Back to Home / #mythtv / 2008 / 03 / Prev Day | Next Day
#mythtv IRC Logs for 2008-03-20

---Logopened Thu Mar 20 00:00:21 2008
00:09-!-dm-madman [] has quit ["BitchX: causing all sorts of havok!"]
00:38-!-gnome42 [] has quit [Remote closed the connection]
01:26-!-Netsplit <-> quits: briand, feiner_, jmk, Dave123, jams, carvajal_, gbee, ead, jhulst, Captain_Murdoch, (+4 more, use /NETSPLIT to show all of them)
01:28-!-BathoryQuorthon [] has quit ["Great Hall Awaits a Fallen Brother"]
01:28-!-Netsplit <-> quits: matty-, asjoyner, jarle, mzb_d800, j-rod, purserj, GreyFoxx, jamesd
01:28-!-anykey_ [] has joined #mythtv
01:29-!-ahbritto [] has joined #mythtv
01:29-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
01:29-!-briand [] has joined #mythtv
01:29-!-jams [] has joined #mythtv
01:29-!-Captain_Murdoch [] has joined #mythtv
01:29-!-_charly_ [] has joined #mythtv
01:29-!-carvajal_ [n=carvajal@] has joined #mythtv
01:29-!-feiner_ [] has joined #mythtv
01:29-!-Dave123 [] has joined #mythtv
01:29-!-jmk [] has joined #mythtv
01:29-!-ead [] has joined #mythtv
01:29-!-gbee [] has joined #mythtv
01:29-!-danielk22 [] has joined #mythtv
01:29-!-carvajal_ [n=carvajal@] has quit [SendQ exceeded]
01:30-!-Netsplit over, joins: purserj, jarle, asjoyner, mzb_d800, jamesd, j-rod, matty-, GreyFoxx
01:30-!-asjoyner [] has quit [Connection reset by peer]
01:30-!-asjoyner [] has joined #mythtv
01:40-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Connection timed out]
01:41-!-grokky_ [] has quit [Connection timed out]
01:43-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
03:24-!-gardz [] has joined #mythtv
03:29-!-xris [] has quit []
03:43-!-Tanthrix [] has quit [Read error: 110 (Connection timed out)]
04:28-!-mattwire [] has joined #mythtv
04:34-!-purserj [] has quit [Read error: 104 (Connection reset by peer)]
05:08-!-mjj29 [] has joined #mythtv
05:24-!-Tanthrix [] has joined #mythtv
05:34-!-joobie [] has joined #mythtv
05:51<joobie>guys how to set mythtv's player to internal?
05:54<joobie>.. went through the playback settings, cant find it
05:56<mjj29>joobie: file types, player Internal, don't use default player
05:59<joobie>where is file types?
05:59<joobie>btw how come not to use internal player?
06:01<gbee>joobie: just set the player command to Internal (capital I)
06:05<joobie>ahh k
06:05<joobie>how to enable socket cmds
06:05<joobie>on port 6546
06:23-!-fargiolas [] has joined #mythtv
06:23-!-fargiolas [] has left #mythtv ["Ex Chat"]
06:24-!-mykeul [n=mykeul@] has joined #mythtv
06:24-!-dapper-daniel [] has joined #mythtv
06:25-!-dapper-daniel [] has left #mythtv ["Konversation terminated!"]
06:30-!-dagar [] has quit [Connection timed out]
06:32-!-dagar [] has joined #mythtv
06:47-!-mzb_d800 [] has quit ["Time to quit"]
06:48<joobie>anyone seen some odd results when u cat /dev/hiddev0 and get numerous ???'s ?
06:48<laga>joobie: topic
06:49<joobie>when i press a button on my apple remote and i cat that dev.. i get like 6 ?'s coming up
06:49<joobie>ahh k
06:49<joobie>thanks laga
06:49<joobie>i guess u aint gona help me yea?:P
06:58-!-mattwire [] has quit [Remote closed the connection]
07:04-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Connection timed out]
07:04-!-MrGandalf [] has quit ["work"]
07:21-!-grim[GameOp]_ [] has joined #mythtv
07:30-!-gardz [] has quit [Read error: 101 (Network is unreachable)]
07:41-!-JoeyBorn [] has joined #mythtv
07:45-!-kurre2__2 [] has joined #mythtv
07:52-!-kurre2__4 [] has quit [Read error: 113 (No route to host)]
07:52-!-Dibblah [] has quit [Remote closed the connection]
07:54-!-grokky [] has joined #mythtv
07:55-!-grokky [] has quit [Client Quit]
07:59-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
08:14-!-grim[GameOp]_ [] has quit []
08:14-!-gardz [] has joined #mythtv
08:16-!-mykeul [n=mykeul@] has quit ["Leaving."]
08:17-!-croppa [] has joined #mythtv
08:19-!-croppa [] has quit [Remote closed the connection]
08:19-!-croppa [] has joined #mythtv
08:49-!-joobie [] has quit ["This computer has gone to sleep"]
08:58-!-clever [] has quit [Read error: 110 (Connection timed out)]
09:00-!-clever [] has joined #mythtv
09:17-!-Dibblah [] has joined #mythtv
09:23-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv
09:52-!-reynaldo_ [] has quit [Read error: 113 (No route to host)]
09:53-!-PointyPumper [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
09:54<Chutt>gbee, you fixing the visualizers?
10:06<gbee>Chutt: I was going to look at it but I don't know what needs to be done
10:06-!-renatofilho^_ [n=renato@] has joined #mythtv
10:07<Chutt>it's something that can be done after merge
10:07<Chutt>what major issues are left?
10:07<Chutt>the mysql binding thing?
10:08-!-purserj [] has joined #mythtv
10:08<gbee>binding thing and a strange mythdialog hang that I can't now reproduce but janneg started to see last night
10:09<Chutt>the custom record screen?
10:09<gbee>I really need to look at a backtrace to work out what's going wrong there, but until I can reproduce it ...
10:09<gbee>custom record screen for me and the gallery (single image view) for janneg
10:13<gbee>ok, the visualiser stuff might not be so bad, but like you said, it can wait
10:13<Cardoe>has the qt4 merge begun?
10:13<Chutt>not quite yet
10:13<Chutt>the bindings issue is fairly major, i think
10:13<gbee>doesn't actually stop mythmusic from working
10:13<Cardoe>Chutt: mysql stuff?
10:14<Chutt>Cardoe, qt stuff
10:14<Cardoe>They changed the way they integrate QSql with stuff
10:14<Chutt>qt3 emulated named bindings, qt4 transforms them to positional
10:14<Chutt>so qt3 could reuse a binding multiple times in one query
10:14<Chutt>qt4 can't
10:15<Chutt>and qt4 has issues if there's more bindings added than actually occur in the query string
10:15<Cardoe>that's no good.
10:15<Chutt>no =)
10:19-!-MrGandalf [] has joined #mythtv
10:23-!-mo0dbo0m [] has quit [Client Quit]
10:23<janneg>Chutt: I have a regexp testing for multiple use of a binded value
10:23-!-gardz [] has quit []
10:23<Chutt>oh cool
10:24<janneg>I suspect the mythdialog bugs might be caused by the use of enter_loop (deprecated since qt3.1)
10:26<janneg>qt has a helpful error message if more bindings are added then in the query
10:29<Chutt>janneg, just need a way to make that exec() blocking
10:29<Chutt>that's all enter_loop is doing there
10:31<Chutt>janneg, gbee, nice work on all this, btw =)
10:32<Chutt>i'm surprised it's stable so fast
10:32<Chutt>well, mostly =)
10:33<gbee>janneg gets all the credit
10:37-!-gardz [] has joined #mythtv
10:44-!-mo0dbo0m [] has joined #mythtv
10:48<sphery>Is the enterLoop() stuff used to make the MythDialog modal? If so, perhaps
10:52-!-nordenm [] has quit []
10:54-!-lsobral [n=sobral@] has joined #mythtv
10:56-!-mykeul [n=mykeul@] has joined #mythtv
11:06-!-reynaldo [] has joined #mythtv
11:12-!-reynaldo [] has quit ["leaving"]
11:12-!-reynaldo [] has joined #mythtv
11:13-!-eharris_ [] has joined #mythtv
11:25-!-eharris [] has quit [Read error: 110 (Connection timed out)]
11:28-!-aeha [] has joined #mythtv
11:32-!-eharris [] has joined #mythtv
11:45-!-eharris_ [] has quit [Read error: 110 (Connection timed out)]
11:54-!-PointyPumper [i=Pintlezz@] has joined #mythtv
11:58-!-mykeul [n=mykeul@] has quit [Read error: 104 (Connection reset by peer)]
12:22-!-Gumba [n=unspoken@] has joined #mythtv
12:34<janneg>thanks, but it would not have been as fast without help
12:46-!-xris [] has joined #mythtv
12:51<Chutt>yay, silicondust replied finally about my dead HDHR
12:53<jams>they replaced my hdhr even after the warranty had expired.
12:53<Chutt>that's cool.
12:53<Chutt>i'm still way under warranty
12:54-!-JoeBorn [] has joined #mythtv
12:57<jams>not to long after it was replaced time warner decided to encrypt everything
12:57<Chutt>i still get the locals
12:57<Chutt>which is all i watch, mostly
12:58<jams>someday i will connect it to an antenna(after i buy one)
13:02-!-gnome42 [] has joined #mythtv
13:08-!-_gunni_ [n=Gunni@] has joined #mythtv
13:10<janneg>gbee: I can't reproduce problems in custom record
13:10-!-Gumba [n=unspoken@] has quit [Read error: 104 (Connection reset by peer)]
13:13<janneg>anyway, I think I'll start the merge after some housework
13:13-!-JoeyBorn [] has quit [Read error: 110 (Connection timed out)]
13:13<janneg>I'll send another warning email before doing so
13:16<Chutt>can we wait until tomorrow?
13:16<Chutt>send the warning today, though?
13:17<Chutt>need to merge daniel's big change in anyway :/
13:19<danielk22>chutt: you want me to have a go at it? I haven't tried yet, but some of the OSD fix from yesterday could be made prettier under Qt4 as well...
13:19<Chutt>sure, if you don't mind
13:19<Chutt>more people testing the merrier
13:20<danielk22>k, I'll check out the qt4 branch now...
13:22-!-Anduin [] has quit [Read error: 110 (Connection timed out)]
13:22-!-jgarvey [] has joined #mythtv
13:23<gbee>janneg: the dialogcode bug is an odd one, I could trigger it every time I tried to go into the custom record screen, but now it just works ...
13:25-!-dekar1 [] has joined #mythtv
13:30-!-dekarl [] has quit [Read error: 60 (Operation timed out)]
13:47<janneg>Chutt: sure
13:48-!-JoeyBorn [] has joined #mythtv
13:48<danielk22>shoot: it looks like the mythbackend isn't working yet in the qt4 port. I'm getting an assert failure in FileTransfer::DownRef() "A mutex must be unlocked in the same thread that locked it." I'll need to recompile with debugging..
13:49<janneg>danielk22: debug build is working fine here as backend
13:50<danielk22>hmm, so no one has seen this yet but me?
13:52<janneg>not sure if anyone started the backend, Anduin might have
13:53<danielk22>If the debug build works here I'll go ahead with the merge from head and then I'll recompile with --release-type=profile to try to repro this with a more useful stack trace.
14:03<danielk22>k, it must just be my setup..
14:06-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
14:11-!-catinpan is now known as catinpan_away
14:11-!-catinpan_away is now known as catinpan
14:14<janneg>danielk22: maybe it's a recorder type. do you have a hdhr or iptv recorder defined?
14:19<danielk22>i think it might just be a limitation of QMutexes in Qt4, they can only be released by the thread that locked them, so locks like FileTransfer::readthreadLock simply don't work in Qt4. There are probably a few other locks like this as well..
14:22<janneg>it is "Attempting to unlock a mutex in a different thread to the one that locked it results in an error"
14:22<Chutt>should've been an error already
14:23<janneg>yes, the same sentence is in in the qt3.0 docs
14:24<danielk22>#6 0x000000000045dfa0 in MainServer::connectionClosed (this=0x2aaaac0185d0, socket=0x2aaaac011b70)
14:24<danielk22> at mainserver.cpp:4075
14:24<danielk22>#7 0x00002b19f49ff7d4 in readyReadThread_iffound (sock=0x2aaaac011b70) at mythsocket.cpp:740
14:24<danielk22>looking at it now
14:25-!-Anduin [] has joined #mythtv
14:25-!-ToadP [] has quit []
14:26<Chutt>danielk22, might just make the stop/pause locks temporary
14:27<Chutt>protecting a 'don't read' variable or something
14:27<danielk22>heh, I was reading that backtrace wrong..
14:28<danielk22>chutt: did you write that socket code?
14:29<Chutt>i don't remember the cross-thread locks, though :p
14:30<danielk22>ah, ok. well I don't think I have the time/concentration to really get to understand that code right now. I'll look at it tonight if no one else has gotten to it before me.
14:30<danielk22>But using a variable protected by a lock should work fine as a solution from what I can tell.
14:32<danielk22>There is no need to hold the lock like that, we aren't using QWaitConditions or anything like that..
14:50<Chutt>does anyone happen to have a small mpeg4+aac .mov file?
14:53-!-JoeBorn [] has joined #mythtv
15:02-!-ahbritto_ [] has joined #mythtv
15:05-!-ahbritto [] has quit [Remote closed the connection]
15:11-!-JoeyBorn [] has quit [Read error: 110 (Connection timed out)]
15:20-!-skamithi [] has joined #mythtv
15:24-!-skamithi [] has quit [Client Quit]
15:35-!-lcase [] has joined #mythtv
15:38-!-ToadP [] has joined #mythtv
15:41-!-phedny [] has joined #mythtv
15:48-!-ToadP [] has quit []
15:50-!-ToadP [] has joined #mythtv
15:50-!-ToadP [] has quit [Remote closed the connection]
15:50<Cardoe>hmm. I think I found another issue in UPnP
15:50<Cardoe>probably already known though
15:51<Cardoe>basically, shows that don't provide an episode name only export 1 show
15:51<Cardoe>rather then all available episodes
15:51<Cardoe>i.e. The Colbert Report
15:52<Chutt>i don't think i've seen that reported
15:55<danielk22>heh, I found the real problem with FileTransfer.. DownRef deletes the mutex then does a mutex unlock..
15:57-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
16:02<danielk22>Never mind, it was a regression I introduced when testing, used a QMutexLocker.. same error message though, which is interesting.
16:03-!-feiner_ [] has quit [Read error: 104 (Connection reset by peer)]
16:04<danielk22>BTW This patch appears to fix the problem:
16:07<danielk22>I think it may just be the unconditional readthreadLock.unlock() causing the problem.. which would be a much smaller patch.. trying now
16:10<danielk22>nope, that wasn't it.
16:11-!-feiner [] has joined #mythtv
16:11<danielk22>chutt: you want to look at the pastebin? I have a bit of a headache right now, so I'm not fit to review my own code at the moment.
16:11<Chutt>i'm kindof deep in some other code for work
16:11<danielk22>k, I'll look at it later then
16:12<Chutt>you sure the refcount/destructor changes are right?
16:12<danielk22>not sure
16:12<Chutt>doesn't _look_ correct, but..
16:13<Chutt>i don't remember why that class is refcounted in the first place
16:13<Chutt>although, might be fine
16:14<danielk22>The ref count starts at zero not one, which may explain why it doesn't look right, that is an unusual construction.
16:16<janneg>gah, basing QString on QChar in qt4 might break our utf8 database strings
16:18<janneg>QString::fromUtf8(query.value(0).toString()) gives broken strings
16:19<janneg>unfortunately only in german less used 8bit symbols are bad
16:19-!-kormoc [n=kormoc@unaffiliated/kormoc] has joined #mythtv
16:20<janneg>I've checked the db and it has the correct utf8 byte sequences
16:21<Chutt>there's gotta be a way to convert
16:26<danielk22>QString::fromUtf8(query.value(0).toQByteArray().constData()) ?
16:28-!-BathoryQuorthon [] has joined #mythtv
16:29-!-MrGandalf [] has quit ["home"]
16:29<Chutt>gbee, now, you have to get it so it doesn't letter/pillar box those images
16:29<gbee>not happy with the frame image, but I'm just playing around
16:31-!-reynaldo_ [] has joined #mythtv
16:34<jams>gbee- is that your victory garden?
16:36<janneg>danielk22: no, same error. it would proabably work if the qvariant weren't already a QString
16:42<sphery>gbee: I love the missing menu. :)
16:43<gbee>sphery: it had to go, waste of space and not to mention it was inconsistent with everywhere else
16:44<sphery>Exactly! It's especially wasteful on NTSC/PAL resolutions, but just plain ugly on HDTV.
16:45<sphery>Of course, now MythGallery might make the "Why MythTV Sucks" page's list of locations with "hidden menus." (Oh, no!)
16:47<gbee>I suppose the OSD menu would be on that list, I'll get to work reducing the video to a 1/3 of the screen size so that we can display the menu at all times
16:48-!-reynaldo [] has quit [Read error: 110 (Connection timed out)]
16:49-!-yfaykya [] has joined #mythtv
16:51<janneg>qt4 mysql driver converts the result to unicode, which causes trouble if the input is utf8 declared as latin1
16:53-!-keith4_ [n=keith@lust.CC.Lehigh.EDU] has joined #mythtv
16:54-!-keith4_ [n=keith@lust.CC.Lehigh.EDU] has left #mythtv ["Client exiting"]
16:58<Chutt>any way we can override?
17:00-!-jgarvey [] has quit ["Leaving"]
17:06<gbee>hmm, with QT4, embedded mysql is supposedly a drop in replacement for an external mysql server
17:06<janneg>it's not converted if the field type is QByteArray. I've changed one column from varchar to varbinary I get a ByteArray Variant back, but the problem doesn't go away
17:07<gbee>don't know how that differs from QT3 .... but obviously they think it's an improvement because it's listed under qt4 features of the mysql driver
17:07<Chutt>janneg, this obviously is a blocker =)
17:07-!-JoeBorn [] has joined #mythtv
17:08<Chutt>what about a DB update to convert the database from latin1?
17:09<janneg>hah, changing to varbinary solves it. after stopping the backend and clearing the old data
17:11<janneg>Chutt: converting db to charset utf8 causes problems since mysql reseves for every utf8 char 3 bytes
17:11<Chutt>oh, right
17:12<gbee>janneg: is that still true in mysql 5?
17:13<gbee>not read it all the way through yet,
17:13<janneg>it's still true for char(x), and for varchar the problem is iirc the index size
17:13<gbee>: To save space with UTF-8, use VARCHAR instead of CHAR. Otherwise, MySQL must reserve three bytes for each character in a CHAR CHARACTER SET utf8 column because that is the maximum possible length. For example, MySQL must reserve 30 bytes for a CHAR(10) CHARACTER SET utf8 column.
17:13<gbee>so yeah :(
17:14<gbee>can't think of any place that we are using char instead of varchar
17:15<janneg>I think we are mostly using varchars, but if a varchar is in a index mysql was resevering length * 3 bytes
17:16<janneg>and the max index size was something like 512
17:18<janneg>I would be fine with changing varchar to varbinary and text to blob. I don't know if it causes trouble for mythweb or binding though
17:18<xris>max index size in mysql is usually 256
17:18<Chutt>other alternative would be to have our own mysql driver
17:18<xris>mysql converts CHAR stuff to VARCHAR internally in most cases, anyway
17:18<Chutt>if qt still has the ability to register a driver at runtime
17:19<gbee>switching the charset to utf8() would probably be the easiest fix, we'd just have to review the places we are using indexes on titles - it's generally something you want to avoid if possible anyway
17:19<janneg>Chutt: it is, the existing drivers are either plugins or compiled in
17:20<Chutt>ie, copy the mysql driver to local
17:20<Chutt>take out the toUnicode stuff
17:21<janneg>changing the database seems cleaner
17:25<gbee>are we using all of the indexes that are currently in place? There is an index on the title column of recorded and I can't think where it would be used
17:28-!-MrGandalf [] has joined #mythtv
17:30<gbee>wonder if hashing titles etc and searches against the hashes would mean smaller indexes and faster scheduling .... I'd assume if it did that mysql would already be doing it internally
17:31<janneg>more utf8 madness in mysql6:
17:32<janneg>they allowing now 4 byte utf8 symbols
17:33<sphery>hashing would only allow exact matches (then again, I think indexes are only used for LIKE comparisons if the argument doesn't start with a wildcard)
17:41<sphery>Wonder how many databases will be broken if mythconverg is converted to utf8 (some users have already changed their tables to utf8-- )
17:42<kormoc>gbee, actually it likely would
17:43<kormoc>gbee, well, depending on the queries that is
17:44<gbee>yeah, wouldn't be used on all queries, just those that might benefit
17:44<laga>sphery: i deleted a similar article in the german wiki because it's plain wrong to do that
17:45<sphery>At least the page has a big "unofficial and unsupported" disclaimer at the top. Unfortunately, it's only at the top...
17:45<gbee>laga: it was wrong until now because of the way mythtv was using the collation to save space etc, but it might have to change
17:46<gbee>sphery: yeah, I added that ;)
17:46<laga>gbee: true, i was referring to previous stable releases
17:46<janneg>I would prefer the binary/blob way
17:46<sphery>I think he's saying it was wrong because MythTV expected the data in the DB to be utf8 encoded as latin1, but if that changes... :)
17:46<laga>sphery: yup
17:48<sphery>Would there be any benefit to MySQL's being able to sort properly if using utf8 instead of binary/blob? Are we actually using or could we be using MySQL sorting for anything?
17:48<gbee>I think we almost certainly are using ORDER BY in a lot of places
17:49<sphery>I know in playbackbox and MythVideo we re-do the sorting (to remove articles)
17:49-!-JoeyBorn [] has joined #mythtv
17:49-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
17:49<gbee>not sure we'd want to change that either, it's not just the work in involved but the added layer of complexity
17:50<janneg>users had still to configure the right collate to sort correctly. sorting is locale dependent, even in utf8
17:50<sphery>yeah, just saying we might be re-doing the sorting in most places.
17:50<sphery>Oh. I thought the collation could be specified in the connection or query.
17:51-!-joobie [] has joined #mythtv
17:52<janneg>sphery: it might, but then it has to be configurable in mythtv
17:53<janneg>at least if users care about correct sorting for their language
17:53<gbee>it can, so it could be tied in with the locale selected by the user - I've been planning a locale settings idea, a set of default settings according to the users location
17:54<sphery>Well, FWIW, that was really the only difference I (admittedly uninformed) could see between utf8 vs binary/blob.
17:55<janneg>the other difference is utf8 wasting space
17:55<gbee>e.g. The user picks their country where they would currently select their language on startup and things like language, date format, tv broadcast format etc are all preset to the defaults for that locale
17:56<sphery>Right. Not questioning the space thing, but was taking that as a definite difference based on what you'd said. I gues the only other difference.
17:56-!-renatofilho^_ [n=renato@] has quit ["Ex-Chat"]
17:56<gbee>utf8 only wastes space on the indexes, which isn't such a huge issue, especially if we can remove those indexes which aren't really needed
17:59<danielk22>gbee: it also reserves 3 bytes for each character right?
17:59<janneg>not in varchars or texts
17:59<gbee>on with char(), seemingly not for varchar which is what we use anyway
18:00<janneg>the other difference is that we don't need the from/toUtf8() if the databse is utf8
18:01<danielk22>are there any databases we may want to support in the future that don't handle UTF8?
18:01<gbee>if we want to go down the binary/blob route then fair enough, I just think it complicates things just a little more and would involve a lot more work
18:02<danielk22>Seems that UTF8 is cleaner all around then.
18:03<gbee>danielk22: I think we've pretty much settled on using embedded mysql in the future, at least as the default, the postgresql argument goes away if we do that
18:03<janneg>gbee: I think utf8 is a little more work since the field lengths/indices need to be modified
18:03<janneg>but it doesn't matter
18:03-!-Cardoe [n=Cardoe@gentoo/developer/Cardoe] has quit ["Leaving"]
18:06<gbee>field lengths don't need changing? The length refers to the number of characters and not the byte length, or am I wrong?
18:06<sphery>I think MySQL 5 also raised the max key length to 3kB from 1kB, so that should mean our keys should continue to work even if 3x the bytes.
18:08<sphery>gbee: I know some people who did the utf8 conversion ended up with programid of 12 chars (from 40 chars) after the table was converted (truncating old data). I don't know if that was only with MySQL 4 or also with MySQL 5.
18:08<janneg>gbee: yes, they specify the number of characters. if the would specify the number of bytes there would be no difference to the currect situation
18:09*gbee is confused
18:09-!-czth__ [n=dbrobins@nat/microsoft/x-fd5a8fdb73b9e33b] has joined #mythtv
18:09<danielk22>gbee is not alone
18:11<janneg>with latin1 we can store x bytes in a varchar(x). since we store utf8 in the varchar only symbols in ascii (7bit) take one byte. all other symbols take 2-3 bytes
18:12<janneg>so we can store in a latin1 varchar(x) at most x utf8 symbols
18:12<danielk22>janne, I understand that, but how did people lose characters in the conversion to utf8 if the fields are specified in characters, not bytes?
18:12-!-jmk_ [n=jmk@] has quit ["Leaving"]
18:13<janneg>ah, I don't understand that either
18:13-!-lcase [] has quit []
18:13<janneg>maybe a mysql bug or more likely user error
18:14<sphery>I got the numbers wrong (20->6), but it did occur to some users when upgrading to MySQL 5 (utf8) from MySQL 4.x (latin1).
18:15<danielk22>Is anyone other than me getting black rectangles drawn over the video output a few times a second with XVideo and OpenGL video rendering?
18:17<janneg>hmm, I'm still undecided, but leaning towards utf8 charset
18:18<gbee>xv-blit is fine here
18:18<janneg>here too
18:19<gbee>opengl is ok too
18:20<gbee>until it crashed
18:20<danielk22>k, it must be something here then.
18:26<danielk22>chutt: I checked the reference counting, it's correct.
18:26-!-czth_ [n=dbrobins@nat/microsoft/x-9a0c1d18e15858f9] has quit [Read error: 110 (Connection timed out)]
18:30<gbee>eww, mythui doesn't handle overlapping widgets too well if the actions of one update the second
18:31<gbee>e.g. scrolling through a list updating a textarea
18:32<gbee>easy to avoid by just being careful about positioning, but I'll see if we can't avoid it
18:32<danielk22>gbee: there is no enforced paint order?
18:34<danielk22>nah, there must be.
18:34<gbee>draw order is handled in the order that widgets are declared, in this case the textarea was beneath the list which updated it
18:35<gbee>since the order is theme defined there is ample opportunity for themers to shoot themselves in the foot :)
18:35<danielk22>this just means things need to be ordered correctly in the XML?
18:36-!-th1 [] has joined #mythtv
18:36<danielk22>not really a bug then..
18:37<gbee>no, not a bug, just something I'd like to see handled better, even if it's just in the form of a warning to the log - took me a while to figure out the problem and I know the code and wrote the theme
18:42<gbee>didn't help that I initially thought it was connected to a qt4 change, so I was barking up the wrong tree for a while
18:42<Chutt>gbee, what went wrong with it?
18:42<Chutt>shouldn't both things get marked as dirty, and redrawn?
18:42<gbee>in the end it was just my clumsy typing, I'd inverted a couple of digits in a position coordinate
18:43<gbee>Chutt: neither is drawn until a redraw is forced (changing window focus etc)
18:44<Chutt>are they not getting marked as dirty, then?
18:45<Chutt>the animate() timer should pick up on the dirty rect and request a redraw
18:46<gbee>not entirely sure what happens and I'm not planning to look at it for a while, but if you want to reproduce the problem, give mythgallery a spin
18:47<Chutt>err, something's missing a SetRedraw() call, i'd assume
18:47<gbee>the "text" textarea overlaps with the "images" buttonlist, the button list emits the itemselected signal when you move around the list and that is connected to a slot which updates the label/caption text
18:48<gbee>I was just about to commit a fix for the theme
18:48<Chutt>i don't have time now =)
18:48<gbee>ok, it's on my list so I'll get to it eventually
18:49<gbee>damn, maybe that wasn't the problem after all - it's still doing it
18:56<gbee>ahh, ok, that same textarea went offscreen by 20 pixels and that caused the problem
19:01<sphery>gbee: I'm pretty sure that the column size changes/truncation issues were all due to system misconfiguration after the MySQL upgrade causing the server to become confused.
19:01<sphery>MySQL docs say, "For a column that has a data type of VARCHAR or one of the TEXT types, CONVERT TO CHARACTER SET will change the data type as necessary to ensure that the new column is long enough to store as many characters as the original column."
19:03<gbee>sphery: thanks for digging that up
19:07<sphery>I was mainly trying to figure out how their DB's got corrupted, and it seems it was always due to misconfiguration after a MySQL version upgrade and using the MySQL binary data files, so now I'm not concerned about the issues during the change (for people who haven't already broken their DB's :).
19:08<sphery>Wonder how long it will take to rebuild indices after the conversion, though. I also wonder if it's possible to do it any other way than with, "myisamchk -r -q --set-collation=collation_name"
19:13<gbee>janneg: any idea what might be causing this?
19:17<Anduin>gbee: That is what the more bound values than are in the query thing looks like.
19:18<gbee>Anduin: yep, I just can't see what's wrong in this case
19:18<janneg>yes, but it isn't in this case
19:19<janneg>gbee: I have no idea. I have the same error in the scheduler
19:19<gbee>MythContext::SaveSettingOnHost() triggered from UpdateDBVersionNumber(const QString &newnumber) in dbcheck of mythphone
19:20<Anduin>gbee: The error is right there in SaveSettingOnHost...
19:20<gbee>since all plugins use a near identical dbcheck I'd assume they are all affected
19:20<Anduin>gbee: In one version of the query NULL is used but :HOSTNAME is always bound...
19:20<janneg>Anduin: I've fixed that already
19:21<gbee>no it's not - if (!host.isEmpty())
19:22<Anduin>Yes, I was looking at the old code.
19:24<gbee>let me just make clean anyway, I can't see another explanation
19:28<janneg>sphery: we can't use CONVERT TO CHARACTER SET since ouras latin1 declared columns don't hold latin1 data
19:29-!-yfaykya [] has left #mythtv []
19:37<janneg>modify doesn't work either. MODIFY first to binary and than back to char CHARACTER SET utf8 works
19:50-!-jd86 [n=jdio@pdpc/supporter/student/jd86] has joined #mythtv
19:51<sphery>Oh, yeah. I didn't think it through all the way. Makes sense. I shouldn't have read that wiki page to see what others had done--I didn't think to question its approach. Regardless, the column truncation thing was an unrelated issue.
19:53<janneg>I'm still wondering why fromUtf8(value.toString().toLatin1().data()) doesn't work
19:54<janneg>it seems that not all control chars from latin1 are in unicode BMP
19:59<janneg>argh, not control but 0x80-0x9F are unused and the second byte of the utf8 string for the broken chars is in 0x80-0x9F
20:07-!-|gunni| [] has joined #mythtv
20:23-!-_gunni_ [n=Gunni@] has quit [Read error: 110 (Connection timed out)]
20:30<xris>anyone have thoughts on how I can debug this firewire-not-recording stuff?
20:34-!-mzb_d800 [] has joined #mythtv
20:42-!-lsobral [n=sobral@] has quit ["leaving"]
20:55-!-TelnetManta [] has joined #mythtv
21:08-!-|gunni| [] has quit ["KVIrc 3.2.4 Anomalies"]
22:03-!-ahbritto_ [] has quit [Client Quit]
22:17*MrGandalf reminisces about his ol' Amiga.
22:18<MrGandalf>that's what we need, an Amiga MythTV port.
22:25-!-croppa [] has quit [Read error: 113 (No route to host)]
22:27-!-croppa [] has joined #mythtv
22:51-!-kormoc [n=kormoc@unaffiliated/kormoc] has quit []
23:00-!-superm1 [] has quit [Read error: 113 (No route to host)]
23:07-!-dekar1 [] has quit [Read error: 104 (Connection reset by peer)]
23:12-!-dekarl [] has joined #mythtv
23:12-!-dekarl [] has quit [Read error: 54 (Connection reset by peer)]
---Logclosed Fri Mar 21 00:00:24 2008