Back to Home / #mythtv / 2002 / 12 / Prev Day | Next Day
#mythtv IRC Logs for 2002-12-28

00:00<mdz>oh, good, it already does that right
00:00<mdz>don't think anyone's tested whether it still works since it was redone
00:02<pheller>does CVS update touch every file!?
00:03<pheller>hmm.... looks like make is recompiling everything after your last update
00:04<mdz>are you sure that's all you got in your update?
00:04<mdz>there have been a bunch of other things tonight
00:06<pheller>oh, hell, I didn't even pay attention. yeah, there was a bunch.
00:15-!-jrh [] has quit [Read error: 110 (Connection timed out)]
00:32-!-TheAsp [] has quit [Remote closed the connection]
01:41-!-jrh [] has joined #mythtv
02:33<-- pheller( has left #mythtv
02:53-!-jrh [] has quit [Read error: 110 (Connection timed out)]
03:20-!-V_laptop [] has quit [Read error: 113 (No route to host)]
03:42-!-jrh [] has joined #mythtv
08:21-!-giuppi [] has joined #mythtv
08:41-!-jrh [] has quit [Read error: 110 (Connection timed out)]
09:07-!-giuppi [] has quit [Read error: 110 (Connection timed out)]
09:27-!-Universe [] has joined #mythtv
09:43-!-jrh [] has joined #mythtv
10:51-!-rkulagow [] has joined #mythtv
10:51<rkulagow>Morning. Isaac, you there?
10:53<rkulagow>Isaac, can you resync the web page documentation to cvs again please? I've added some initial myth-0.8 instructions a few days ago, so hopefully it will cut down on folks discovering the hard-coded IP in backend_settings needs to be modified.
10:59-!-jrh [] has quit [Read error: 60 (Operation timed out)]
11:04-!-m0tion [] has joined #mythtv
11:05<m0tion>hey, i'm kinda confused, i'm browsing the CVS tree and there is no option to download a tarball
11:05<m0tion>it used to be in the lower-left corner if I remember correctly, but now it isn't there
11:08<m0tion>anybody know whats up?
11:10<_shad>it's gone
11:10<_shad>it was taking up too much of chutt's bandwidth
11:10<_shad>use cvs and update :P
11:11<m0tion>are there any instructions for getting it using cvs?
11:12<_shad>somewhere :)
11:12<_shad>cvs -d login
11:12<_shad>cvs -z3 -d checkout MC
11:14<m0tion>do i need a password?
11:15<m0tion>werkz =)
11:17-!-pheller [] has joined #mythtv
11:43-!-jrh [] has joined #mythtv
12:01<Chutt>'Great work otherwise.'
12:07<_shad>Lego Andy?
12:07<_shad>heh. I saw that
12:07<Chutt>that's just insulting
12:07<_shad>I agree
12:10<_shad>I haven't had time to play with mythtv lately :(
12:12<_shad>I have a hard choice... a digital camera or a new computer.. grr
12:17<rkulagow>Isaac, have you thought about setting up one of those Amazon wishlists or a paypal account? Who knows, maybe enough micropayments would add up?
12:19<Chutt>i'd feel funny accepting people's money
12:19<Chutt>i'm just doin this for fun
12:20-!-pheller [] has quit [Read error: 110 (Connection timed out)]
12:20-!-jrh [] has quit [Read error: 110 (Connection timed out)]
12:21<Chutt>the docs on the website should be updated now
12:21<_shad>Chutt: I'll supply you software :)
12:23<_shad>Chutt: you play mame games?
12:23<Chutt>not as often right now
12:23<_shad>Chutt: you have all of them?
12:24<_shad>heh. I think I'm missing 14 of them
12:24<Chutt>mdz, yo
12:25<mdz>lots of criticism on the mailing list this morning :-)
12:25<Chutt>yeah, like i said 'Great work otherwise.'
12:26<rkulagow>Well, as long as you're doing it for fun, "Great work otherwise." Is that going to be the new "in-joke"?
12:27<Chutt>it's just annoying =)
12:28<_shad>Chutt: start charging him for use :)
12:28<rkulagow>Isaac / MDZ, any thoughts on maintaining a "We know this doesn't work, don't complain about it unless you've got a patch" for myth-0.8? As an example, I ran into the same issue that Lego Andy got with the insert cursor jumping around when I was trying to change the path from /mnt/store to /var/video in the setup module. However, since I don't know enough about QT to fix it, I did't post anything to the list.
12:29<Chutt>well, i didn't know anything about that until last night =)
12:30<mdz>dunno why it does that
12:30<mdz>first I'd heard, too
12:30<mdz>does that actually happen?
12:30<mdz>oh, he's using the mouse
12:30<mdz>well wtf is that
12:31<mdz>it should filter mouse events
12:31<Chutt>no, it happens with the keyboard as well
12:31<rkulagow>Hrmm. I get the problem with the left/right cursor keys.
12:31<mdz>you can't move around in the text box with the keyboard
12:31<mdz>oh, maybe in lineedit
12:31<mdz>not in the combobox though
12:31<Chutt>the path names
12:32<mdz>well, not the video or audio device paths
12:32<Chutt>the other ones
12:32<mdz>dunno why it does that
12:33<rkulagow>I'll use cursor left to get to the /mnt/video, and hit "DEL" to get rid of "mnt". The cursor might delete one char, then jump to the end of the text. I usually have to highlight everything and delete and then just type it all in again.
12:33<Chutt>yeah, i dunno either
12:33<mdz>it passes that stuff straight through to qlineedit
12:33<Chutt>think it might be a circular signal
12:34<Chutt>the lineedit widget sends a signal to lineeditsetting
12:34<Chutt>which sends a 'settext' back to lineedit
12:35<mdz>what is he talking about "program icons"?
12:35<mdz>ah, that's probably true
12:36<mdz>we could fix mythlineedit so that it doesn't act so silly if settext is called with the same text that's currently in the box
12:36<mdz>because it's a pain otherwise; it's important that the lineedit get updated if the setting is changed
12:37<rkulagow>Q: on backend settings. Isaac, in the ML, you said "backend_settings.txt is only used by the backend server, not the frontend.
12:37<rkulagow>It's the address that's given to the client as what to connect to for
12:37<rkulagow>playback of files local to the backend server, and to connect to for live tv." How does the frontend find the backend if it doesn't use the backend_settings.txt file and it's on a different box? Broadcast?
12:38<rkulagow>Wait, let me check setup first and make sure that I'm not asking a dumb question.
12:40<Chutt>the initial connection is just made to localhost
12:40<Chutt>that's hardcoded in the frontend for now
12:41<Chutt>there's two different connection types here
12:41<Chutt>which is why that ip address in backend_settings.txt is used
12:41<Chutt>the initial connection is supposed to be to the master server
12:41<Chutt>which handles all high-level requests for stuff
12:42<Chutt>when it goes to do live-tv, or playback a file
12:42<Chutt>it could be to a different backend server
12:42<Chutt>so it uses that setting
12:43<Chutt>so the backend says to the frontend client 'here, connect to this server for live-tv' or 'the file you requested is on this server, that path'
12:44<rkulagow>OK, so there can be multiple back-end servers, but only one master. The front-end connects to the master back-end, which will then send it to the correct back-end. Did I get that right?
12:45<Chutt>though the multiple backend servers haven't been written yet
12:45<Chutt>but the support for them in the frontend is all finished
12:46<mdz>hmm, qlineedit is already smart about when the text has not actually changed
12:46<rkulagow>Is that the reason for the ServerMode: Master in backend_settings?
12:46<Chutt>does it change the cursor position, though?
12:46<Chutt>rkulagow, yup, though it doesn't do anything yet =)
12:46<mdz>yes, whenever setText is called
12:47<Chutt>backend_settings'll be gone soon, though
12:47<rkulagow>Got it, thanks.
12:47<Chutt>and that'll go into the setup program
12:47<mdz>sucks to change it, though
12:47<mdz>I guess I could override setText and save/restore the cursor position
12:47<mdz>otherwise we need to fork qlineedit entirely
12:47<Chutt>don't want to do that =)
12:48<Chutt>forking the wizard stuff was bad enough
12:48<mdz>save/restore looks easy
12:49<Chutt>override the slot, save the position, call the parent's slot, restore the position?
12:50<mdz>yeah...I don't think it can work right in all situations, though
12:50<mdz>like, what if the new text is shorter?
12:50<rkulagow>Isaac, do you remember the idea I had about self-organizing backends? Backend sends a broadcast to the local LAN saying, "I'm a back-end" every 5 seconds. Frontend listens for broadcasts. Multiple back-ends fight it out about who will be Master based on IP address or some other algorithm.
12:50<mdz>rkulagow: please, no :-)
12:50<Chutt>the master needs to do different stuff, though
12:50<mdz>rkulagow: let's not do NetBIOS
12:50<Chutt>like, it'll have the mysql server on it, nothing else needs one
12:50<rkulagow>OK, I'll stay out of the discussion for now so that you guys can figure out the settings stuff.
12:51<Chutt>mdz, only restore if the new length is the same as the old length
12:51<Chutt>or, hmm
12:51<Chutt>i dunno
12:51<Chutt>that won't work
12:51<mdz>it should work ok for the simple insert case
12:51<mdz>but not for replacing the selection or anything like that
12:51<Chutt>well, you can't replace without deleting first
12:52<mdz>er, no, it'll be wrong for insert too
12:52<Chutt>why will it be wrong?
12:52<mdz>never mind
12:52<Chutt>the cursor and the text should be set properly from when the user entered it
12:52<Chutt>it's already changed it and all that
12:52<Chutt>it's just changing it again =)
12:53<mdz>I did the save/restore thing and it behaves for inserts
12:53<mdz>hmm, also seems to work for changing the selection
12:53<mdz>hey, it all works
12:54<paperclip>has anyone determined what the lowend CPU would be for use with the mjpeg cards? I just searched the list but didn't really find much..
12:54<Chutt>well, a p3-550 can't playback full-sized mjpeg video
12:54<Chutt>that's the only datapoint i have
12:54<mdz>and as a bonus, it'll never select all the text
12:54<mdz>because moving the cursor deselects it
12:54<paperclip>i saw that it was working with the duron 650 w/96mb ram over nfs =)
12:55<Chutt>right, but probably not at full size
12:55<Chutt>probably works at quarter size like it does on my p3
12:55<paperclip>what's quartersize?
12:55<paperclip>nice for VCD, eh?
12:55<mdz>selection is nice for the numeric entry boxes
12:55<mdz>but this should'nt affect those
12:56<mdz>that is, if I ever get around to fixing that key entry problem with the spinboxes
12:56<mdz>would be nice to be able to key things in
12:56<Chutt>then you'd have to check for valid numbers and stuff
12:56--> Spark^( has joined #mythtv
12:57<paperclip>i think i'll order a matrox g200.. should be nice to have anyway =)
12:57<Chutt>setting up the g200 was a pain in the ass
12:57<Chutt>there's half a dozen different driver sets and crap
12:57<paperclip>the tv out?
12:57<Chutt>and i haven't even used the tv out
12:57<paperclip>or just the card in general.. oh..
12:58<paperclip>do any of the matrox tv cards do mjpeg and dual head?
12:58<paperclip>i recently set up xinerama and that would be nice =)
12:58<Spark^>quickie question: latest cvs seems to want lots more setting to be in the db, such as capturecard and channels, etc. which step have i missed (there's no longer a setup program, is there?). ie. what should i run before mythfilldatabase?
12:59<mdz>lego andy is infuriating
13:00<mdz>he called all of the configuration options, even the ones which can be graphically configured on-screen, "hardcoded"
13:02<mdz>checking for valid numbers would be easy
13:04-!-Tuscany0_ [] has joined #mythtv
13:06* paperclipis amazed how many people don't clip extranious info from their email replys..
13:06<mdz>oh, you already yelled at him for the 'hardcoded' bit :-)
13:06<paperclip>;search_string=matrox;guest=4784&t=search_engine#4546 <- perfect example
13:07<rkulagow>Gentlemen, ran into a problem doing a "from scratch" install. Compiled CVS, did a "drop database", then ran the mc.sql script. Went into setup/, did a ./setup:
13:08<rkulagow>would you like to clear settings, etc, before starting: y
13:08<rkulagow>pure virtual method called
13:08<rkulagow>Known issue, or a rkulagow fuckup somewhere?
13:09<mdz>rkulagow: did you do 0-7-to-0-8.sql?
13:09<mdz>shouldn't cause that, of course
13:09<mdz>if you were building from a directory you used before, do a make clean before building
13:10<rkulagow>no; hold on. I thought that the mc.sql had incorporated all of the 0-7 changes. (I had dropped the old 0-7 mythconverg database before starting)
13:10<mdz>it has the 0.7 changes. it does not have all of the post-0.7 changes
13:12<rkulagow>Doing a make clean and seeing what's going on, then will re-run. I guess I thought that the mc.sql had the 0-7 to 0-8 changes (saw the "insert" for recording profiles at the end)
13:13<mdz>it's a mixed bag
13:14<rkulagow>OK, then my bad apparently. I'll let the compile finish, then go and update the HOWTO. Assuming a from scratch install, the sequence needs to be compile, database/mc.sql, database/0-7 to 0-8, setup.
13:18<-- Spark^has quit (Read error: 110 (Connection timed out))
13:21<Chutt>what's missing from mc.sql?
13:22<rkulagow>I think that's a MDZ question (about what's missing from mc.sql), correct?
13:22<Chutt>i dunno
13:22<Chutt>i'm pretty sure i got everything in it when i updated last night
13:24<rkulagow>I'll let you know in a minute or two. I've done a make clean, and the make is running now. I'll drop the mythconverg database, then run mc.sql and see if setup aborts again.
13:25<Chutt>setup shouldn't abort like that because of a db problem
13:26<rkulagow>OK, running mc.sql adds 48 rows to mythconverg. Running 0-7 to 0-8 aborts: "ERROR 1060 at line 19: Duplicate column name 'hostname'"
13:26<rkulagow>I seem to remember it dropping 51 rows when I did the drop database mythconverg before starting.
13:28-!-pheller [] has joined #mythtv
13:28<rkulagow>Actually, I just checked, and 54 rows were affected by the drop database command.
13:28<Chutt>might have had some stuff setup already
13:28<Chutt>i dunno =)
13:40<rkulagow>ok, setup; clear: y, Capture card, new capture card. Still aborts the same way. I'm going to try running the 0.7 mc.sql, then the 0.8 0-7 to 0-8 update script and see what happens.
13:42<Chutt>code problem, most likely
13:44<-- tinsley( has left #mythtv ("chix dig openbsd")
13:44<Chutt>if you could run it in gdb, that might help
13:44<pheller>is this a seg fault problem in setup?
13:44-!-jrh [] has joined #mythtv
13:44<Chutt>it's not a segfault
13:55<rkulagow>well, if I drop database, 0.7 version of mc.sql, 0.8 version of 0-7 to 0-8.sql and run setup: clear "y", it still gives the "virtual method" error when i add the new card. if i setup: clear:"n" it segfaults. I'll recompile with debugging and run in gdb next.
13:56<rkulagow>ok, edited, make clean; make is running now.
13:59-!-ummmmmm [] has joined #mythtv
14:00<-- ummmmmm( has left #mythtv
14:14-!-jrh [] has quit [Read error: 110 (Connection timed out)]
14:31<yebyen>Chutt: i must not be much of a quality freak
14:32<yebyen>Chutt: my recordings are 400mb/hr and they're gorgeous
14:33<Chutt>you are using the tv as a display, right?
14:33<Chutt>definately helps
14:33<yebyen>if I weren't, i'd be pissier
14:33<yebyen>i ordered an irman
14:34<yebyen>had enough of tis remote with only 21 buttons
14:34<Chutt>my mythtv box doesn't have the cable hooked up at the moment
14:34<Chutt>no tv for me
14:37<yebyen>i can record a shitton more now
14:37<yebyen>now that the average show is 200MB
14:38<yebyen>ok, shit
14:38<yebyen>my laptop's 512mb of ram and 256mb swap is filling up
14:38<yebyen>something is leaking memory
14:38<yebyen>and i'm pretty sure it's Xft2
14:39<Chutt>mdz, ok, i can't figure out how to easily set the hostname in the capture card stuff
14:39<yebyen>*closes galeon and watches memory usage go down to 20%
14:40<yebyen>Chutt: is 0.8 still a long way off?
14:40<Chutt>i dunno
14:40<Chutt>not too much more to do
14:46<yebyen>118 messages!
14:58<rkulagow>well, don't you hate when that happens? i recompiled with debugging on, and now setup _doesn't_ crash with the "virtual method" call. If I run straight from the command line, it _does_ crash with the same "virtual method" error.
14:59<rkulagow>this is with the database/mc.sql from 0.8 CVS.
15:00<rkulagow>Same thing happens if I use the 0.7/mc.sql and the 0.8 0-7 to 0-8 update script after dropping the database.
15:01<rkulagow>meaning, if doesn't crash if I do a $ gdb ./setup (run) versus $ ./setup
15:09<rkulagow>chutt, mdz, I've got a backtrace on that virtual method error. paste it here, email, or send to list?
15:09<Chutt>just paste it here
15:10<rkulagow>here we go: this is with the 0.8 cvs version of mc.sql.
15:10<rkulagow>[mythtv@pvr setup]$ mysql -u root
15:10<rkulagow>Welcome to the MySQL monitor. Commands end with ; or \g.
15:10<rkulagow>Your MySQL connection id is 56 to server version: 3.23.52
15:10-!-rkulagow [] has quit [Excess Flood]
15:10-!-rkulagow [] has joined #mythtv
15:11<rkulagow>yeah, cute.
15:11<rkulagow>the abort happens after I go to define a capture card and do "new card"
15:12<Chutt>you didn't manage to paste the backtrace
15:12<rkulagow>I didn't? I think trillian doesn't like me then.
15:12-!-rkulagow [] has quit [Excess Flood]
15:13-!-rkulagow [] has joined #mythtv
15:13<Chutt>ok, email it to me
15:13<rkulagow>ok, that's annoying. i'm going to send the text file to you Isaac.
15:14<rkulagow>just mailed it from my rocketmail account.
15:34<rkulagow>Isaac, did you receive the email?
15:38<Chutt>yeah, sorry
15:38<Chutt>was away for a minute
15:40<rkulagow>no proble, just wanted to make sure that I didn't crash yahoo/rocketmail the same way that I've been crashing trillian when I attempt to paste in the bug report. :-)
15:43-!-bigguy [] has joined #mythtv
15:45-!-jrh [] has joined #mythtv
16:14--> Spark^( has joined #mythtv
16:15<Spark^>is anyone else experiencing tv_grab_na being called during setup when it shouldn't be (ie. when i'm set to tv_grab_uk)? i think i've tracked down the function/class that's being called, btu can't figure out why... probably due to not understand qt well enough
16:23-!-bigguy [] has quit ["Client Exiting"]
16:36-!-jrh [] has quit [Read error: 110 (Connection timed out)]
16:42<-- Spark^has quit (Read error: 110 (Connection timed out))
16:48<mdz>Chutt: couldn't you just call gethostname?
16:48<mdz>Chutt: or are you talking about how to get it into the db?
16:50<Chutt>how to get it into the db
16:51<mdz>lego andy is doing us the great favor of a backtrace
16:51<mdz>if and when he gets around to it
16:51<Chutt>yes, i noticed
16:51<mdz>until then, we are on our own
16:51<Chutt>i can't wait
16:51<mdz>dunno what we're going to do until then
16:51<Chutt>probably leave it be
16:52<mdz>check mail really often I guess
16:52<Chutt>and just because robert here started writing docs, it means that cvs is suddenly not development
16:52<Chutt>at least, no more under development than the last release
16:57<yebyen>err, lol
16:57<yebyen>is that sarcasm? i haven't used cvs, so I don't know :)
16:57<yebyen>well, successfully
17:07-!-QSECOFR [] has quit []
17:40<mdz>qmake's "make clean" isn't very clean
17:41<yebyen>how bout make distclean?
17:41<mdz>it should at least get rid of *.a and *.so.*
17:41<mdz>a regular clean should remove everything created by the build
17:42<Chutt>make distclean removes the targets
17:42<Chutt>make clean just removes the object files
17:42<Chutt>mdz, if you get a chance, could you see if tv and stuff still work for you?
17:43<Chutt>i finally got around to doing the libavcodec update
17:43<Chutt>with all the direct rendering junk
17:43<Chutt>i'm not doing some of it right, but =)
17:44<Chutt>everything seems to work for me, at least
17:44<mdz>that's a silly distinction for clean and distclean
17:45<mdz>I'm doing a new build now, I'll go and and install it then
17:45<Chutt>wasn't too hard, there's now a 'release buffer'
17:45-!-jrh [] has joined #mythtv
17:46<Chutt>that we could use to make sure that it's not drawing the osd on something that's still needed
17:46<Chutt>but i'm just ignoring it for now, basically
17:47<mdz>I added the hostname to the capturecard stuff
17:47<mdz>about to commit it
17:48<mdz>I guess the capturecardeditor should probably only show cards on the local host
17:48<mdz>haven't done that yet
17:48<Chutt>no prob, nothing else for the multiple encoder boxes is done either, really =)
17:52<mdz>does setup break for you if you try to edit the same capturecard twice?
17:53<mdz>I'm not sure what's going on there, I get some weird stack traces
17:53<mdz>one of them seems to be infinitely recursing
17:53<mdz>#5 0x402fd1f5 in QBoxLayout::setupGeom () from /usr/lib/
17:53<mdz>#6 0x402fc01d in QBoxLayout::minimumSize () from /usr/lib/
17:53<mdz>#7 0x402c222d in QLayout::totalMinimumSize () from /usr/lib/
17:53<mdz>#8 0x402c0bfe in QLayoutItem::heightForWidth () from /usr/lib/
17:53<mdz>#9 0x402c152a in QWidgetItem::minimumSize () from /usr/lib/
17:53<mdz>#10 0x402fd2f5 in QBoxLayout::setupGeom () from /usr/lib/
17:53<mdz>and so on
17:53<Chutt>probably related to robert's pure virtual error =)
17:54<mdz>a libqt3-mt-dbg would be nice
17:54<Chutt>something's missing a configWidget()
17:54<mdz>heh, stack frame #17655 is mythtv
17:55<mdz>I made configWidget not pure virtual anymore in my last commit
17:55<mdz>since it's OK for stuff not to have a configWidget if it's not visible
17:56<mdz>should be easy enough to set a breakpoint there if something is hitting that
17:56<mdz>whoa, it's blowing up in showFullScreen()
17:56<mdz>that's the qt entry point in that recursive crash
17:57<mdz>is something getting added with itself as parent or something?
17:57<Chutt>where it's drawing all the widgets and stuff
17:57<Chutt>why would it work once, though?
17:57<mdz>sometimes it happens the first time
17:57<mdz>just now it did
17:58<mdz>different stack though
17:58<mdz>#0 0x0064002f in ?? ()
17:58<mdz>#1 0x4009f552 in VerticalConfigurationGroup::configWidget (this=0xbffff440,
17:58<mdz> cg=0x0, parent=0x80eaad8, widgetName=0x0) at settings.cpp:93
17:58<Chutt>the pure virtual stuff
17:58<mdz>in an empty configuration group
17:58<Chutt>it's calling something that doesn't exist
17:59<mdz>it's dying inside the loop over the children in VerticalConfigurationGroup::configWidget
17:59<mdz>but gdb says that children.size() is 0
17:59<mdz>so that shouldn't have even run
18:00<mdz>time for a distclean maybe
18:03<Chutt>i was supposed to clean up a little
18:06<mdz>something seriously weird is happening
18:06<mdz>it's in that loop with i=1 and children.size()=0
18:08<Chutt>is children.size() == 0 before the loop as well?
18:09<mdz>checking that now
18:10<mdz>no, it isn't
18:10<mdz>maybe gdb is just confused
18:10<mdz>though the inside of the children object sure looks like an empty one
18:11<mdz>I'm pretty sure this used to work
18:25<mdz>it doesn't seem to happen with the cardinput dialog at all
18:26<mdz>which is very similar to the capturecard dialog
18:26<mdz>and it only happens randomly in capturecard
18:30<mdz>hmm, I think I found it
18:31* yebyennotes that he has a 4 hour seinfeld marathon waiting for him
18:45-!-jrh [] has quit [Read error: 110 (Connection timed out)]
19:19<rkulagow>Hi guys, segfault on setup. I'll mail the backtrace to Isaac. Action: make clean, make distclean, make, make install, drop database, mc.sql
19:19<rkulagow>gdb ./setup, run, clear: y, ENTER, ENTER (segfault)
19:19<mdz>send the backtrace to the list
19:19<mdz>it's probably in my stuff anyway
19:20<rkulagow>problem is that when I paste, trillian crashes, and it doesn't get sent to the list. ok?
19:21<rkulagow>Matt, is that a good email for you?
19:24<mdz>rkulagow: thanks for the update to the docs :-)
19:24<mdz>rkulagow: yeah, that email is fine
19:24<mdz>rkulagow: I meant the mailing list, not the channel, but sending it to me is fine
19:25<rkulagow>ok, just sent it to you and Isaac from my rocketmail account.
19:30<mdz>ok, this is the same crash that I've been puzzling over for a while
19:30<mdz>someone also reported it to the list recently
19:31<rkulagow>ok then, my bad. wasn't sure if it was a known issue yet.
19:32<rkulagow>Anything you'd like to know about the system, libs, etc?
19:32<mdz>a couple of people have reported it to the list as well
19:32<mdz>not really...I am easily able to reproduce it here
19:32<mdz>I just don't know what's causing it
19:32<mdz>if you can find a system where it's *not* happening, I'd like to know what libs are on it
19:33<mdz>but my assumption so far is that it is a mythtv bug
19:33<mdz>it surely worked before
19:33<mdz>I think maybe Chutt broke it :-P
19:33-!-TheAsp [] has joined #mythtv
19:34<TheAsp>live tv seems to be working ok
19:36<rkulagow>Hrmm. Problem is that I've got a Mandrake 9.0 mono culture here in the house. Got a PIII/733, dual Celeron 450 and PIII-M; I'll check the Celeron box, but I doubt that it's going to be CPU dependant.
19:39* TheAspadjusts the osd fontsize
19:50<mdz>Chutt: maybe something isn't getting destroyed properly...
19:52<mdz>I can reproduce the bug just by instantiating a cardinputeditor, doing an edit, destroying it, creating a new one, and trying to do an edit there
19:53<mdz>er, a capturecardeditor
19:55<mdz>seems to happen with cardinputeditor eventually, but not nearly as often
19:55<mdz>it takes 10 or so iterations
20:00<TheAsp>let me guess, notification that recording is going to start while watching livetv is currently broken . . . :P
20:12-!-jrh [] has joined #mythtv
20:12<mdz>somebody figure out what is causing this crash, fix it, and send me the patch
20:12<mdz>sincerely, mdz. tx.
20:23<mdz>Chutt: it happens even if I take all of the configWidget calls out and fill it with QLabels instead
20:24<mdz>or even just one QLabel
20:35<rkulagow>is that a good "aha", or a reference to the one-hit wonder?
20:35<mdz>it's the setInsideMargin call
20:35<mdz>well, the problem goes away when I take it out anyway
20:35<mdz>but I see no reason why it should be incorrect to have it there
20:35* TheAspblames qt
20:38<mdz>it sure looks like a qt bug
20:42<mdz>rkulagow: I just committed a workaround, can you test?
20:44<rkulagow>Sure. Also, I've got another error; don't know if it's related to the one that we've been troubleshooting. If setup doesn't segfault when the card is defined, if you go back in to edit the card that you just created, it will segfault. Want the backtrace for that email to you and Isaac?
20:46<rkulagow>OK, just did a checkout. make clean and make distclean are done, and the make is running now.
20:52<yebyen>hm, mythbox is about to hit 2 weeks of uptime
20:52<yebyen>in 3 hours
20:56<mdz>rkulagow: that's the same bug
20:56-!-jrh [] has quit [Read error: 110 (Connection timed out)]
21:00<rkulagow>ok. just re-ran twice, and I don't get the crash anymore. I'm able to go back into the capture card dialog and edit my sound capture rate.
21:00<TheAsp>strange error flushing buffer ... OH NO! :P~
21:01<mdz>rkulagow: that matches what I see
21:01<mdz>I only figured it out by trial and error
21:01<rkulagow>Next issue: defined a source called "cable". grab_na worked, and gave me the correct provider. However, when I go to associate a source with an input, (I'm doing Television->Cable), "cable" is available twice in the drop down.
21:01<mdz>what drop down?
21:02<mdz>the video source drop down?
21:02<mdz>are you sure you only have one video source named 'Cable'?
21:02<rkulagow>Associate input with source screen: video source. Choices are (None) the default cable, cable.
21:02<mdz>there's no check for duplicates presently, I don't think
21:02<mdz>the video source menu should show them all; you can also check the database
21:02<mdz>select * from videosource
21:03<rkulagow>Well, this is a clean build, no previous database should exist since I've been dropping mythconverg;
21:05<rkulagow>ok; there's only one "cable" on the video source screen; I've got (Define new video source) and cable listed there. However, when I go to associate /dev/video (Television), the "Video Source" drop down has None, cable, cable.
21:07<Chutt>mdz, was there any reason you're using boxes instead of layouts for those vertical/horizontal groups?
21:08<rkulagow>Well, no more troubleshooting for me tonight; see you tomorrow.
21:08<mdz>Chutt: boxes instead of layouts?
21:09<Chutt>i meant for stuff like the comboboxsetting
21:09<mdz>it's a boxlayout inside a groupbox; I think I used a groupbox to provide a label
21:09<Chutt>how it uses a qhbox, then packs the label and the combobox inside of it
21:09<mdz>now that the label is optional, I guess you could forget the groupbox altogether if there is no label
21:10<mdz>oh, that
21:10<mdz>because it was the first thing I ever used Qt for? :-P
21:10<Chutt>ah, ok
21:10<mdz>qhbox seemed like the right tool
21:10<Chutt>i just had to do that setStretchFactor() stuff
21:10<mdz>but I don't entirely understand the difference
21:10<Chutt>well, the layout's more dynamic
21:10<mdz>Chutt: any idea why setinsidemargin broke the world?
21:10<Chutt>no idea
21:10<mdz>that was a huge pain to find
21:10<Chutt>yeah, sorry
21:11<mdz>not your fault it broke everything :-)
21:11<Chutt>going to respond to that guy that just independently reported it?
21:11<mdz>which one?
21:11<mdz>there have been about 3 :-)
21:11<Chutt>j.r. hyde
21:11<mdz>I already responded to a couple
21:14<mdz>hmm, removing that margin call made things a little ugly
21:14<Chutt>that's why it was there =)
21:14<mdz>that sucks
21:14<Chutt>i'll try to pretty it up a little
21:14<mdz>the default should be to leave enough space for the label
21:14<Chutt>i think it does, but it doesn't take the font size into account
21:14<mdz>are you running qt from unstable, or some unofficial stuff?
21:15<Chutt>unofficial 3.1 debs compilied against g++ 3.2
21:16<Chutt>i believe they're the same as what's in unstable
21:16<Chutt>just compiled with the different compiler
21:17* TheAspdrools over 3.2 qt debs...
21:18<mdz>will QLabel do word wrap?
21:18<Chutt>yup, if you tell it to
21:18<Chutt>check out the help text stuff in mythwizard
21:18<Chutt>it helps to give the label the exact sizes you want, else it does weird things when it wraps sometimes
21:19<mdz>I just want to add a little label for the other grabbers saying that it will run something in the terminal to configure it
21:19<TheAsp>what property sets the menu font for the exit menu?
21:20<TheAsp>fontsize that is...
21:20<Chutt>the medium fontsize
21:20<Chutt>the same size is used pretty much everywhere
21:20<TheAsp>in osd.txt?
21:21<Chutt>which exit menu?
21:21<TheAsp>the one that asks if you want to save position, etc
21:21<Chutt>oh, different one =)
21:21<TheAsp>i made the fonts small so they fit in the boxes
21:21<TheAsp>but now the menu looks like crap :)
21:22<TheAsp>i'm running at 512x384
21:22<Chutt>it uses the info_font size
21:25<TheAsp>well, ill have to wait till next time im not recording to see if 13 works...
21:37<TheAsp>i changed channel and fastforwarded/rewinded a bunch today in live tv
21:37<TheAsp>seemed stable
21:37<TheAsp>slow fastforward/rewind though
21:37<Chutt>i think it's possible to mess it up fast forwarding off the end sometimes, though
21:37<mdz>Chutt: oh, btw, I'm running current CVS now
21:37<Chutt>haven't really looked into it yet, should be something simple
21:37<mdz>nothing obviously broken
21:37<TheAsp>off the end of livetv?
21:37<Chutt>theasp, yeah, past 'now'
21:38<TheAsp>didnt seem to affect me
21:51<mdz>argh, conflicts
21:51<mdz>mumblemumble semicolons :-)
21:51<Chutt>yeah, he's anal about that
21:51<mdz>is that actually necessary?
21:51<mdz>I was pretty sure it wasn't in the API
21:51<mdz>only in the CLI
21:51<Chutt>but, i figure it keeps him happy
21:52<Chutt>and he has contributed some decent stuff
21:52<Chutt>remember that linuxdvrnews guy?
21:54<Chutt>he posted to the freevo list that he's going to use it because python makes more sense for such a project than c
21:54<mdz>he's right
21:54<mdz>python makes more sense for a menu like freevo
21:54<mdz>which is apparently what he wants
21:54<Chutt>they've got a very rudimentary recording setup now, i think
21:55<Chutt>a scheduler thing in python that's cronned to run every minute or something like that
21:55<mdz>so it will launch mencoder at the specified time?
21:55<mdz>python is totally reasonable for that
21:55<mdz>whether that is a reasonable approach to implementing a PVR is another question
21:56<Chutt>what conflicted?
21:56<Chutt>oh, you had some other change that you hadn't checked in yet?
21:56<mdz>I had deleted the entire function containing your edit :-)
21:57<mdz>amazingly, your commits to didn't do any harm
21:57<mdz>I had the same thing of course
21:58<mdz>and so it merged it in with the "already contains the changes..." message
21:58<mdz>and I would have expected that the second commit would wipe out my change
21:58<mdz>but somehow it was smarter than that
21:58<Chutt>it knows its a local change
21:58<mdz>I thought the "already contains" bit would disregard it as a local change, since it went into the repository
21:59<mdz>wow, bunch of patches queued up
21:59<Chutt>that's everything, except the two button channel change stuff
21:59<mdz>oh, some of those are me
21:59<Chutt>which i don't really like
21:59<mdz>seems like it could be done better
22:00<mdz>what do you think about setup prompting and then running mythfilldatabase?
22:00<mdz>I kinda prefer to run it separately, but I may not be a typical user
22:00<Chutt>i'd prefer to keep it separate as well
22:00<mdz>but if it prompts, maybe that's best all around
22:02<mdz>I think the label for the fixedaspectration thing should be more descriptive
22:02<Chutt>the help text can be
22:03<Chutt>i'm still surprised no-one's asked about i18n
22:04<mdz>heh, what would you say if they did?
22:04<mdz>I'm guessing you're not a gettext fan
22:06<Chutt>qt's got a bunch of that all built in, apparently
22:10<mdz>the only thing that's still bugging me in the setup stuff is the lack of any way to exit
22:10<Chutt>1500 messages this month
22:10<mdz>I'm surprised nobody's whined about it yet
22:10<Chutt>can't you force on the allow exit stuff?
22:10<mdz>I mean it's totally unprofessional
22:11<mdz>does that work with escape?
22:11<Chutt>err, yup
22:11<Chutt>either plain escape, or with a modifier key
22:11<mdz>oh, I misread the test
22:11<mdz>I thought it was killable && modifier
22:12<Chutt>or just call themedmenu->setKillable()
22:12<Chutt>like the music stuff does
22:12<mdz>yep, I'll do that
22:15<mdz>do you keep a mythtv todo list anywhere?
22:15<mdz>something lower-level than what's on the website?
22:15<Chutt>not right now
22:15<Chutt>i used to have one written down, but it's old
22:16<mdz>I have a nagging feeling there were some other things I meant to fix in the frontend setup
22:16<Chutt>everything on it's done
22:16<mdz>I know there are a couple of new things in settings.txt that need to go in
22:16<Chutt>oh, check out settings.txt
22:16<Chutt>and theme.txt
22:16<Chutt>everything left in there is not in the ui
22:16-!-Edgan [[] has joined #mythtv
22:16<mdz>does the xinerama stuff only affect the Xv window, or everything?
22:16<mdz>yeah, I noticed you committed a change to it
22:17<mdz>and that there were things still in there
22:17<Chutt>should probably decide which settings are going to be host-based
22:18<mdz>are you planning to do per-host stuff for 0.8?
22:21<mdz>theme.txt still seems to have everything in it
22:22<TheAsp>chutt: how hard would it be to put a "system status" screen in the player?
22:22<mdz>I think all that stuff excetp epgtype is in there
22:22<TheAsp>thats hows whats being played, whats beging recorded, free space, etc...
22:23<mdz>Chutt: so that should just be a checkbox, "use alternative epg layout"?
22:25-!-jrh [] has joined #mythtv
22:25<Chutt>theasp, planning on it for the master backend server
22:26<Chutt>mdz, yes, for 0.8 =)
22:26<Chutt>and yea, just a checkbox for it
22:26<Chutt>oh, i left the theme in since stuff broke if it wasn't set is all
22:26<Chutt>and themeqt and guiwidth/height
22:26<Chutt>but the date formats aren't, that i saw
22:27<mdz>oh, there are classes for them but I didn't put them in the dialog
22:27<mdz>EPGType defaults to on or off?
22:27<mdz>settings.txt seems to say on
22:27<Chutt>default it to off
22:28-!-pheller [] has quit [Read error: 110 (Connection timed out)]
22:28<mdz>not that they can be usefully edited in the frontend anyway
22:31-!-giuppi [~giuppi@] has joined #mythtv
22:31-!-witten [] has joined #mythtv
22:31<mdz>what should the range be for the timezone thing?
22:32<-- witten( has left #mythtv
22:32<rkulagow>Gentlemen, have you thought about doing the "contrast" patch;search_string=contrast;guest=4110&t=search_engine#5091
22:32<Chutt>rkulagow, that's in cvs
22:32<Chutt>mdz, i dunno, it's not needed over here
22:32<mdz>I know, I'm just not sure what time zones are valid
22:33<mdz>well, it won't account for all of them anyway
22:33<mdz>but -12 to 12 or something
22:34<mdz>but I don't think they span that whole range
22:35<mdz>0 for default will do the right thing, right?
22:35<Chutt>make sure it uses the same format that's in the settings.txt file
22:36<Chutt>the code's silly and parses it
22:36<Chutt>by character position
22:37<mdz>oh crap
22:37<mdz>that means I can't use a spinbox
22:37<mdz>then what should the default be?
22:37<mdz>empty string?
22:38<mdz>or "0000"?
22:41<Chutt>should be fine empty
22:41<mdz>empty string
22:42<Chutt>it's an xmltv config
22:42<Chutt>well, it configures filldatabase
22:43<mdz>mythfilldatabase does open the db before it gets that setting, right?
22:43<Chutt>i believe so
22:43<mdz>yep, looks like it
22:44<mdz>so the last thing on my list is doing something about initial frontend setup
22:45<Chutt>what'cha wanting to do next?
22:45<mdz>after that, I want to do the editing of scheduled recording parameters
22:46<mdz>what do you think about the frontend setup question?
22:46<Chutt>i dunno
22:46<mdz>maybe make a list of things which must be set in order for mythtv to work
22:46<rkulagow>You're right - I just saw the commit for the keys.txt file for the contrast patch, which is why I thought that it hadn't been committed yet. Next time, I'll sync to cvs before posting
22:47<Chutt>forcing the user to go through things initially is good, but so is having all those settings set automatically
22:48<Chutt>rkulagow, it's been in cvs for quite awhile =)
22:48<mdz>looking at it, the only thing they should _need_ to set up is the audio device
22:48<mdz>and the default will probably be correct in 99% of cases
22:48<Chutt>and the path names
22:48<mdz>well, those should be moved to setup, right?
22:48<Chutt>ah, right
22:49<mdz>along with tvformat and freqtable
22:49<mdz>I'll add a section for them and do that now
22:49<mdz>dunno what to call it though
22:49<mdz>it'll have recordfileprefix, tvformat and freqtable
22:53<Chutt>i gotta stop using these cvs kde debs
22:54<mdz>need instructions on using CVS -> probably should not use CVS versions of software
22:56<mdz>what's wrong with the kde debs?
22:56<Chutt>konq's having display problems
22:56<Chutt>or something is
22:57<Chutt>some pages don't display until i hide the window
22:57<Chutt>expose event handling's wrong or something
22:58<mdz>have you heard from the VCD guy lately+
22:59<Chutt>no email from him, hasn't joined the channel
23:00<mdz>cleaning out old mythtv mail
23:00<mdz>lots of new features that were never followed up on
23:01<Chutt>i may go and write the radio stuff sometime
23:01<yebyen>which part, the internet stream or the btradio stuff?
23:02-!-giuppi [] has quit ["Lost terminal"]
23:02* yebyenfollows the mailing list, but not that closely
23:02<Chutt>i dunno, probably just the btradio stuff
23:03<yebyen>i'm in a basement, I get next to zero reception :D
23:03<yebyen>in a basement in a valley even
23:03<Chutt>i'd like to do radio timeshifting stuff
23:04<Chutt>i wonder if there's any source of radio show times
23:04<yebyen>i somehow doubt it
23:04<yebyen>if the quality of radio shows around here is anything like the rest of the world
23:05<yebyen>hrm, this recording plays back hyper-slow with no sound
23:06<yebyen>Delaying to next trigger: 562048435
23:06<yebyen>Audio buffer overflow, audio data lost!
23:06<yebyen>i thought that stuff might go away when I switched to mpeg4
23:17-!-Edgan [] has quit ["Client Exiting"]
23:20<mdz>hmm, BufferName should go with the backend too, huh
23:24<mdz>Chutt: should MaxBufferFill go with the backend or the frontend?
23:24<mdz>seems like backend to me
23:25<Chutt>and the buffer size stuff as well
23:26<mdz>and the PIP stuff
23:27<yebyen>hrm, debian's making my life too easy
23:27<mdz>Bryce C reposted his message, changing only the subject
23:28<yebyen>mdz: HEH
23:28<Chutt>let's write everything in one big long paragraph
23:29<mdz>heh, just read that
23:29<mdz>well, skimmed
23:29<mdz>anything with that large a paragraph is unreadable
23:29<yebyen>what an intriguing email
23:29<mdz>I caught the bit about CD and line in, and the word "vfat"
23:29<mdz>I stopped reading after "vfat"
23:29<yebyen>From: bje <>
23:29<yebyen>Subject: A IE 6.0 patch
23:29<yebyen>Content-Type: application/octet-stream;
23:29<yebyen> name=login[1].htm
23:29<yebyen>Content-Transfer-Encoding: base64
23:29<yebyen>Content-ID: <KTJIXlDR>
23:30<yebyen>i wonder why someone at would be sending me a patch for IE6
23:30<Chutt>because it's not someone at redhat
23:30<yebyen>Chutt: sorry, sometimes I forget that sarcasm doesn't translate in plain text ^_~
23:33<yebyen>Return-Path: <>
23:38-!-jrh [] has quit [Read error: 110 (Connection timed out)]
23:43-!-Edgan [[] has joined #mythtv
23:43<TheAsp>yebyen: that can be faked too
23:46<yebyen>TheAsp: yeah, but why would he fake a different address for that? :)
23:47<TheAsp>because it's not a human being?
23:47<yebyen>Received: from Satqfqi ([]) by
23:47<yebyen>can that be faked? (err, it can, but is it at all likely)
23:48<TheAsp>whats the next one?
23:48<TheAsp>the received lines tell a story, at any point it could become fantasy, backwards...
23:54-!-Echo465 [] has joined #mythtv
23:56<mdz>mythtv cvs is still sloooow
23:57<mdz>Chutt must be apt-get upgrading
23:58<TheAsp>pausing an inprogress recording uses more cpu then playing it now
23:58<TheAsp>using yesterday's build
23:59<yebyen>TheAsp: they're all verizon
23:59<TheAsp>what about the one you got from your isp's (or you're) server