Client Exiting"]
11:17-!-Edgan [] has joined #mythtv
11:32-!-bigguy [] has joined #mythtv
11:42-!-choenig [] has quit [Remote closed the connection]
11:50-!-choenig [] has joined #mythtv
11:57-!-PeteCool [] has joined #mythtv
13:05-!-mdz [] has quit ["Client Exiting"]
13:13-!-Captain_Murdoch [] has joined #mythtv
13:33<PeteCool>hmm, mythbackend dies for no visible reason
13:35<PeteCool>Chutt: are the contrast-setting-errors supposed to be fatal?
13:36<PeteCool>Chutt: also, are the constrast settings dependent on v4l1 (vs v4l2) ?
13:36<Soopaman>what are the lowest specs required to record/time shift with myth?
13:39<Captain_Murdoch>Soopaman: depends on the resolution you want to record at. I'm running on a Athlon 700 and record at 352x480, but can't record and playback at the same time. If I bumped the resolution down I could though. 352x400 seems to use about 75% cpu which gives me enough room to playback at the same time and not cause problems.
13:40<PeteCool>It depends on the hard disk speed too
13:40<PeteCool>I couldn't do much until I swapped drives around, increasing hdparm -tT scores by 10 mbytes/s
13:42<Soopaman>well I am hoping to use my dual xeon 450 scsi system as my new myth box
13:42<Captain_Murdoch>I run my storage directory over nfs. 100Mbit ethernet
13:44<Captain_Murdoch>Soopaman: you might only be able to do 320x240 with that, I'm not sure. check out the pvr hardware site to see what others say.
13:49<Captain_Murdoch>make sure you look at the rating field at the end. there are some people on there who try to do 640x480 mpeg4 on a P2-450 and then give it a rating of 1 cause it doesn't work.
13:52<Soopaman>yeah, it seems that rtjpeg is the best bet for my machine
13:53<Soopaman>or mjpeg
13:54-!-mdz [] has joined #mythtv
13:55<mdz>why does the i18n directory create an executable?
13:56<PeteCool>any ideas why mythbackend would jsut die?
13:57<moegreen>PeteCool: is it seg faulting or just exiting?
13:57<PeteCool>pixie root # mythbackend
13:57<PeteCool>Starting up as the master server.
13:57<PeteCool>Probed: /dev/video0 - Television
13:57<PeteCool>Probed: /dev/video0 - Composite1
13:57<PeteCool>Probed: /dev/video0 - S-Video
13:57<PeteCool>Probed: /dev/video0 - Composite3
13:57<PeteCool>VIDIOCSPICT: : Invalid argument
13:57<PeteCool>VIDIOCSPICT: : Invalid argument
13:57<PeteCool>VIDIOCSPICT: : Invalid argument
13:57<PeteCool>Segmentation fault
13:58<PeteCool>I've been running with the VIDIOCS errors for over a week without problems
13:58<moegreen>PeteCool: have you ran it in gdb to get a backtrace?
13:58<PeteCool>not yet
13:58<PeteCool>I'm emerging gdb now
13:59<moegreen>did you recently update CVS?
13:59<mdz>mythbackend from 0.8 has been running continuously since 03/17 for me
14:00<moegreen>yeah, I haven't had any problems with it - PeteCool: if you did update CVS, make sure you have all the tables in your mysql database
14:01<PeteCool>I don't remember seeing a "U" besides mc.sql since last week
14:02<PeteCool>I just ran cvs.sql, and it still segfaults
14:03<moegreen>Make sure you have'CONFIG += debug' uncommented in your - if not your backtraces won't be very helpful
14:03<PeteCool>11 meg source for gdb, this'll take quite a while to compile :(
14:04<mdz>I thought gentoo had binary packages now
14:04<moegreen>PeteCool: isn't there a precompiled package for your distro?
14:04<mdz>because installing gdb from source is just silly
14:04<Chutt>mdz, same reason the themes dir makes an executible, it's just there to trick qmake
14:05<mdz>Chutt: surely there must be a better way
14:06<Chutt>probably, but i didn't see one quickly
14:06<Chutt>it is the simplest possible c program, so it should compile relatively quickly :p
14:08<vektor>his name is chutt
14:09<vektor>hey hey!
14:09<vektor>Chutt: how's it going?
14:09<Chutt>pretty good
14:09<vektor>been watching TV?
14:09<Chutt>was looking into stealing some of your cc code
14:09<vektor>i recommend
14:09<Chutt>since someone submitted a pretty halfassed patch to mythtv
14:09<Chutt>but i figure if i combine that and your stuff, it'll be good
14:10<vektor>i did some work today on freetype rendering stuff
14:10<vektor>we badly need work on cc rendering though
14:10<vektor>rendering is the difficult part
14:10<vektor>i rewrote my freetype code
14:10<vektor>we had some bugs that were causing badness with cc rendering i think
14:10<Chutt>you're still rendering at higher res, though, right?
14:10<vektor>i'm trying to get your code now
14:10<PeteCool>gentoo doesn
14:10<vektor>rendering at higher res?
14:10<Chutt>well, 4:4:4
14:10<mdz>vektor: E: Couldn't find package tvtime
14:10<PeteCool>gentoo doesn't really have precompiled packages
14:10<Chutt>instead of 4:2:0
14:10<mdz>vektor: :-(
14:11<vektor>mdz: dude, i know, simon has been slow, he's trying to be too 'correct' about it.
14:11<vektor>mdz: /msg coleSLAW and ask how it's going :)
14:11<mdz>vektor: tvtime: A high quality television application, 110 days in preparation.
14:11<Chutt>mdz, oh, got my zaurus yesterday =)
14:11<mdz>Chutt: nice, running OZ now?
14:11<Chutt>not yet, no
14:11<vektor>Chutt: we render at 4:4:4 then have a 4:4:4-to-4:2:2 composting routine.
14:11<vektor>Chutt: i think that's faster anyway
14:11<mdz>there are a few things that bug me about it
14:11<Chutt>i'm still waiting for it to charge fully
14:12<Chutt>manual said 24 hours, so
14:12<vektor>is the mythtv cvs somewhere new since i last got it?
14:12<mdz>but it's the best thing going I think
14:12<Chutt>vektor, probably
14:12<vektor>where is it now?
14:12<vektor>or stilly 'mythtv' ?
14:12<Chutt>same login info as before
14:12<vektor>what's the password?
14:12<Chutt>mythtv, i think
14:12<Chutt>i dunno, though
14:13<vektor>ok :)
14:13<Chutt>it's in the docs
14:13<vektor>i'm getting it happily now.
14:13<vektor>btw, the bounding box code was wrong
14:13<Chutt>i don't think i've touched the font rendering stuff
14:13<vektor>in the code you had before
14:13<Chutt>probably was
14:13<vektor>then i can probably just plug this right in
14:13<vektor>and you'll kick ass.
14:13<vektor>cvs tvtime uses dscaler dlls now.
14:14<vektor>for many filters.
14:14<Chutt>i updated a few days ago
14:14<vektor>cool :)
14:14<vektor>we also have code to not render when the window isn't visible. seems like a good idea.
14:15<vektor>have you been following the xfree86 political stuff?
14:15<vektor>damn, it's a blast :)
14:15<PeteCool>mdz: gdb compiled faster than MC with debug :)
14:15<Chutt>i don't think any of my linux boxes have usb enabled in the kernel
14:15<mdz>PeteCool: gdb is written in C
14:16<Chutt>mdz, if i totally fuck up a flash of the zaurus, is there a way to recover?
14:16<mdz>Chutt: depends on how totally
14:16<bigguy>vektor: on debian-devel or where?
14:16<mdz>if you were to, say, zero out the whole thing, you'd be fucked
14:16<mdz>unless it was still up and running
14:16<Chutt>so the code to reflash it lives in what it flashes?
14:16<Chutt>not in a rom somewhere
14:17<mdz>but there are two ways to flash it
14:17<mdz>you can either reflash the entire device, bootloader, flash util and everything
14:17<mdz>or just have it replace the kernel and initrd
14:17<mdz>I always do the latter because it's safer
14:17<Chutt>which is the oz way?
14:17<Chutt>seems that they just have a kernel and initrd image
14:18<bigguy>vektor: about the removal of Packard?
14:18<mdz>then it will do it the second way
14:18<bigguy>or the Open Governance thread?
14:18<Chutt>ok, cool
14:18<mdz>it does it based on what you put on the flash card
14:18<vektor>bigguy: the whole list has been interesting.
14:19<Chutt>need to get usb working, so i can talk to the thing first =)
14:19<Chutt>then i'll just wipe a cf card from my camera
14:19<mdz>I heard somebody got ejected from xfree86, but that's it
14:19<Chutt>you don't read slashdot like a good open source fanatic?
14:19<mdz>um, no
14:20<mdz>I go to slashdot when I want to see how a new browser renders a crazy table structure
14:20<Chutt>the upper right corner of the screen is darker than the rest of it
14:22<Chutt>moegreen, do you do an admin -kb of the .png files after you check them in?
14:22<Soopaman>vektor, how goes your project?
14:23<vektor>Soopaman: very well, you should try it out.
14:23<vektor>Chutt: I've been trying to catalog alot of performance problems and compatibility problems.
14:23<vektor>Chutt: some very interesting stuff on xvideo performance, btw.
14:24<Chutt>cool, got a page up somewhere?
14:24<Soopaman>vektor, url?
14:24<vektor>i have some things to add though that i've found out
14:24<Chutt>i heard there was some mythtv bashing in #livid the other day :p
14:25<moegreen>Chutt: no I didn't, but I should have
14:25<vektor>it wasn't bashing :)
14:26<Chutt>vektor, i can confirm #13 -- you apparently can't draw stuff over the Xv output unless it would cover it completely
14:26<Chutt>well, at least going by bug reports i get
14:27<vektor>i can't believe how much that one sucks
14:27<mdz>Chutt: I am fixing up the external channel change / file descriptor thing
14:27<Chutt>mdz, great, thanks
14:27<PeteCool>how do I bakctrace a specific thread? I forgot
14:28<Chutt>vektor, and, cool that you've got nvidia working on that video corruption
14:28<mdz>PeteCool: "thread <thread no>", then "bt"
14:28<vektor>with overscan?
14:28<PeteCool>the error is in EncoderLink::isConnected()
14:28<vektor>yeah, mark was really surprised that the bug exists :)
14:28<mdz>PeteCool: line number?
14:28<Chutt>well, with scaling, really
14:28<vektor>he says it exists with everything
14:28<mdz>that function really cannot segfault
14:29<mdz>unless your build is broken
14:29<PeteCool>mdz: 58
14:29<mdz>PeteCool: your build is broken
14:29<mdz>oh, 58 is the first line of the function. the object could be null or something
14:29<Chutt>every can checkout moegreen's new playback stuff now =)
14:29<PeteCool>mdz: and how could I unbreak it? I didn't really change anything
14:30<Chutt>everyone, rather
14:30<mdz>PeteCool: go up one level and see if the object being called is null or otherwise invalid
14:30<mdz>PeteCool: make distclean and a full rebuild usually unbreaks it
14:30<moegreen>Chutt: Should I put the PlayBoxTransparency option on the appearance setup page?
14:30<mdz>and removing any old libmyths
14:32<Chutt>near the 'theme qt widgets' option
14:37<Chutt>moegreen, buncha errors about rectTopLeft being undeclared
14:38<Chutt>well, similar variables to that, too =)
14:40<moegreen>Chutt: yeah, I just noticed that - I forgot I made some changes, I'm working on it now :)
14:42<Chutt>heh, np
14:46<Captain_Murdoch>anybody have any ideas on how to throttle cpu usage for the function I have that scans the recorded file for commercials? right now I am playing around with having it read /proc/stat to get cpu idle time and if < 10% idle it will sleep every X frames (X changes depending on idle time)
14:47<moegreen>Chutt: ok, the changes are in, let me know if you have any other problems
14:48<moegreen>I just tested it w/ 640x480 resolution and everything scaled nicely
14:48<Chutt>i just rebooted my dev box
14:48<Chutt>so, i'll check it out shortly
15:04<Chutt>moegreen, it only seems to want to display two lines of description text
15:05<Chutt>and it's possible to move the cursor off of the program subtitles =)
15:06<Chutt>that's it for bugs i've seen
15:08<moegreen>ok, got the description text show more lines
15:08<moegreen>err to show more lines
15:08<Captain_Murdoch>Chutt: do you have any problems with me reading /proc/stat for cpu idle time inside of NuppelVideoPlayer.cpp?
15:08<Chutt>that's fairly hacky
15:09<Captain_Murdoch>lol, thought so.
15:09<Captain_Murdoch>can you think of any better way for me to throttle the cpu usage of the nice-ed thread that is searching for commercials. I don't think it's good for it to go full-bore even if it is niced.
15:10<Captain_Murdoch>I can either try to intelligently throttle it or just have it sleep a little every frame so it uses little or no cpu time but that means it takes the full length of a show to process it for commercials.
15:11<Captain_Murdoch>if I don't throttle it, and you have back to back recordings it uses as much cpu as is available and can cause problems with the 2nd recording.
15:12<PeteCool>Captain_Murdoch: why not pause it every 5~10 frames? It wouldn't use full cpu, but would be shorter
15:13-!-thor_ [] has quit [Read error: 113 (No route to host)]
15:15-!-thor_ [~thor@] has joined #mythtv
15:17<Captain_Murdoch>well, even that could be too much. consider a machine where you only have 5-10% cpu free while recording (whether it's 1, 2, 3, etc. cards at the same time). I'll just eliminate the idle time calculation stuff and see how much it uses if I bump it up a frame at a time (ie, test it with sleeping every X frames, changing X for each compile/test cycle)
15:17<moegreen>Chutt: both of those fixes are in
15:18<PeteCool>if it's nice'd it'll still interfere with the recording (not freeing it's cpu usage fast enough?)
15:18<Captain_Murdoch>I've got it running right now on my Athlon 700 and it's recording and checking the previous show for commercials at the same time. still with 15% cpu free. 352x400 res w/ mpeg4
15:18<Captain_Murdoch>PeteCool: it's already niced and I think it causes problems with framedrops sometimes since even through it's niced it will use all available cpu.
15:19<thor_>Captain_Murdoch, what about nicing it and putting it on a timer?
15:19<Captain_Murdoch>you mean tying it to some fps rate?
15:20<Captain_Murdoch>or setting a timer so it only does X fps
15:20<thor_>the latter
15:20<thor_>Have a Qt timer fire every 200 ms or something, process a few frames, and then go away
15:22<Chutt>thor, the playlists patch you posted seems to be working nicely
15:22<thor_>Chutt, I haven't seen any hiccups, but have not really played with it much
15:23<Captain_Murdoch>I would think it's be more desirable to just sleep for Y seconds every X frames (and less complex as well)
15:23<Captain_Murdoch>Y milliseconds that is.
15:24<thor_>That's probably equivalent ... although I seem to remember that Qt timers will *not* fire if there's no idle left on the CPU
15:24<Chutt>thor, one minor thing
15:24<Chutt>when you create as a new playlist, could the current queue be replaced with a reference to the new list?
15:25<Chutt>instead of just keeping all the individual tracks in there
15:25<thor_>Uhm ... probably
15:25<PeteCool>mythbackend is still segfaulting :(
15:26<Captain_Murdoch>hmm, maybe I'll look into that then, thanks.
15:26<Chutt>and could there be a 'clear current queue' button in the popup for it?
15:26<Chutt>that'd clear out the current playqueue list, and also clear the (previous list) status
15:27<Chutt>those are the only two things i'd like to see before it goes into cvs =)
15:27<Chutt>and, really, this could all go in now, if you'd want
15:28<thor_>No ... let me make those changes and improve some of the places where things could easily be sped up ... then send you a tarball for final sanity check ... later tonight ... tomorrow
15:29<thor_>One other thing
15:30<thor_>if the Current Queue is full of stuff and you copy over from a playlist, you lose all that info ... was thining about storing what's in the queue when you push on a playlist so that when it pops off you go back to the way the Active Queue was before... thoughts"
15:30<PeteCool>my segfault shouldn't be coming from a locked device or something, since tvtime is able to read it alright
15:31<Chutt>yeah, that'd work
15:31<PeteCool>mdz: could you explain how to do " go up one level and see if the object being called is null or otherwise invalid" please?
15:31<thor_>may even persists if (default_playlist_pushed_storage or something)
15:31<thor_>persist it
15:31<thor_>using a different keyboard (lots of typos)
15:34<mdz>PeteCool: that would require too many prerequisite explanations I think
15:35<mdz>it sounds like either your build or your database is broken
16:02<PeteCool>one feature I'd like to have is to manually override one channel and anywhere recordings... sometimes I already saw it, sometimes I want to use the comp for something else
16:02<PeteCool>currently it removes the channel or anywhere status
16:04<Captain_Murdoch>PeteCool: just go into the playback screen and delete the inprogress recording. that will stop just that one instance of it and it will record next time it comes on.
16:05<Chutt>moegreen, that's something, does the new code mark in-progress recordings in red like the old stuff?
16:06<moegreen>Chutt: that a flag in the ProgramInfo class?
16:09<Chutt>i think conflicting is set to true
16:10<moegreen>ok, that should be easy to add - I'll do that a little later
16:12-!-nyquiljer [] has joined #mythtv
16:18<Captain_Murdoch>in the new playbackbox code, I cursor right to go to the individual show listing on an item where there's only one show and I can cursor down off the 1 show. if I cursor down far enough, a scroll up arrow is drawn in the window on the right.
16:19<Chutt>captain_murdoch, he just fixed that a little bit ago
16:19<moegreen>Captain_Murdoch: I updated the CVS shortly after the initial import and that should hvae been fixed
16:19<Captain_Murdoch>ok, must have been after I updated cvs. sorry.
16:19<Chutt>moegreen, don't quote me on that confliting being set to true or not =)
16:19<Chutt>i don't remember what i did
16:21<Captain_Murdoch>btw, nice job moegreen
16:21<mdz>Chutt: I believe it is conflicting
16:25<moegreen>if (pginfo->conflicting) _cg.setColor(QColorGroup::Text, Qt::red);
16:25<Chutt>moegreen, also, does it keep the same position if filllist is called?
16:25<mdz>it wouldn't be such a bad idea to add a flag for that rather than abusing conflicting
16:27<moegreen>Chutt: Yeah, it should keep the position. I have it reset the position after a delete though
16:27<Chutt>reset everything?
16:28<Chutt>or just the current position in that program's section
16:28<moegreen>currently it resets everything (goes back to all programs, top of the list)
16:28<Chutt>you reverted the i18n stuff, btw
16:28<rmj>chutt:I've got something to submit for the deinterlacing, its not as good as I'd hoped but after talking with vektor just now getting anything better is going to be a fair bit harder so I'd like to submit what I have now and investigate better options. Its a quality tradeoff at the moment, for people with progressive displays (or weird cloned setups like me) its probably worth having, you'll probably want to stick with the default mode thoug
16:29<Chutt>rmj, i was testing on my monitor, before
16:30<Chutt>moegreen, any way to make it not do that? =)
16:30<moegreen>probably - i just wasn't being very imaginative :)
16:30<rmj>yup, I've done some more testing here, I've noticed doubling my bit rate helps, the issue (I believe) is the
16:31<rmj>field jumping is chewing up lots of bits, couple this with the fact xvideo is scaling the half height image, the resulting mpeg4 artefacts are twice as visible.
16:32<Chutt>makes sense
16:32<Chutt>did you fix rtjpeg?
16:32<rmj>(I've fixed it so it works with RTJpeg again also). Overall I still think its useful for users who need deinterlacing but linear blend doesn't cut it for them (linear blend is only really useful on very static material).
16:33<Chutt>yeah, send in a patch
16:33<Chutt>i'll get it applied
16:34<rmj>anyways its not on by default and it doesn't change too many code paths (I made one tiny tweak to the timing code in the player where there was an assumption that frame time was 40ms irrespective of fps in the file).
16:35<Chutt>well, assuming my box boots after this kernel recompile
16:35<rmj>I'll post the patch to -dev coz I wan't to send in some explanations on how to make it useful (like the mpeg4 bitrate issue). I've also included some modes which output full frames at 50fps but my CPU 1800 athlon is probably 20% too slow to keep up with them. I've also added in the nuppel area based deinterlacer which gives decent output at 25fps for sports material.
16:37<rmj>I'll have to think about how to get the next level of quality out of the thing, tvtime is very careful with the way it times output. There is also an issue with sending the half height fields with xvideo as you currently can't tell xvideo which is the top and bottom fields so it cant scale them appropiate (pushing the bottom field down half a scanline) [ vektor pointed this out to me], apparently directx can do this but xvideo can't yet.
16:38<rmj>I'm planning to look at vektor's directfb tv output code with a view to making directfb an output option for mythtv.
16:38<PeteCool>moegreen: going into "watch recordings" crashes mythtv for me
16:38<Chutt>you updated with cvs -z3 update -PAd
16:38<Chutt>and did a make distclean?
16:38<rmj>I think I may need to upgrade to a 5000 Athlon to acheive some of what I want though ;-)
16:39<PeteCool>-PAd ?
16:39<Chutt>guess not.
16:39<Chutt>gotta specifically tell cvs to download new directories
16:39<Chutt>else, it doesn't
16:40<Chutt>and he added some new directories.
16:44<Captain_Murdoch>does the new playbackbox allow deleting videos from the playback screen or just from the delete screen? 'D' on the playback screen isn't doing anything here.
16:45<Ndiin>Yeah, I was just looking at that code..
16:45<Ndiin>P and D aren't bound
16:46<Ndiin>It can easily be fixed though
16:47<Ndiin>Just change the prototype of selected() to selected(int blah=0), make the slot emit 1 for Key_P, 2 of Key_D, check the int before the switch, etc.
16:47<moegreen>i wasn't aware of those key bindings
16:48<Captain_Murdoch>P & D work from both screens, P always plays, D always deletes
16:48<Captain_Murdoch>"worked" that is. :)
16:57<mdz>is the tree view in mythmusic a stock tree widget?
16:58<mdz>it doesn't accept enter as a synonym for OK like the rest of myth
16:58<Chutt>i thought i had fixed that
17:00<mdz>likewise for the buttons on the playback screen
17:00<Chutt>those i know are broken
17:00<inman>yeah, petecool was complaining about that a couple weeks ago.
17:01<inman>i looked at it, but i don't see the problem.
17:01<Chutt>the qtoolbutton stuff treats enter differently
17:01<inman>i know, but that seems like a bug to me.
17:01<mdz>can it be fixed in qtoolbutton, or do we need to derive a new one?
17:02<Chutt>we're already deriving a new button class for it
17:02<inman>i tried adding enter as an accel but i think i was only able to get one button to accelerate even given focus changes.
17:02<inman>what do we lose by using Button versus ToolButton, though?
17:02<mdz>ah, I see
17:02<Chutt>inman, looks diferent
17:02<inman>that is easily fixed, isn't it?
17:03<inman>toolbutton is derived of button; we just gotta do some extra legwork.
17:03<moegreen>Chutt: I'm not seeing 'conflicting' being set for programs that are being currently being recorded
17:04<Chutt>did i set it to conflicting in the old filllists?
17:04<moegreen>hmmm...I'll have to look
17:05<Ndiin> [ "conflicting" not found ]
17:05<Ndiin>in old playbackbox.cpp
17:05<mdz>Chutt: do you know if toggle() is the right slot for the toolbutton?
17:05<Chutt>like i said, i dunno what i had it doing
17:05<Chutt>mdz, for what?
17:05<Chutt>it getting selected?
17:05<mdz>what space does
17:05<mdz>and what enter should do
17:06<Chutt>not right off, no
17:06<inman>mdz: depends on the button
17:06<moegreen>heh, I see, it was being set in the ProgramListItem class - heh
17:06<inman>some buttons have state, some don't...
17:06<Chutt>moegreen, yeah, that's right
17:07<mdz>inman: for qtoolbutton of course
17:08<-- nyquiljer( has left #mythtv
17:09<inman>mdz: i'm not sure what you're asking.
17:09<PeteCool>Chutt: I did a cvs update -PAd but the new playbackbox still crashes... weird
17:10<PeteCool>with make distclean of course
17:10<Chutt>so, get a backtrace...
17:12<PeteCool>OT: does a binary with debug symbols run noticeably slower?
17:12<mdz>inman: qtoolbutton has an action that it takes when space is pressed. mythtoolbutton should take that same action when enter is pressed.
17:12<Chutt>of course it does
17:13<mdz>the question is whether it uses qbutton::clicked(), or toggle()
17:13<Chutt>probably clicked
17:13<Chutt>i think
17:13<Chutt>i dunno
17:15<Chutt>should be pretty easy to test if it's right, though
17:15<Chutt>either the pause button will stay depressed, or it won't =)
17:15<mdz>it looks like qtoolbutton doesn't override the keypress event handler, so it must be the same as for qbutton
17:16<inman>right, that's about where i stopped looking.
17:16<inman>should the delete-recordings screen update the disk-used after each deletion?
17:16<PeteCool>I think yes
17:17<PeteCool>I delete recordings to free space
17:17<Chutt>it does
17:17<Chutt>not immediately, though
17:17<Chutt>unless moegreen's not got that hooked up
17:17<mdz>but ah, in qbutton, enter -> clicked() and space -> pressed()
17:18<inman>so we just need to bind to both signals?
17:18<moegreen>it gets updated here
17:18<mdz>I'll just have enter send pressed()
17:18<mdz>and leave everything else the same
17:18<moegreen>right away :)
17:19<PeteCool>anybody tried >=mysql-4.0 yet? I'll be doing so soon, I think
17:20<Ndiin>eeek, thor, What's D bound to in your playlist patch?
17:20<Ndiin>because "I", jumps to the I's, "D", deletes the item..
17:21<Ndiin>from the active queue
17:24<PeteCool>is it actually possible to scroll the list to jump to a particular song?
17:24<PeteCool>I didn't see that previously
17:24<Ndiin>the treelist widget does that, i think
17:24<Ndiin>but he's rebinding a few keys, as well
17:24<PeteCool>not when editing the playlist, but during playback
17:24<Ndiin>oh, no =(
17:25<Ndiin>Not yet, at least.. ;)
17:25<Chutt>hit next a bunch.
17:25<Ndiin>That crashes it..
17:25<Ndiin>next repeatedly, that is
17:25<Chutt>nice that people report bugs like that.
17:25<Chutt>since it obviously doesn't do that for me.
17:26<Ndiin>I haven't reported it because I haven't looked into it yet
17:26<PeteCool>It never crashed for me
17:26<PeteCool>repeatedly = how many times?
17:27<Ndiin>Like 10 times in less than a second, or so.
17:29<inman>the epg seems to include channels from alternate inputs. i don't think it should.
17:29<Chutt>sure it should
17:30<Chutt>what if you just want to go mark something for being recorded?
17:30<Ndiin>thor_: hmm, Enter and/or Space don't close the Text entry box for playlist names
17:30<PeteCool>Chutt: the "watch recordings" crash doesn't happen in gdb
17:30<inman>what if you have two inputs? one sat and one cable.
17:30<PeteCool>I just have two windows with empty theme backgrounds
17:30<Chutt>petecool, that indicates that you didn't build it properly
17:30<PeteCool>but it still doesn't work
17:30<Chutt>and didn't install the new images.
17:31<PeteCool>the first time it wasn't a debug build
17:31<PeteCool>ok, let me check that
17:31<moegreen>PeteCool: the images should be in /usr/local/share/mythtv/themes/default/
17:31<moegreen>maybe slightly different if your install location is different
17:32<PeteCool>I do have those
17:32<PeteCool>uparrow, filler, etc
17:32<Chutt>mdz, the alsa maintainer needs to get his act together.
17:33<inman>thanks for the p->play d->delete bindings, moegreen.
17:33<Chutt>i like how his source packages just refuse to compile
17:36<moegreen>PeteCool: so you're saying there is no text being written to your screen?
17:36<PeteCool>moegreen: exactly
17:36<Chutt>and no images.
17:36<PeteCool>moegreen: but I just dumped my CVS dir, starting from nothing just to be really sure that's not the problem
17:37<PeteCool>Chutt: only iulius's background, yes
17:39<moegreen>PeteCool: are you using the solids background or the default transparent backgrounds
17:39<PeteCool>moegreen: transparent
17:39<moegreen>at line 206 add cout << transShowing << endl;
17:39<moegreen>(to playbackbox.cpp)
17:49-!-Edgan [] has quit ["Client Exiting"]
17:49-!-inman [] has quit ["Client exiting"]
17:56-!-Soopaman [] has quit [Read error: 104 (Connection reset by peer)]
17:57-!-Soopaman [] has joined #mythtv
18:02-!-Edgan [] has joined #mythtv
18:02<PeteCool>moegreen: I'm getting this: /usr/share/mythtv/themes/default/trans-showings.png
18:03<PeteCool>moegreen: but, even with "transparent boxes" unchecked it crashes
18:08<Ndiin>But it doesn't happen through gdb?
18:10<PeteCool>even with a release build, in gdb mythfrontend doesn't crash
18:11<PeteCool>I do get this with both release and debug though:
18:12<PeteCool>0x405a64fd in QString::QString(QString const&) () from /usr/qt/3/lib/
18:13<PeteCool>moegreen: did you handle accented strings in your code?
18:13<PeteCool>moegreen: those: \xE0\xE9\xEA\xE8
18:14<PeteCool>it *might* be that
18:16<Chutt>he doesn't need to handle them
18:16<PeteCool>or it might be because I have no shows yet
18:16<Chutt>qt does automatically
18:17<Chutt>time for kernel recompile #6
18:17<PeteCool>it doesn't handle empty recorded shows list
18:18<PeteCool>now it's fine
18:18<PeteCool>and cool, too
18:18<moegreen>hmmm...i *guess* I could check for the empty recording list :)
18:18<bigguy>Chutt: what'd you miss in the last 5 recompiles?
18:18<Chutt>randomass stuff
18:18<Chutt>trying to get the zaurus talking to the computer properly
18:19<Ndiin>ccache is your friend ;)
18:19<moegreen>Anyone else notice anything missing in the playbackbox stuff?
18:19<Chutt>that should be it
18:19-!-merf [] has joined #mythtv
18:19<merf>can someone help me with an odd mythepg bug?
18:19<Chutt>as long as it doesn't lock up on remote preview playback
18:20<moegreen>does it now?
18:20<Chutt>i dunno, can't test it at the moment
18:20<merf>for some odd reason, every now and then (actually, pretty often) mythepg doesn't display any show information
18:20<merf>apparently it's either not connecting to mysql or can't get the data for some reason
18:20-!-nyquiljer [] has joined #mythtv
18:20<moegreen>Chutt: I've only got one machine so I guess I can't test that, right?
18:21<Chutt>you can sort of test it by going into mythbackend/mainserver.cpp
18:21<nyquiljer>anyone have any opinions on what the best remote solution is, assuming i have a tuner card with no remote? i'm looking thru but tehres so many options.
18:21<merf>nyquiljer: the cheapest :)
18:21<Chutt>and comment out the three lines starting with 424
18:21<moegreen>nyquiljer: I just use a Sony DirectTV remote that I had lying around
18:22<Chutt>where it decides if the thing's local or not
18:22<nyquiljer>moegreen: well, i mean the otehr end actually, the recieiver
18:22<moegreen>Chutt: which pathname to keep? mythtv:// ?
18:23<nyquiljer>moegreen: i want to be able to use one of the hundred of remotes thats laying around
18:24<rmj>chutt: forgot to mention a detail when I sent that pactch in, I haven't tried to make this work for hardware mpeg (as I cant test it), I've tried to leave it alone so it shouldnt affect it with deinterlace set to default (DEINTERLACER_NONE) but I'm not sure what will happen to hardware mpeg capture if deinterlacing is enabled.
18:24<moegreen>nyquiljer: i got this a while ago:, seems to work fine for this remote, I would have to assume others
18:25<Chutt>only silly people are using that, anyway
18:25<merf>does mythtv 0.8 use settings.txt or mysql to store its recording settings?
18:25<moegreen>Chutt: I see the problem with the missing subtitle information now
18:25<nyquiljer>moegreen: excellent, thanks very much.
18:25<moegreen>merf: mysql
18:25<Chutt>it went away for me with the version you checked in =)
18:26<merf>presumably if it couldn't connect to mysql it'd die, right?
18:26<moegreen>Chutt: yeah, I didn't see it when I have it running locally ... odd?
18:26<Chutt>i dunno
18:26* merfcurses mythepg
18:27<nyquiljer>moegreen: how to you have the sensor portion of it mounted? does it reliably get every button press you send it?
18:28<moegreen>nyquiljer: I have high ceilings so I have it up about 12 feet on the wall, gets every button I press from a wide range of angles
18:28-!-inman [] has joined #mythtv
18:29<moegreen>nyquiljer: that one comes w/ a few feet of cable and on the end is the small receiver
18:29<merf>my reciever is sticking out of a shoddy hole drilled in a drive bay cover :D
18:29<nyquiljer>moegreen: nice, that will work very well for me
18:32<merf>has anyone in here gotten snes9x to work with fullscreen?
18:48<PeteCool>moegreen: Enter doesn't seem to start playback in the new playbackbox
18:49<PeteCool>moegreen: it did beofre
18:51<PeteCool>only pause does
18:52<Captain_Murdoch>pause meaning space right?
18:53<PeteCool>eh, yes
18:53<moegreen>ok, i've got enter added, I'll commit it shortly
18:56<PeteCool>Captain_Murdoch: you're the one doing the ad skip code right?
18:56<Captain_Murdoch>sorry, pause == p, space also works.
18:57<PeteCool>Captain_Murdoch: did you ever those ads that are a black screen with a few white words in the middle?
18:58<PeteCool>Captain_Murdoch: those aren't skipped currently
19:00<inman>my problem with deleting programs after viewing them seems to be gone.
19:02<PeteCool>moegreen: great work :)
19:03<Captain_Murdoch>yeah. they get detected as blank frames, but shouldn't get marked as commercials because it checks time lengths between frames.
19:09<Captain_Murdoch>pete, moegreen, just noticed a bug where I deleted a program (3rd from the top I think) in "all programs" and when it redrew, it listed the top item (a currently recording item) twice.
19:09<Captain_Murdoch>oops., meant just moegreen on that.
19:11<moegreen>Captain_Murdoch: indeed ... I see that as well
19:11-!-choenig [] has quit [Read error: 104 (Connection reset by peer)]
19:12<Captain_Murdoch>also getting segfaults around line 655 concerning i == curTitle after doing a delete occasionally.
19:12<Captain_Murdoch>in playbackbox.cpp, occurred first around 651 actually, then on 655 with the !=
19:15<PeteCool>moegreen: I think the lower portion of the screen should be blank when you're choosing between program titles (all programs or something else)
19:22<Captain_Murdoch>moegreen: more debugging info. it segfaulted on 655 again with h = -2, i = -3, and curTitle = 0. I'm deleting from the all programs list if that makes any difference.
19:23<Captain_Murdoch>I think it may have to do with deleting the last episode of a show
19:25<Captain_Murdoch>don't have any more shows left to delete to test but I'm recording now and have a bunch of stuff setup tonight to record that I can delete after testing some other stuff.
19:27<moegreen>alright, i'm looking into it - thanks for the info
19:28<Chutt>stupid zaurus
19:28<Captain_Murdoch>ok, I can put more debug statements in later if need be once I've got more stuff to delete.
19:29<PeteCool>Chutt: doesn't work?
19:29<Chutt>works, but i'm having problems talking to it
20:11<mdz>Chutt: the alsa maintainer got preempted...there are several people working together on it now
20:11<mdz>the packages are in a lot better shape than they were 6 months ago
20:13<Chutt>current unstable doesn't compile at all
20:13<Chutt>well, it does if you go in and manually configure the source
20:13<Chutt>then manually do a make clean
20:13<Chutt>then run make-kpkg modules_image two or three times
20:17<Chutt>obviously, things will be better with 2.5
20:18<PeteCool>did you try 2.5 yet?
20:19<PeteCool>it works okay here on some of my comps
20:19-!-schwin97 [] has quit ["Client exiting"]
20:21<Chutt>i use it every now and then
20:21<Chutt>haven't put it on my main box recently, though
20:21<Chutt>need it to work too much to waste time futzing with unstable kernels
20:38-!-nziarek [] has joined #MythTV
20:38<nziarek>meogreen: saw that last commit -- thanks :)
20:39<moegreen>heh, np :)
20:39<nziarek>I am in the process of reinstalling Mandrake, so I can't play, but how theme able is it?
20:39<moegreen>it's not right now, but I'll be working on that
20:40<nziarek>sweet :) can't wait to see it in action - looks great, good work
20:40<moegreen>eventually you will be able to define all the graphics and the placement of everything on the screen
20:41<nziarek>great. wish this isntall would hurry
20:44<PeteCool>moegreen: your news screens look great in SasQuatch
20:50<moegreen>PeteCool: yeah? I think the text should be a different color - or the dropshadow needs to be extended some
20:50<moegreen>then again - maybe it's just my TV's brightness/contrast settings
20:50<nziarek>what are "news screens"
20:51<moegreen>nziarek: I assume he's referring to the Watch Recordings and Delete Recordings screens
20:52<nziarek>oh - thought I had missed MythNews on my break :)
20:53<PeteCool>nziarek: yeah news should have been "new" here
20:54<nziarek>Sorry - wasn't trying to be nit picky, I really did think I missed a new Module or something. I came back to 400 messages and didn't read them all :)
20:55<moegreen> does show some promise for a news modules
20:55<moegreen>err module
20:56<nziarek>yeah, I love that site. if they had local, I'd never go anywhere else
20:58<Chutt>welp, that didn't work
21:10<Chutt>md5's don't match the oz site
21:19-!-SoopaMech [] has joined #mythtv
21:24-!-Soopaman [] has quit [Read error: 60 (Operation timed out)]
21:24-!-SoopaMech is now known as Soopaman
21:28-!-nziarek [] has quit [Read error: 110 (Connection timed out)]
21:54<Chutt>how was my message to that dumbass not civil?
21:58<vektor>on the list? :)
21:58<vektor>damn i miss uncivil messages :)
21:58<Chutt>he said 'I wasn't trying to record, just view tuner output.'
21:58<Chutt>among other things
21:58<Chutt>so in my reply, i say 'Watching tv is recording.'
21:58-!-NonToxic [] has joined #mythtv
21:59<Chutt>apparently, that wasn't civil
22:00<NonToxic>grumble... can mythtv use an older version's DB?
22:00<Chutt>no, it needs to be upgraded
22:00<NonToxic>uhm, where is mysql's database stored?
22:00* NonToxicneeds to do some RM'ing
22:01<Chutt>why delete it?
22:01<NonToxic>i've fubarred stuff...
22:01<NonToxic>I don't know enough about SQL to repair it, so I'm starting from scratch
22:01<Chutt>just run the 'mysql' command line client and 'drop database mythconverg;'
22:01<NonToxic>er, ok
22:01<Chutt>or something close to that
22:02<NonToxic>ok, there we go
22:02<Chutt>mdz, so, how the hell does one exit vi with this keyboard?
22:02<NonToxic>:wq :)
22:02<Chutt>on a zaurus
22:02<Chutt>need to hit escape, no escape button, escape button on the on screen keyboard doesn't work
22:02<NonToxic>pop the onscreen keyboard
22:02<Chutt>yeah =)
22:02<NonToxic>that's weird...
22:03<NonToxic>zaurus or openzaurus?
22:03<NonToxic>use oz...
22:03<Chutt>i'm in the process of installing that
22:03<Chutt>it'd be _nice_ to edit resolv.conf, is all
22:03* NonToxicis running Familiar on his ipaq with opie instead of qpe
22:03<NonToxic>oh, that,.
22:03<NonToxic>uhm, don't you ahve a serial console you could be doing this over?
22:04<Chutt>well, that would involve turning on telnetd
22:04<Chutt>which would involve editing inetd.conf, and that brings me back to how to exit vi =)
22:04<Chutt>ah well, i'll not need to in a few minutes
22:04<NonToxic>er, serial console.
22:05<NonToxic>not telnet...
22:05<NonToxic>or does the zaurus not have serial?
22:05<Chutt>i don't have a serial cable, no
22:05<Chutt>i'm really just waiting for a cf card to finish emptying
22:05<Chutt>then i'll just transfer the oz stuff to that and reflash
22:06<NonToxic>hmmm, I'm having problems with ./setup
22:06<NonToxic>it doesn't seem to be able to get a provider list...
22:06<NonToxic>tis a server down or something?
22:07<Chutt>not that i'm aware of
22:07<NonToxic>zap2it gave us a page with no service provider options
22:07<Chutt>running xmltv 5.8?
22:07<NonToxic>uhm, likely not
22:08<NonToxic>where do I get this xmltv?
22:10<NonToxic>god I hate sforge's download system
22:14<mdz>Chutt: the Cancel key is escape
22:14<Chutt>mdz, you're not on -users, still are you?
22:15<mdz>Chutt: I never was
22:23<Chutt>mdz, broke my z once already =)
22:23<Chutt>had to reflash with the official oskit stuff
22:26<PeteCool>Chutt: to save and exit vim try ZZ
22:26<PeteCool>that's the only key set I know :)
22:26<Chutt>i use vim for all my dev work
22:27<PeteCool>can't you caps lock the on-screen keyboard?
22:27<Chutt>mdz, and whose idea was it to make ifdown usb0 also ifdown eth0 while it was at it?
22:27<PeteCool>then it'd be easier than :wq
22:27<Chutt>it's rather annoying =)
22:30* NonToxic<3 screenscrapers
22:54<NonToxic>looks good,b ut the gui is horribly laggy...
22:54<NonToxic>I saw that both in freevo and mythtv.
22:54<Chutt>lower the res it's running at
22:55<NonToxic>eh, how would I do that?
22:55<Chutt>use the arrow keys to modify the resolution settings
22:55<Chutt>by default it resizes to full-screen
22:56<-- nyquiljer( has left #mythtv
22:56<Chutt>and since qt uses the render extension to alpha blend pngs like in the menu, it's slow
22:56<Chutt>it's quite snappy at normal tv-out resolutions, and that's all i really care about =)
22:56<moegreen>NonToxic: is there any one part of the interface or is it overall slow?
22:57<NonToxic>moegreen: the channel selection, mostly
22:57<NonToxic>Chutt: hm... rewind and fast forward aren't as fast as they used to be..
22:57<moegreen>NonToxic: The program guide or while watching live TV?
22:57<Chutt>that bit's slow because of the category colors getting blended onto the background
22:58<NonToxic>moegreen: program guide is slow at 1024x768, watching live TV also tends to be a little bit slow in the controls
22:59<moegreen>NonToxic: if you're running the CVS version you can disable the alpha blending in the program guide
23:00<Chutt>the guy that just emailed (twice) about voice control
23:00<Chutt>emailed the freevo list a couple days ago
23:00<Chutt>basically the exact same thing
23:01<NonToxic>moegreen: eh... I actually mostly don't use the program guide, I gotta watch TV and that's it :)
23:01<moegreen>at first I thought he was drumming up support for help on his senior project
23:01<NonToxic>moegreen: so running it at 800x600 is fine
23:01<NonToxic>Chutt: some non-coder demanding voice control?
23:01<Chutt>no, someone working on a school project
23:02<Chutt>i'm not really sure what he wants from me or anything
23:02<NonToxic>moegreen: only mythbackend has to be running for it to record, right?
23:02<moegreen>that's correct
23:03* NonToxicwrites an initscript for it
23:03<NonToxic>can mythbackend be run in daemon-mode?
23:03<Chutt>there's a command line option for it to daemonize itself
23:03<NonToxic>which is?
23:04<NonToxic>wiseguy586@bluefire setup $ mythbackend -?
23:04<NonToxic>Invalid argument: -?
23:04<Chutt>i dunno
23:04<NonToxic>-h, -?, --help, all don't seem to work
23:04<Chutt>i don't think there's a help command =)
23:04<Ndiin> } else if (!strcmp(a.argv()[argpos],"-d") ||
23:04<Ndiin> !strcmp(a.argv()[argpos],"--daemon")) {
23:05<Chutt>there's also a -l <filename> arg to log output to a file
23:05<NonToxic>ah, it appears to be -d
23:05<Ndiin>-p or --pidfile for pidfile output, as well, I see
23:05<NonToxic>Ndiin: ah
23:07<Chutt>time to add someone else to the banned from posting to my mailing list list
23:08<moegreen>Chutt: what's the infraction this time?
23:08<Chutt>moegreen, read the most recent post =)
23:09<NonToxic>are there archives?
23:09<Chutt>two sets, actually
23:09<Chutt>both linked to from the webpage
23:13<NonToxic>Chutt: this on -dev, or -users?
23:13<Chutt>"Jay <>"
23:14<Chutt>now if only mailman's regexp ban_list stuff would work.
23:15* NonToxicnow has an initscript for myth
23:16<Chutt>just make sure that it'll restart the backend in case it goes dead
23:17<NonToxic>my machine isn't on all the time, so it's not a worry
23:19<poptix>Chutt: heh.
23:21<Chutt>poptix, i almost ordered a wintv-pvr this morning
23:22<Chutt>but i figured i just spent enough on a zaurus, so, didn't =)
23:31<Ndiin>hmm, Anyone know the sql syntax offhand to delete a specific entry? Say where chanid=1004?
23:33<moegreen>Ndiin: delete from <table> where ... I think
23:33<Ndiin>sounds right..
23:33* Ndiintries
23:33<Ndiin>yep, thanks
23:35* Ndiindoes some channel housekeeping
23:36<Ndiin>There, much better
23:41<Chutt>moegreen, heh, nice one line change to playbackbox.cpp
23:41<Ndiin>wow, weird.. just tried mythbackend's -l stuff.. it defaulted the permissions on it to 044, with set GID
23:41<Ndiin>no, SUID, rather, not gid
23:41<Chutt>looks like it opens it at 0644
23:42<Ndiin>---sr--r-- 1 root root 100 Mar 23 21:39 /var/log/mythbackend.log*
23:42<Chutt>in the code
23:43<mdz>Chutt: dunno what you mean about usb0 and eth0, works fine here
23:43<Ndiin>0664 in the code.. strange.
23:43<_shad>15 hours of work sucks
23:43<moegreen>Chutt: yeah, I figured I fire some credit Nathan's way :)
23:43<mdz>Chutt: are you using usbdnet or usbnet?
23:43<Chutt>mdz, whenever i unplugged the zaurus from the cradle, eth0 would go away
23:44<m0tion>anyone here have any experience with the PerlQt library by any chance?
23:45<mdz>Ndiin: you're probably running old code
23:45<mdz>Ndiin: I fixed that a bit after the 0.8 release
23:45-!-bigguy [] has quit [Read error: 113 (No route to host)]
23:45<mdz>in CVS and in the debs
23:48<Ndiin>mdz: yep, the code I was looking at was a fresh co, just checked what was last installed ;)
23:52<Chutt>mdz, think you'll have time to do the recording profile stuff anytime soon/
23:57<mdz>Chutt: maybe, I'm still a bit undecided about how it should look nda work
23:59<Chutt>i think a list of scheduled recordings
23:59<Chutt>that when selected pops up a second screen of settings should work
23:59<Chutt>let you select the recording profile from a combo box