Back to Home / #mythtv / 2003 / 03 / Prev Day | Next Day
#mythtv IRC Logs for 2003-03-07

00:00-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
00:27<rkulagow>mdz, are you here?
00:27<mdz>rkulagow: yeah, but I'm going to bed right now
00:28<rkulagow>ok, NP. wanted to talk to you about the pointer you sent me earlier when i was talking to chutt about the two columns for the mythmusic setup program i'm doing. visualizations is the last thing, and i'm stuck, and my scrollback is gone. we can talk in the morning.
00:34-!-PeteCool [] has joined #mythtv
00:38-!-PeteCool [] has quit [Client Quit]
00:50<rkulagow>chutt: are you back?
00:50<Chutt>going to bed soon
00:52<rkulagow>ok; don't let this keep you up. i'm working on those visualizations. got them into two columns. i'm trying to figure out how to work with the checkboxes without actually storing the data to the database, because that's not needed.
00:52<rkulagow>i thought i could assign a variable to the result of each checkbox status, then build a string, and write _that_ to the database.
00:53<rkulagow>but i'm stuck, and now i think i'm just breaking stuff.
00:53<rkulagow>mdz had mentioned something earlier, but it scrolled off the top of my scrollback, and i didn't note what he pointed me to.
00:54<rkulagow>anyway, everything else is done, other than that.
00:54<Chutt>i think it was overriding the save() function, like the xmltv stuff does
00:55<rkulagow>ok, i'll check there then. thanks.
00:58-!-planetjay [] has quit [Read error: 60 (Operation timed out)]
01:07-!-nick_hudson [] has joined #mythtv
01:12<nick_hudson>anyone around?
01:15<Captain_Murdoch>not for long..
01:16<nick_hudson>i just installed mythtv and tried recording a show and is there anyway that i can make the file size smaller?
01:16<Captain_Murdoch>you mean all the time or just that 1 file?
01:16<nick_hudson>a 30 minute show was around 3.5 gigs
01:17<Chutt>perhaps you might want to, i dunno, adjust the encoding parameters?
01:17<Captain_Murdoch>what speed cpu do you have?
01:17<nick_hudson>I thought about trying a lower resolution but I wanted to ask before I did that
01:17<nick_hudson>AMB Xp 1800
01:17<Captain_Murdoch>go into setup and setup a recording profile. try changing to mpeg4 instead of rtjpeg. you can bump the resolution to 480x480 if you want as well.
01:18<nick_hudson>ok I have it set already to 480x480 .. but wasnt sure about mpeg4
01:18<Captain_Murdoch>Chutt, I haven't found any scaling stuff in libavcodec that I think I could use. I'm thinking that if I scale the yuv data instead of the rgb data that I'll save 25% of time right off the bat.
01:19<Captain_Murdoch>nick_hudson: I'm recording at 352x480 and take about 900Megs per hour using mpeg4. I'm on a lowly Athlon 700 though that's why the lower resolution.
01:19<Chutt>can't use the img_resample stuff looked like it could work for me
01:19<Captain_Murdoch>also make sure sound is being encoded as mp3 not raw.
01:19<Chutt>err, two different sentences there, but you get the idea
01:20<nick_hudson>ok cool I got it changed ... thx for the help
01:20<Captain_Murdoch>I don't see a depth in there anywhere, maybe I'm missing something. I guess that's yuv though so I wouldn't see a depth?
01:20<nick_hudson>I just thought that mpeg4 would have a larger file size maybe I read that wrong somewhere
01:21<Chutt>the default encoding parameters are set for fairly low compression like that because otherwise i'd get even more people bitching about their slow machines not being able to handle things
01:21<Chutt>captain_murdoch, that's taken care of in the avpicture struct, no?
01:22<Chutt>but, yeah, i believe it only works on yuv data
01:22<Captain_Murdoch>I'll play around with it. I've never messed with libavcodec any except for trying to track down that buffer bug a couple weeks ago. :)
01:23<Captain_Murdoch>scaling the yuv is faster anyway, only 75% of the size of full rgb at 16bit.
01:23-!-Edgan [] has joined #mythtv
01:23<Captain_Murdoch>I see their sample main() at the bottom of imgresample.c so I'll play around with that.
01:24<Captain_Murdoch>I need to setup a 2nd box bad... :) I can't compile anything cause I'm recoring something.. lol
01:24<Captain_Murdoch>time for bed anyway.
01:27<Captain_Murdoch>I'll take a look at it tomorrow and try to get it working. I'd like to get scaling working faster before 0.8.
01:28<Captain_Murdoch>later all...
01:28-!-Captain_Murdoch [] has quit ["X-Chat!"]
03:28-!-paperclip [] has quit []
03:28-!-paperclip [] has joined #mythtv
06:26-!-choenig [] has joined #mythtv
06:52<yebyen>captainmurdoch uses 352x480 on an Athlon 700?
06:52<yebyen>i'm on an athlon 700 and I use 320x240
06:53* yebyenconsiders that making the vertical resolution 480 might save him from deinterlacing, and use no additional cpu...
06:56<yebyen>actually, wow, it works fine either way
06:57<yebyen>no wonder, I was only using 25% cpu!
07:09-!-schwin97 [] has quit ["Client exiting"]
08:15-!-choenig [] has quit [Remote closed the connection]
08:25-!-nziarek [] has joined #MythTV
08:44<mdz>rkulagow: I think the right approach is to create a MultiSelectSetting which stores its value as a delimited string
08:45<mdz>rkulagow: then derive a setting from that which represents it as a group of checkboxes (or whatever)
08:53-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
08:58<Chutt>mdz, so i updated my debian install yesterday, and now all my fonts in konsole look like ass
08:59<Chutt>and i can't get it back to something useable
09:33<mdz_>Chutt: solution: don't use konsole
09:34<mdz_>xterm looks fine here
09:34<Chutt>mdz, solution, braindead debian maintainers should be removed from duties =)
09:34<Chutt>fontconfig is only setup to use /usr/share/fonts
09:34<Chutt>_nothing_ else
09:34<Chutt>so all the normal X fonts are not being seen
09:35<Chutt>finally got it working
09:35<Chutt>i only spent 4 hours trying to get my environment back
09:37<eli>what settings do u use for recording & live tv?
09:38<mdz_>complain to #debian-qt
09:38<mdz_>they're breaking everything these days
09:38<Chutt>the issue was fontconfig
09:38<Chutt>it's X, not qt
09:39<Chutt>well, maybe qt since it's told to use it
09:40<mdz_>heh, fontconfig seems like the simplest part of the whole thing
09:40<mdz_>"I have some fonts"
09:40<mdz_>figures that would break
09:40<Chutt>the whole issue was that it wasn't configured to look in the right dirs anymore
09:41<Chutt>and has a big 'DO NOT EDIT THIS FILE' warning on the top of the config file, so i didn't think someone would be changing things
09:41<mdz_>wow it really is unreadable
09:41<Chutt>it seriously pissed me off last night
09:42<mdz_>I thought that was hyperbole
09:42<mdz_>why would changing the font path cause that?
09:42<Chutt>because it's not using the font it was before
09:42<Chutt>his 'good.png' is with a fixed bitmap font
09:42<Chutt>which obviously wasn't found after the update
09:43<Chutt>so it was replaced with a ttf
09:43<Chutt>which doesn't scale down that small well =)
09:44<Chutt>basically, none of the fixed pcf fonts were in the path anymore
09:46<mdz_>isn't that the whole point of scalable fonts?
09:46<mdz_>that they should look good at many different sizes?
09:46<Chutt>not with AA turned on
09:51<Chutt>freetype's AA for tiny sizes really only works on a lcd
09:51<Chutt>when it can use subpixel hinting
09:53<Chutt>otherwise, it just looks bad =)
09:55<rkulagow>mdz, a multiselect would be nice, but i'm afraid that i'm just not going to have the cycles in the short term to even attempt anything like that.
09:57<rkulagow>chutt: i've done the "grunt work" of the mythmusic setup, but the subtleties of the visualizations is beyond me. 3 month old baby + no sleep + work == no cycles. i'll send you what i have, and you can either finish it, or not. or just post it to the list and let someone more talented finish it.
09:57-!-nziarek [] has joined #MythTV
09:57-!-nziarek [] has quit [Client Quit]
09:58<Chutt>rkulagow, i may wait until after 0.8 to do that, if that's ok?
09:59<rkulagow>you're asking _me_ if it's ok to wait? <amusement> uh, sure, i give you my permission to wait until after 0.8. :)
10:00<Chutt>it's something you spent time on, didn't know if you wanted it in right away :p
10:12-!-justin_ [] has quit [Read error: 60 (Operation timed out)]
10:18<rkulagow>well, it does 90% of the what the old text setup did. again, the easy way out is a line edit for the visualizations and just have the user specify them. the harder way (to me, since i'm definately not a C++ programmer) is to setup a bunch of checkboxes, override the save function, then use the boolean values of the checkboxes to build a string that gets saved to the database. ATM, too much "new stuff" that i have to wrap my head around. it can wait.
10:20<rkulagow>chutt: in other news, nathan ziarek has sent me a .pdf version of the HOWTO. It's conveniently printable, and if he can keep up with the changes, then it's a nice option for users to use. mind if i commit it?
10:20<rkulagow>(i've tried to generate pdf
10:20<Chutt>maybe not
10:20<Chutt>how big is it?
10:20<rkulagow> files on my box, but somewhere in the docbook ->pdf it gets messes up.)
10:20<rkulagow>hold on, let me check.
10:21<Chutt>i'd rather not, then
10:21<Chutt>that's rather big
10:21<Chutt>we can make it available from the website, however
10:21<rkulagow>ok, howabout a link and i get it externally hosted.
10:22<rkulagow>ok, i'll get that setup and send you the URL. are you doing the day-to-day maintenance on the website, or transitioned to someone else?
10:22<Chutt>mdz, i've confirmed that the latest fontconfig removes /usr/lib/X11/fonts from the directories given in /etc/fonts/fonts.conf
10:24<rkulagow>i finally pulled the trigger on a mythbackend box. dell poweredge 600SC, p4 2.4ghz, 128MB dram, 40 GB hard drive. $302 after rebate.
10:24<rkulagow>i was running the numbers, and can't build it cheaper, so...
10:25<rkulagow>crucial DRAM is cheaper, and between staples/best buy/compusa, i can get a bigger drive for less than dell would charge me. now, to sneak it into the house... :)
10:32<mdz_>too bad about the P4 though
10:32<mdz_>but that's all you can buy from Dell these days
10:34<eli>how does a P4 compair to a XP for video encoding?
10:35<rkulagow>right. amd would be nice (or for mdz, an epia 10000-M itx box with lindows), but the effort of building my own out of AMD parts would be > $302. (and since it's a server, dell is shipping without OS, so no extra stupid MS license i don't need anyway)
10:35<rkulagow>mdz: smiley at you.
10:37<rkulagow>eli: if you've got enough CPU cycles, it doesn't matter what CPU you're using. if you don't have enough CPU cycles, it doesn't matter either.
10:38<rkulagow>both architectures will produce good video. i think it comes down to cost and philosophy.
10:38<rkulagow>go to if you want to see charts and graphs of AMD vs. x86 for various tasks if you're thinking about buying something and you think it matters.
10:40<eli>what cpu do u have now?
10:42<rkulagow>celeron 450 and a piii/733
10:43<eli>can you encode mpeg4 at the highest quality with that?
10:52<mdz_>rkulagow: that is a pretty good deal
10:52<mdz_>rkulagow: do you have a URL for the rebate info?
10:53<Chutt>is that one of the business only deals?
11:03<rkulagow>yes, but just put your name as the business. they don't care.
11:04<rkulagow>, e-value code: 64732-S20303
11:04<rkulagow>eli: no, i can't encode at highest quality. i can do 480x480 MPEG4 on the Piii/733 at 2200. can't watch at the same time though.
11:05<eli>my wife always bitchs about the quality of the video..
11:06<eli>i've got it turned up about as much as my amd 1600 can do
11:08<rkulagow>add 5% off coupon at checkout: 8D2755970FD4
11:08<rkulagow>free shipping, but tax added (too bad i don't live in oregon/washington state)
11:10<rkulagow>came to $402-$100 MIR=$302. gigE nic, no sound, onboard video 4 64bit PCI slots, keyboard, mouse. perfect for development, without the temptation of playing games. (no AGP slot)
11:15<Chutt>andy davidoff made his ratings patch useful
11:18<Chutt>kevin slater's email
11:18<Chutt>about the capturecard entry not having a hostname
11:31-!-thor_ [~thor@] has joined #mythtv
11:33<Chutt>hey thor
11:34<thor_>hey Chutt, how's it going
11:34<Chutt>just fine =)
11:34<Chutt>you don't mind me applying patches to mythweb, do you?
11:34<Chutt>they all seem to be pretty much correct
11:34<thor_>No, not at all ... actually I felt a little guilty about having promised to get a patch in and then disappearing for a few days (work)
11:35<Chutt>yeah, i'm the same way =)
11:35<thor_>Got single pane mythmusic w/ playlists about 80% done ... will be working on it for the rest of the day
11:35<Chutt>mind holding off on committing that until after 0.8?
11:35<thor_>Nice to get that functionality in before 0.8
11:36<Chutt>i was planning on calling a feature freeze tonight
11:36<thor_>Oooops ... parrallel thinking ... fine, I'll just do a patch to dev, not a commit
11:37<Chutt>that can go in shortly after the release, then
11:37<Chutt>i figure on a release sometime next week, then
11:38<thor_>Suspect that 0.9 (0.85?) will be a minor extension and bugfix release that will probably come quite soon
11:38<Chutt>it'll be 0.9
11:38<thor_>Hmmm .... dangerously close to a 1.0?
11:38<Chutt>next'll be 0.10
11:39<thor_>Hey, it's all semantics
11:39<Chutt>1.0 is reserved for feature completeness
11:39<thor_>Right before the ground up rewrite =)
11:39<Chutt>hoping to never do that =)
11:39<Chutt>well, any more than this 0.8 was
11:40<Chutt>i think andy davidoff must be running things on an ancient machine
11:40<thor_>Got a second MythBox up and running this morning (takes a while) ... how's multibackend in practice ?
11:41<Chutt>his 'major speedup' increased the listings page from 'near' instantaneous to 'virtually' instantaneous on my box
11:41<Chutt>did you see bruce markey's email about all that last night?
11:41<Chutt>he went over things fairly well
11:41<thor_>I'll have a look
11:42<Chutt>there's apparently something wacky in the conflict resolution stuff still, but that should hopefully be fairly easy to fix
11:43<thor_>Poor Andy, I think he just hates to see even unimportant "inefficiency", not a member of the "make it run -> then make it better --> only then make it efficient" school
11:43<Chutt>hey, for mythmusic
11:44<Chutt>how bad would it be to cache all the metadata entries in ram, all the time?
11:44<thor_>All the time? As in as long as mythmusic is in existence, or as long as mythfrontend is in existence?
11:44<Chutt>as long as mythmusic is in existance
11:45<thor_>That's how I'm doing it (main() builds a list of playlists, and passes it down, writes on exit)
11:45<Chutt>like, during the file existance check at the beginning
11:45<Chutt>just load the entire tree into ram
11:45<Chutt>should speed up creation of the selection tree quite a bit
11:46<thor_>Yup, that's how I'm doing it ... only drawback is that I have to have a class hierarchy outside of Qt that parrallels the class hierarchy for the ListViewItem widgets
11:46<thor_>'cause the lists have to exist whether they are drawn or not
11:46<Chutt>but, that's pretty much the only way to do that
11:46<thor_>but it's working
11:47<Chutt>makes it more of a data/view design, anyway
11:47<thor_>Yeah ... so all the delay (such as there is) is at startup ... that's we have a progress bar !
11:53<bigguy>Chutt: you looked at Andy's latest patch?
11:53<Chutt>i'll probably apply it
11:53<bigguy>looked decent to me
12:37<rkulagow>chutt: can you please sync the website to CVS docs?
12:38<Chutt>in a bit
12:38<Chutt>got some work to do first
12:38<Chutt>but soon
12:38<rkulagow>NP. thanks.
12:40<thor_>Oh crap, time to move to sourceforge
12:42<thor_>What's the size of the pipe to CVS?
12:42<Chutt>my dsl line
12:42<Chutt>which i'm using for other stuff at the moment
12:43<Chutt>cvs is staying here, though
12:43<mdz_>yes | cvs
12:43<mdz_>looks like about 1/4 inch maybe
12:43<mdz_>but the size of the pipe depends on your font size
12:43<thor_>Not 3/8's
12:43<Chutt>sf's cvs stuff sucks.
12:44<Chutt>can't move files around
12:44<Chutt>have to email someone if there's a stale locking file left over somehow
12:45<thor_>I wrote an LJ article about 10 years ago, and got a wopping web hosting bill shortly thereafter (in the days when that kind of stuff was expensive). It drives a lot of traffic
12:46<Chutt>i don't think it'll be any worse than /.
12:46<thor_>I take your point
12:52<Chutt>rkulagow, docs updated
13:07-!-nziarek [] has joined #MythTV
13:10<rkulagow>nathan, you here?
13:13<nziarek>btw - that PDF should only be 330KB...I embedded the thumbnails in the one you had (ooops)
13:19<rkulagow>nziarek: ok, thanks. can you do a new pdf based on the updated docs as of 30mins or so ago?
13:21-!-rickter [] has joined #mythtv
13:23<nziarek>just sent it...
13:40-!-rickter [] has quit ["Trillian ("]
13:49-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
14:13-!-justin_ [] has joined #mythtv
14:57-!-r1ckt3r [] has joined #mythtv
15:46-!-dweezle37 [] has joined #mythtv
15:47<dweezle37>Getting some segfaults and want to make sure i'm not missing something obvious before i recompile for debugging..doing a bit of debugging from work... I have a frontend on here at work conecting to my backend at home..
15:47<dweezle37>I can connect fine.. but whenever i try to watch my recordings I get:
15:48<dweezle37>Could not connect to server
15:48<dweezle37>Could not connect to server
15:48<dweezle37>Segmentation fault
15:48<dweezle37>Any ideas?
15:48<Chutt>the 'could not connect to server' doesn't tell you anything?
15:49<bigguy>well it'd be kinda hard to watch from work anyway wouldn't it?
15:49<dweezle37>it doesn't tell a whole lot... i'm able to connect to the database (or is 'server' something else)?
15:50<Chutt>the backend server
15:50<dweezle37>I don't expect the video to be watchable.. but i shouldn't be segfaulting..
15:50<Chutt>feel free to fix it, then
15:50<thor_>Uhm ... is there any chance you actually have the port open at home (is home not behind firewall/NAT)'
15:51<dweezle37>what port is it trying to connect on?
15:51<Chutt>whatever port you told it to connect on
15:51<dweezle37>BackendServerPort | 6543
15:51<dweezle37>TCP connection i assume?
15:52<Chutt>what's backendserverip set to?
15:52<dweezle37>port should be open at firewall..
15:52<thor_>try a telnet
15:52<dweezle37>does backendserverip resolve hostnames? or does it have to be an IP?
15:52<Chutt>it has to be an ip
15:52<Chutt>what's it set to?
15:53<dweezle37>(a hostname)...hmm.. perhaps that's my prob... i'll give that a whirl...
15:53<Chutt>i believe i always refer to it as an ip in the help text for that setting
15:54<dweezle37>yea.. i was just hoping it could resolve.. that way i wouldn't have to keep changing it (dynamic ip)
15:55<dweezle37>nope.. no luck.. interesting thing is that i can start the front end fine..
15:56<dweezle37>I see: connecting to backend server: XXX.XXX.XXX.XXX:6543
15:56<Chutt>make sure masterserverip is the same as backendserverip
15:57<dweezle37>is "/usr/local/share/mythtv/mysql.txt" still being used? or is it all in the DB now?
15:57<Chutt>it uses that
15:58<dweezle37>k... just updated that.. trying again.
15:58<dweezle37>nupe.. same thing...
15:59<thor_>Chutt, perhaps you could configure the DB to get it's connection values from the .... DB .... oh never mind :-)
15:59<dweezle37>well you could run setup, get DB ips, then push all setting over after you are done w/ setup..
16:00<thor_>dweezle, can you "telnet 6543"?
16:00<Chutt>dweezle, the code you'd want to look at is in libs/libmyth/remotefile.cpp
16:00<dweezle37>thor_: yup.
16:00<thor_>so ports are open, what's the bandwidth in between
16:01<dweezle37>telnet XX.XX.XXX.XX 6543
16:01<dweezle37>Trying XX.XX.XXX.XX...
16:01<dweezle37>Connected to XX.XX.XXX.XX.
16:01<dweezle37>Escape character is '^]'.
16:01<dweezle37>Chutt: K.. i'll take a look.
16:01<dweezle37>bandwidth prob isn't great.. being pushed out on a cable modem (from home)
16:02<thor_>Should be enough to get the front screen up though ...
16:02<Chutt>thor, it's only dying after it tries to play
16:03<thor_>Welcome to the age old problem of streaming
16:03<Chutt>it's not even streaming, though =)
16:03<Chutt>not getting that far
16:03<dweezle37>i'll recompile w/ debug...
16:04<Chutt>you shouldn't need to
16:04<Chutt>a couple couts should be fine
16:04<dweezle37>good point.. i'll throw some in.. see what I get...
16:04<Chutt>see what it's doing
16:05<dweezle37>also interesting i'm getting "Could not connect to server" twice. Code only looks like i should get once.
16:06<Chutt>it's opening two sockets
16:06<Chutt>one for control, one for data
16:07<dweezle37>ah.. gocha.
16:08<dweezle37>alright.. i'll have to take some time to stare at the code.. this is the 1st i've actually had a chance to crack the case open... (it's been on my TODO)
16:09<mdz_>has anyone had any scheduler feedback?
16:09<mdz_>I guess that's good
16:09<mdz_>I missed a recording yesterday and don't know why
16:09<mdz_>it's not my fault as far as I can tell
16:09<mdz_>I ran the query and it came back OK
16:10<mdz_>it's suspicious, though, because it was a program where the title and subtitle matched, but the description was different (two-part episode)
16:10<mdz_>but I quite clearly test that in the query, and it gives the expected results
16:11<Chutt>it's seemed ok to me, but i don't have anything in the near future that would be an old record
16:14<mdz_>all of the duplicates that I see show up in viewscheduled are perfectly legit
16:15<mdz_>it's a little bit too much of a coincidence that it was the day after I put that code in though
16:16<mdz_>I should have put a printscheduled() in there so it would get logged
16:16<mdz_>printlist that is
16:24<Chutt>quiet list today
16:25<thor_>badly ... playlists container with playlists with pointers to widgets that hold pointer back to playlist entries blech
16:26<Chutt>might be overcomplicating things =)
16:26<Chutt>if the widgets are in there, at least
16:27<thor_>yes, but playlists exist whether they are drawn or not, so you need to only new the ListviewItem's when and if the Listview exists
16:27<thor_>then you need a way to get back to the main hierarchy from user presses
16:27<thor_>and no connects()
16:28<Chutt>well, i'll see it when you post the patch =)
16:28<thor_>and then hopefully someone who knows what they're doing will improve it ... back to coding
17:03-!-danc_ [~danc``] has joined #mythtv
17:08<-- danc_(~danc`` has left #mythtv
17:08<dweezle37>Ok.. a some quick and dirt debug statements in libs/libmyth/remotefile.cpp show that openSocket is trying to sock->connectToHost(host, port) where host is (Internal IP of backend).. where is this 'host' coming from?
17:09<Chutt>it should be coming from the backendserverip
17:09<Chutt>you do have to restart the backend for changes in that to take affect, of course
17:32<dweezle37>clear something up for me..
17:32<dweezle37>when will BackendServerIP and MasterServerIP ever be different?
17:33<Chutt>multiple backends
17:34<dweezle37>ahh.. ok.
17:34<Chutt>backendserverip is a per-host setting, it's what is sent out for other machines to connect to when they want to talk to it
17:35<Chutt>masterserverip is what everything connects to on startup
17:36<thor_>Chutt, is there a reason you chose QValueList over QPtrList whenever you held collections of Metadata (other than the fact that you tend to get less pointer deletion segfaults)
17:37<thor_>sorry ... this in mythmusic
17:37<Chutt>no reason whatsoever
17:38<thor_>I'm trying to load the music data once, but want to make searching over it as fast as possible ... would seem to me QPtrList should be faster
17:38<Chutt>it would be
17:38<dweezle37>Chutt. Thanks for the help debuging.. i have a much better undersanding of front/back end settings and how they play.
17:38<Chutt>i think i just did it so i wouldn't have to keep track of who owned what metadata object
17:38<Chutt>dweezle37, np
17:39<Chutt>but if there's only one metadata object for any given track, stored in ram, that's not an issue
17:39<thor_>No, I want to have an AllMusic object that knows all available tracks. Then other things can query it instead of asking the DB
17:40-!-dweezle37 [] has quit ["Client exiting"]
17:41<thor_>Sorry, we were saying the same thing
18:32-!-nick_hudson [] has quit [Remote closed the connection]
18:46-!-nziarek [] has joined #MythTV
19:05-!-nick_hudson [] has joined #mythtv
19:13-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
19:20<mdz>Chutt: I redid things with -D_FILE_OFFSET_BITS=64 and nothing broke, ok to commit?
19:21<mdz>I'm hoping this will fix my stdio ringbuffer problems, but I backed that stuff out to do this separately
19:22<mdz>I remembered not to commit my standing change this time :-D
19:22-!-paperclip [] has quit []
19:22<Chutt>heh, thanks
19:22<mdz>it would be nice if that could be specified at build time
19:23-!-paperclip [] has joined #mythtv
19:31<mdz>Chutt: the handleDuplicate dialog should show the program info, I think...what's the preferred format for displaying that?
19:32<mdz>damn, that didn't fix it
19:33<Chutt>there's not really a format for that
19:33<Chutt>so, whatever you want
19:33-!-dweezle [] has joined #mythtv
19:34<dweezle>how do you know if your video card supports xv?
19:34<mdz>dweezle: xvinfo
19:34<Chutt>mythtv complains when it doesn't
19:35<dweezle>yea.. but is there some way to know if the old video card sitting on my desk supports it? A hw database somewhere?
19:36<Chutt>not that i'm aware of
19:48<Chutt>basically most of the neuros mp3 players are getting recalled for a hardware fix
19:49-!-justin_ [] has quit [Read error: 110 (Connection timed out)]
19:57-!-justin_ [] has joined #mythtv
19:59<mdz>I can't figure out why this stdio stuff doesn't work
19:59<mdz>unless fseeko is totally broken
20:09<mdz>when I skip ahead (30 sec), it seems to seek about the right amount
20:09<mdz>but is clearly getting desynced
20:09<mdz>from the frame headers
20:11-!-danc_ [~danc``] has joined #mythtv
20:11<-- danc_(~danc`` has left #mythtv
20:15<mdz>ah, I'm dumb
20:19-!-planetjay [] has joined #mythtv
20:49<Chutt>mdz, got it working?
20:57<mdz>Chutt: yep
20:57<mdz>dunno if it'll help
20:58<mdz>hopefully will cut down on a read() or two
20:58<mdz>cleans up the code a bit too
21:02<mdz>I'm not really sure what a reader thread for ringbuffer would look like
21:03<mdz>read up to a fixed size buffer, sleep until it's emptied some, then try to fill it again?
21:03<mdz>throw it away on a seek
21:04<mdz>am I the only one who sees this behaviour when watching and recording at the same time?
21:05<mdz>I forget who the other guy is who's using NFS
21:08-!-planetjay [] has quit [Read error: 110 (Connection timed out)]
21:09<justin_>what behavior?
21:09<justin_>i usually watch recordings rather than use livetv
21:09<justin_>it works better that way
21:09<justin_>at least, cvs from a week ago:)
21:10<thor_>mdz, you trying to buffer against NFS latency problems
21:42<mdz>thor_: um, yeah
21:54<mdz>thor_: you using NFS?
21:55<thor_>No, at least not yet, just wondering which version/implementation of NFS you were using
22:04-!-meth [] has joined #mythtv
22:09<meth>cant get it to watch tv :/
22:09<meth>VIDOCGMBUF:: Invalid argument
22:09<meth>VIDIOCSAUDIO: Invalid argumen
22:17-!-nziarek [] has joined #MythTV
22:18-!-r1ckt3r [] has quit ["Trillian ("]
22:31<mdz>thor_: Linux, nfsv3
22:34<thor_>mdz: of well, so much for the obvious (100BaseT as well I suppose)
22:35<meth>VIDOCGMBUF:: Invalid argument whats this error from
22:35<mdz>thor_: yes
22:36<mdz>it works quite well for the most part
22:36<justin_>meth: means something isnt working
22:36<mdz>but when I'm both recording and playing back at the same time, I get the occasional stutter
22:36<thor_>I haven't done the maths, but I'm surprised that's not enough on it's own ... hows your router
22:36<meth>lol justin_
22:36<mdz>my theory is that the read() is blocking on the playback side
22:36<mdz>meth: what kind of capture card?
22:37<mdz>thor_: what do you mean? what's not enough?
22:37<justin_>i did the math once, posted it as a reply on the ML
22:37<justin_>1 gig an hour, comes out to like a few hundred k/s
22:38<thor_>mdz: I'm surprised 100BaseT & nfs3 is not enough to keep up
22:38<mdz>thor_: it is
22:38<mdz>there is plenty of bandwidth
22:38<mdz>but the latency will always be higher than a local disk read
22:38-!-dweezle [] has quit ["Client Exiting"]
22:39<mdz>and since there's no read buffer, it will occasionally skip
22:39<mdz>that's the theory I'm working under anyway
22:39<thor_>no read buffer ... yet :-)
22:39<justin_>what are your nfs options?
22:40<meth>i figured that out :)
22:40<meth>what do i need to do to fix it
22:40<justin_>do you use the rsize=8192,wsize=8192 ?
22:40<meth>ati wonder
22:40<meth>not all-in wonder
22:40<meth>works on my redhat install perfectly fine
22:40-!-meth [] has quit [" | Urban Terror Freestyle!"]
22:41-!-meth [] has joined #mythtv
22:41<meth>meh freenode lagged out
22:41<meth>i have a ati wonder-ve
22:41<mdz>justin_: of course not. I'm on ethernet with an MTU of 1500
22:42<mdz>there's no sense in sending 8k packets when they'll need to be fragmented and reassembled
22:43<justin_>meth: is the bttv driver loading correctly?
22:43<meth>xawtv works
22:44<meth>and i have mythv working in redhat
22:44<meth>do i did v4l2?
22:45<justin_>did you try chaning the channel
22:48<meth>i cant load mythtv at all
22:58<meth>heh :/
23:00<justin_>does tvtime work?
23:00<meth>i dont have tvtime
23:01<justin_>get it:)
23:01<meth>after imagemagick
23:06<meth>tvtime works fine
23:07-!-nziarek [] has quit [Read error: 60 (Operation timed out)]
23:10<Chutt>sounds like you're attempting to capture at an invalid resolution.
23:10<meth>default config
23:10<Chutt>what's that have to do with anything?
23:10<justin_>well um
23:10<meth>int Width=480
23:10<meth>int Height=480
23:10<justin_>does default config mean you havent configured it?
23:10<meth>i configured it
23:10<meth>but default config screen size
23:10<meth>for capture
23:11<Chutt>again, what's that have to do with anything?
23:11<Chutt>change it and see if that helps
23:11<Chutt>if you want help, don't question the person trying to help
23:11<justin_>tvtime using 720x580 by default i think
23:11<justin_>or whatever ntsc is
23:11<justin_>though i have the tv wonder(not ve) and use 480x480
23:12<meth>at 720x580 same error
23:12<meth>Changing from None to WatchingLiveTV
23:12<meth>VIDOCGMBUF:: Invalid argument
23:12<meth>VIDIOCSAUDIO: Invalid argument
23:12<meth>ill look at my redhat config
23:13<Chutt>really, though, if VIDIOCGMBUF fails, there's something wrong with bttv
23:13<Chutt>that's the second call in the v4l init sequence
23:14<Chutt>and it's not even setting anything, it's just getting information
23:14<Chutt>'course, you could have the capture device set wrong
23:16<justin_>in the setup
23:16<meth>its /dev/video
23:16<justin_>do you also have a webcam or something?
23:17<meth>also /dev/v41/video0
23:17<meth>ln -s in /dev/video0
23:17<meth>and ln -s for /dev/video
23:18-!-meth [] has quit ["reboot"]
23:20-!-meth [] has joined #mythtv
23:25<meth>heh o well
23:26<meth>now it sits at: Changing from None to WatchingLiveTV
23:26<Chutt>audio, most likely
23:26<Chutt>probably blocking trying to open the device for output./
23:26<justin_>meth: killall esd ;killall artsd
23:26<meth>ah xmms was open
23:26<meth>back to the same old err
23:27<meth>Changing from None to WatchingLiveTV
23:27<meth>VIDOCGMBUF:: Invalid argument
23:27<meth>VIDIOCSAUDIO: Invalid argument
23:27<justin_>i'd strace it
23:27<justin_>see what its trying to do exactly
23:27<justin_>strace -e ioctl mythbackend
23:28<meth>no strace
23:28<meth>id have to insatll it :/
23:28<meth>or use mine from redhat
23:28<justin_>what are you using if not redhat?
23:29<meth>ioctl(5, VIDIOCGTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:29<meth>ioctl(5, VIDIOCSTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:29<meth>ioctl(5, EXT2_IOC_GETVERSION, 0xbffff8d0) = -1 EINVAL (Invalid argument)
23:29<meth>ioctl(5, EXT2_IOC_SETVERSION, 0xbffff8a0) = -1 EINVAL (Invalid argument)
23:29<meth>ioctl(5, VIDIOCSCHAN, 0xbffff8a0) = -1 EINVAL (Invalid argument)
23:29<meth>--- SIGRTMIN (Unknown signal 32) ---
23:30-!-nziarek [] has joined #MythTV
23:30<Chutt>what device is it trying to open for the tuner card?
23:30<Chutt>you see that in the strace?
23:30<justin_>i dont think you can do strace -e "open|ioctl"
23:30<justin_>you'll have to strace -o strace.out mythbackend
23:31<justin_>egrep "open|ioctl" strace.out should work
23:32<meth>ioctl(5, VIDIOCGTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:32<meth>that stuff?
23:33<justin_>there should be a
23:33<meth>ioctl(5, VIDIOCSTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:33<meth>ioctl(5, EXT2_IOC_GETVERSION, 0xbffff8d0) = -1 EINVAL (Invalid argument
23:33<justin_>open ............................ = 5
23:33<meth>ioctl(5, EXT2_IOC_SETVERSION, 0xbffff8a0) = -1 EINVAL (Invalid argument)
23:33<meth>ioctl(5, VIDIOCSCHAN, 0xbffff8a0) = -1 EINVAL (Invalid argument)
23:33<justin_>what is the .........?
23:33<meth>thats it
23:33<meth>let me look
23:33<justin_>if you did strace -o strace.out
23:33<justin_>it has to be there
23:33<meth>open("/usr/share/mysql/charsets/Index", O_RDONLY|O_LARGEFILE) = 5
23:33<meth>open("/dev/dsp", O_RDWR) = 5
23:34<justin_>whats the one right before
23:34<justin_>ioctl(5, VIDIOCGTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:35<justin_>should be open("/dev/video"
23:35<meth>/dev/video" not found
23:35<justin_>can you send me the strace.out?
23:37<justin_>it has
23:37<justin_>open("/dev/dsp", O_RDWR) = 5
23:37<justin_>ioctl(5, VIDIOCGTUNER, 0xbffff910) = -1 EINVAL (Invalid argument)
23:37<justin_>you have /dev/dsp as the video device
23:38<meth>str AudioDevice=/dev/dsp
23:38<meth>where does it store the video info
23:38<justin_>your using cvs eh?
23:38<justin_>its all in the database
23:38<Chutt>in the database when you ran setup.
23:38<justin_>did you never run setup?
23:38<meth>let me run setup again
23:38<meth>but i told it /dev/video
23:38<meth>Use which video device for card 1? [/dev/video]
23:38<meth>Use which audio device for card 1? [/dev/dsp]
23:39<justin_>thats not cvs then
23:39<justin_>thats 0.70
23:39<justin_>what exactly are you using?
23:39<meth>its a gentoo ebuild
23:39<meth>should be from a tar.gz source
23:40<justin_>what did you run?
23:40<justin_>strace what
23:40<justin_>then thats not cvs...
23:40<meth>no its not
23:40<justin_>though in the log it has
23:40<meth>its a ebuild
23:41<justin_>write(4, "Q\0\0\0\3SELECT videodevice,audiodev"..., 85) = 85
23:41<justin_>hard to see wtf is going on
23:41<justin_>looks like the ebuild did something really wrong
23:41<meth>i ran setup again
23:42<Chutt>there's an ebuild linked to in the docs that supposedly works
23:42<Chutt>the one in gentoo doesn't.
23:42<meth>i have the ebuild from the docs
23:43<Chutt>or that's what i've heard, at least
23:43<Chutt>i don't support anything but straight source, or the debs
23:44<Chutt>but, it really looks like the video device is not getting set to the right thing somehow.
23:44<Chutt>could always check in the database and see that it's setup properly
23:44<justin_>yeah, there is no "video" in the strace.out
23:45<meth>cant open /dev/dsp now :/
23:45<justin_>it opens /dev/dsp and then runs the ioctl on it
23:45<justin_>meth: something else using it?
23:47<meth>tvtime works
23:47<Chutt>tvtime doesn't touch /dev/dsp
23:48<meth>nothing is loaded to use /dev/dsp
23:48<Chutt>so, what's the capturecard table look like in the database?
23:50<meth>i got it to work
23:50<meth>i guess when ir an xmms it still had dev/dsp locked
23:53-!-nick_hudson [] has quit ["Client exiting"]
23:53<meth>really choppy
23:53<meth>but it works heh
23:53<Chutt>you'll probably need to lower the quality settings
23:55<meth>hm how do i exit out
23:57<meth>that wont get me out of the first menu
23:57<rkulagow>chutt: i saw the feature freeze announcement for 0.8. i'm assuming that the multiple backend stuff is working sufficiently now? i'm going to get my boxes together and start working it out to get the docs updated
23:57<rkulagow>meth: ALT-F4
23:58<meth>no alt f4
23:58<meth>doesnt work in enlightenment for some reason
23:58<Chutt>seems to be
23:58<meth>man 100% cpu
23:58<meth>this ran better on my dual 514's
23:59<Chutt>rkulagow, i'll attempt to help with the docs, too =)
23:59<rkulagow>thanks, appreciate it.
23:59<justin_>lol i found it funny that someone said redhat/mandrake were good on the ML because the docs had a lot for them