Back to Home / #mythtv / 2008 / 02 / Prev Day | Next Day
#mythtv IRC Logs for 2008-02-10

---Logopened Sun Feb 10 00:00:25 2008
00:02<kormoc>xris, perhaps later. Plenty to do right now :)
00:13<kormoc>xris, ooh, btw. I profiled the listing's popups. There's no way with how we currently do the schedules to speed it up. 75% of cpu time is spent talking to the backend
00:15-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
01:23-!-gnome42 [] has quit [Remote closed the connection]
01:37-!-ahbritto [] has quit [Remote closed the connection]
01:37-!-ahbritto [] has joined #mythtv
01:39-!-ahbritto [] has quit [Client Quit]
01:53-!-Netsplit <-> quits: RaYmAn-Bx, prg3, |gunni|, anykey_
01:54-!-okolsi [n=mythtv@] has joined #mythtv
01:55-!-emja [] has joined #mythtv
01:55-!-anykey_ [] has joined #mythtv
01:56-!-prg3 [] has joined #mythtv
01:56-!-RaYmAn-Bx [] has joined #mythtv
01:56-!-emja [] has left #mythtv ["Ex-Chat"]
02:13-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
02:34<xris>kormoc: huh? like I said, don't talk to the backend. create an ajax responder script that bypasses that stuff
02:35<kormoc>xris, given how we tell what's scheduled, we load up the entire schedule and then get the matches. I'm gonna need to look into some more backend commands to see if I can just grab info for one record and not all of them
02:37<xris>oh, that's right.. the popup info contains schedule info.. forgot about that.
02:37<xris>there's an xml method to pull up info for just one recording.
02:37<xris>I wanted to migrate to the xml interface for most of that stuff, but it's not a small task
02:38<kormoc>If I keep up my current random spirt, could get it done by next week :P
02:40<xris>you've been pretty productive. heh
02:42<kormoc>Yeah, I'm running out of stuff, likely will finish the valid html stuff and then move onto finishing up the objects if I don't lose steam
02:42<kormoc>If I write in the XML stuff into the objects now, it'll be really easy to swap over to and have it just magically speed up.
02:44<kormoc>xris, we're bumping the php version after this release, right?
02:44<kormoc>5.2 by any chance? :P
02:44<xris>but yeah, I'd say do it about 5 mins after the release. :)
02:44<xris>5.2 is fine with me
02:45<xris>actually, no
02:45<xris>centOS is still using 5.1.6
02:45<xris>should probably use that as a baseline
02:45<kormoc>5.2's built in JSON stuff is amazingly fast, tis why I'd want to go that high, but fair 'nuff
02:45<kormoc>it'll be easy enough to swap over whenever centos updates
02:45<xris>could we do a compat thing like you did for php5 support?
02:46<kormoc>Yeah, I can do a easy wrapper
02:46<kormoc>the autoloader will do the magic for us then as well
02:46<kormoc>new JSON()
02:46<kormoc>if it's built in, the autoloader won't trigger, and thus built in, else it'll load the wrapper class
02:48<xris>so yeah. 5.1.6 min, 5.2.x recommended
02:49*kormoc nods
02:49<kormoc>So what's the thoughts on the iPhone interface stuff?
02:55<xris>the "doesn't work on the iphone" ticket?
02:55<kormoc>well, it's extending the wireless device stuff to be specific for certain platforms.
02:56<xris>in that case, no.
02:57<xris>fixing the mobile support in general would be good, but not at the expense of making the code a lot more complex for a small minority of users.
02:57<kormoc>well, to be fair, it doesn't actually make it specific yet, but it's the framework to start allowing specific device templates
02:58<xris>wap and wml are standards...
02:58<xris>I'd rather just stick with wap, lite and wml
02:58<xris>er, wap lite and default
02:59<xris>maintaining all of the extra template files is more a pain even now...
02:59*kormoc nods
02:59<hads>Could do template choice from a cookie or something if people want to target specific templates to secific devices.
02:59<xris>hads: the settings already allow that
02:59<hads>heh was just thinking that it might :)
02:59<xris>hads: the problem is maintaining an entire different template (or a subset thereof) specifically for the iphone.
03:00<xris>and/or another for the blackberry, or opera mobile, or a treo, or .....
03:00<hads>Oh I agree, I don't do platform specific stuff myself.
03:00<kormoc>I'll pick and choose from his patches then, cause some of it I like
03:00<xris>WAP template definitely needs some help. I've been meaning to poke at it on my blackberry, since imho it's completely unusable on a small screen
03:00<kormoc>Yeah, I can't really help you there :P
03:00<xris>but there have to be better ways to handle things
03:01<hads>A choice by cookie should be sufficient. People that want a specific template can obviously make their own.
03:01<xris>you need a cell phone. :)
03:01<kormoc>I have one! :P
03:01*kormoc totally thinks they forgot they pay for it
03:01<hads>Is WAP still alive? I thought most everything supported HTML/XML these days.
03:02<xris>hads: WAP is sort of like html
03:02<xris>just a subset
03:02<kormoc>does WAP or WML support JS?
03:02<hads>Yeah, I know WAP, I just haven't supported it for ages.
03:02<kormoc>I'm such a JS whore these days...
03:03<xris>kormoc: blackberry supports rudimentary js. dunno how standard it is, though
03:03<xris>I'd say it's best to leave it off, though
03:03*kormoc nods
03:03<xris>XHTML Mobile Profile (XHTML MP), the markup language defined in WAP 2.0, is made to work in mobile devices. It is a subset of XHTML and a superset of XHTML Basic. A version of cascading style sheets (CSS) called WAP CSS is supported by XHTML MP.
03:04*kormoc wonders if it's called XMP
03:04<xris>I say we shoot for that
03:04<xris>but not much we can do before .21 unless I get really un-lazy.
03:04*kormoc nods
03:05<xris>honestly, my arm/shoulder have been killing me for the last few weeks. I don't like to hang out in front of my computer much when that happens.
03:05<kormoc>gonna clone wap or start entirely fresh?
03:05<kormoc>you should hookup the voice coding stuff
03:05<xris>probably take WAP and upgrade/fix it.
03:05*kormoc nods
03:05<xris>the xhtml-mp stuff is still "wap"...
03:06<kormoc>Hrm. I didn't think it was backwards compatible tho, but that may not be a problem
03:07<hads>WAP 2.0 :)
03:07<xris>no reason to support ancient WAP.
03:07<hads>Ancient WAP is weird anyway.
03:08<xris>that's WML, not WAP
03:08<xris>we *could* kill the WML stuff, though.
03:08<xris>even for .21
03:08<hads>Wel, it's WML which is in WAP
03:08<xris>yeah, all that stuff sort of confused me
03:08<kormoc>xris, I'm for that (killing wml...)
03:09<kormoc>Ooh, I still need to clean up more JS from the lite templates too...
03:10<xris>I think on a lot of that, I was waiting to finish the look/feel on the Default templates, and copy them over to lite as I remove the javascript, etc.
03:11<kormoc>A lot of the lite templates I updated the JS to use the prototype stuff so I could remove the other stuff, but the JS didn't work before the conversion cause it was only half added/removed (whichever way it went)
03:11<kormoc>so it's rather sucky right now
03:15<xris>shouldn't have any js
03:16<xris>heck, I'd personally love to just remove it, but I've had enough people threaten to riot if I pulled it
03:19<xris>but that's why I've set it so that it only does the tv stuff.
03:20<xris>heck, could theoretically turn the lite thing into xhml-mp
03:20<xris>strip it down, paginate the listings page, etc.
03:21*kormoc nods
03:22*xris goes to make another hat pattern for RPP
03:22<xris>(and I'm aware that kormoc is the only one who will know what that means)
03:29<rooaus>xris, kormoc: Just curious, has there been any thought about adding the autoexpire list of recordings to mythweb?
03:30-!-jamesd [] has quit [Read error: 104 (Connection reset by peer)]
03:30<kormoc>xris, Even after the rousing endorsement of the indy hat by raedwolf?
03:30<kormoc>robert__, yeah, there has been a ton, but I've just never gotten around to it.
03:31<kormoc>whoos, rooaus, that was to you
03:33<rooaus>kormoc: Cool, I only thought about it the other day... been using mythweb for years (literally) and only just noticed it wasn't there. :)
03:36<kormoc>rooaus, what will likely actually happen is to add a field to the recorded shows page that is sortable
03:36<kormoc>but we'll see if that fits and all that jazz
03:48<rooaus>kormoc: Do you mean for displaying the autoexpire list like in the status of the frontend?
03:53<rooaus>That would be handy, especially with the undelete functionality. Might peek at the php some time.
03:53<rooaus>*new undelete
03:53*kormoc nods
03:55-!-okolsi [n=mythtv@] has joined #mythtv
04:12-!-ma9mwah [] has quit []
04:16-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
04:24-!-|gunni| [] has joined #mythtv
04:31-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
04:43-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Remote closed the connection]
04:53<xris>rooaus: it's not in mythweb because I didn't know it existed...
04:53<Chutt>is mythweb safe to update?
04:53<xris>Chutt: should be
04:53<xris>kormoc's fixed most of the issues from the last week
04:53<Chutt>lotta files changed..
04:53<xris>pages are 30-60% smaller
04:53<xris>javascript is cleaner, etc.
04:53<Chutt>is the bug where it kills X when you try to toggle autoexpire gone? =)
04:53<kormoc>html is mostly valid now as well
04:54*kormoc laughs
04:54<kormoc>it's worth a try, we're using prototype for all of that now
04:54<Chutt>i was thinking of updating and trying to see what's wrong with preview generation
04:55*xris can't complain about that. :)
04:55<xris>totally forgot about that LANG bug
04:55<kormoc>it's mostly just polishing it up, and hopefully speeding it up a bit so that it's a nice improvement for the release
04:55<Chutt>what LANG bug?
04:56<xris>for some reason, non-english utf8 LANG settings seem to cause mythweb to re-request all of the thumbnails with every page load
04:57<Chutt>which would match what i'm seeing
04:57<Chutt>but i don't have LANG set
04:57<xris>well, that's not en_US.UTF-8
04:57<kormoc>I'm currently hit by the thumbnails can't be generated at all and thus get re-requested ever time
04:57<rooaus>xris: Fair enough, myth is a sufficiently large project that it is easy for that to happen.
04:57<Chutt>well, LANG=C
04:58<Chutt>which is essentially unset :p
04:58<xris>Chutt: yeah. so not sure what's up with that.
04:58<xris>might dig into it tomorrow, but not sure what my day will be like. wife sort of holds my weekend calendar.
05:01-!-beavis [] has joined #mythtv
05:03<Chutt>heh, 5 am
05:05-!-nordenm [] has joined #mythtv
05:05<xris>yeah, yeah.. I know your (work) life involves around us west coast types. ;)
05:07<xris>anyway, time for me to sleep. or at least get away from the computer again.
05:07-!-xris [] has quit []
05:45-!-ma9mwah|tired [] has joined #mythtv
05:46-!-foo8ar [] has joined #mythtv
05:55-!-okolsi [n=mythtv@] has joined #mythtv
06:03-!-nordenm [] has quit []
06:05-!-justinh_ [] has joined #mythtv
06:16-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
06:27-!-foxhunt [] has joined #mythtv
06:32-!-justinh [] has quit [Read error: 110 (Connection timed out)]
06:49-!-foxhunt [] has quit [Read error: 110 (Connection timed out)]
07:03-!-beavis [] has quit ["Verlassend"]
07:03-!-beavis [] has joined #mythtv
07:09-!-MaverickTech [] has joined #mythtv
07:28-!-XChatMav [] has quit [Read error: 113 (No route to host)]
07:32-!-BleedAway [] has joined #mythtv
07:32-!-robert__ [] has quit [Read error: 110 (Connection timed out)]
07:32-!-BleedAway [] has quit [Remote closed the connection]
07:55-!-okolsi [n=mythtv@] has joined #mythtv
07:57-!-jamesd [] has joined #mythtv
07:58-!-robert__ [] has joined #mythtv
07:59-!-justinh_ is now known as justinh
08:02-!-mattwire [] has joined #mythtv
08:15-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
08:24-!-rooaus [] has left #mythtv []
08:26-!-BleedAway [] has joined #mythtv
09:07-!-nordenm [] has joined #mythtv
09:13-!-aquarius [] has joined #mythtv
09:18-!-aquarius [] has left #mythtv ["Ex-Chat"]
09:30-!-JoeBorn [] has quit [Read error: 110 (Connection timed out)]
09:55-!-okolsi [n=mythtv@] has joined #mythtv
09:58-!-nordenm [] has quit []
09:58<gbee>jams: mind opening a ticket for the plugin build problems? Assign it to me and I'll deal with it soon, I'm just a little puzzled that you seem to be the only person seeing these problems with current trunk and I'm wondering if someone else has a clue what change might have started it
10:16-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
10:27-!-nordenm [] has joined #mythtv
10:56<janneg>Chutt, kormoc: as far as I know mythweb doesn't need to be fixed since it uses ProgramInfo and no direct database access
10:58<janneg>mythweb against a patched backend still displays the flags
11:29-!-gnome42 [] has joined #mythtv
11:38<gbee>the additional information on programmes is provided by xmltv as well as EIT, I believe xris also thought there might be a possibility of that information being made available through SD (think surround sound info is already in the raw data) so I'm amused by fact that Bruce keeps refering to SD and calling the EIT code 'broken'
11:39-!-MrGandalf [] has quit [Read error: 110 (Connection timed out)]
11:41<gbee>as far as maintaining the scheduler, that's a role he chose for himself (and jealously guards), if he feels like he is expected to maintain it and objects to changes which cause him more work, then that's his own fault
11:43-!-mattwire [] has quit [Read error: 113 (No route to host)]
11:47-!-Vaelys [i=awong@slammer.cs.Dal.Ca] has quit ["leaving"]
11:55-!-nordenm [] has quit []
11:55-!-okolsi [n=mythtv@] has joined #mythtv
12:06-!-Vaelys [i=awong@slammer.cs.Dal.Ca] has joined #mythtv
12:06-!-mattwire [] has joined #mythtv
12:08-!-onixian [n=xian@] has joined #mythtv
12:16-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
12:16-!-clever[rev] [] has quit [Read error: 110 (Connection timed out)]
12:18<stuarta>i can't see how the patch on 4622 can do much
12:19<stuarta>it just changes de-reference objects for pointers, which at the end of the day is the same thing
12:25-!-clever[rev] [] has joined #mythtv
12:26*gnome42 (who's trying to get a grip on reference stuff) looks at #4622 patch
12:27*stuarta hopes he's not talking rubbish
12:31-!-danielk22 [] has left #mythtv []
12:34<stuarta>janneg: maybe a patch to libavcodec's decoder to make it set AVFrame.pts is in order
12:34-!-danielk22 [] has joined #mythtv
12:34<stuarta>evening danielk22
12:35<stuarta>danielk22: dunno if i'm reading the patch on 4622 wrong, but to me it doesn't do anything
12:36<stuarta>passing objects by reference and pointers to them are effectively the same are they not?
12:36<danielk22>yep, a reference is just a pointer which tries to keep you from shooting yourself in the foot
12:36*stuarta goes back to catching up on todays -dev list traffic
12:37<danielk22>I honestly haven't looked at that patch yet. But any change like that needs to be delayed until after 0.21 anyway..
12:38<stuarta>sure, just no point keeping it open if it doesn't do anything, but another pair of eyes is welcome
12:40<janneg>I don't see a reason within the patch to use pointers instead of a const references
12:40<stuarta>me either
12:41*stuarta sees kormoc has been a very busy boy
12:41<danielk22>hmm yes it looks like this is a no-op patch..
12:41<janneg>I would guess the submitter has almost no C++ experience beyond C
12:43<danielk22>ok, it's closed w/comment now.
12:44<danielk22>stuarta, janne: thanks for looking at that..
12:44<stuarta>np, it's what we tend to do in here :)
12:45<danielk22>i worked pretty hard to keep memory copies to a minimum using C++ references in the mpeg stuff..
12:45<danielk22>I thought he was addressing some of the still existing memory copies..
12:46<stuarta>can't be many of them?
12:51<janneg>stuarta: setting AVFrame.pts in the decoder isn't that easy as I thought when I made the comment in #4621
12:51<stuarta>why am i not surprised...
12:52<stuarta>fyi, i noticed that in one of my recordings that mythtranscode doesn't like the pts warps backwards
12:52<stuarta>haven't debugged it fair enough to know if that's actually the problem
12:52<janneg>stuarta: that's expected with b-frames
12:53<stuarta>since they have a reference to the previous frame?
12:55<janneg>no, b-frames (bidirectional) have reference to a following frame and hence have to be decoded out of order
12:55<stuarta>hmmm, i need to brush up on my mpeg decoding
12:55<janneg>they are stored in decoding order
12:55<stuarta>broadcast too?
12:56*stuarta digs out the logfile
13:01<kormoc>stuarta, just a tad
13:01<stuarta>think that gets the award for most bug fixes in one weekend
13:02-!-jmpdelos [] has joined #mythtv
13:02-!-janneg [] has quit [Remote closed the connection]
13:05-!-janneg [] has joined #mythtv
13:14-!-onixian [n=xian@] has quit [Read error: 104 (Connection reset by peer)]
13:15-!-splAt1 is now known as splat1
13:19-!-MrGandalf [] has joined #mythtv
13:27-!-mattwire [] has quit [Read error: 113 (No route to host)]
13:28-!-mattwire [] has joined #mythtv
13:55-!-okolsi [n=mythtv@] has joined #mythtv
13:58<justinh>heh. I can't wait to get back to work tomorrow. housework, dog walkies, gardening, cleaning all the windows, washing the dog.... Zzzz having a bottle of wine at lunch didn't help :P
13:58<justinh>day of rest indeed!
13:59*kormoc laughs
14:16-!-okolsi [n=mythtv@] has quit [Read error: 110 (Connection timed out)]
14:17-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
14:18<MrGandalf>gardening? heh, I'm looking at an ice storm today which is said will leave an inch of ice on everything.
14:19<kormoc>But! With a wave of your staff...
14:19<MrGandalf>if only..
14:19-!-dekar1 [] has joined #mythtv
14:34-!-xris [] has joined #mythtv
14:35-!-dekarl [] has quit [Read error: 110 (Connection timed out)]
14:35<gbee>unseasonably warm over here at the moment, if I were asked to guess the month based on the weather over the last 3 days I'd have said it was mid April.
14:36<sphery>I'm thinking that ping command (in util.cpp) used by mythcontext to check if the dbhost is accessible is going to cause a lot of issues when 0.21 is released. For users with GNU inetutils version (which uses -t for packet type, not timeout), the fallback to a ping -c 1 (without -t) will run forever (it really never times out) if the user has a firewall that drops instead of rejecting ICMP pings.
14:36<sphery>Guess everyone will have to at least enable ping from myth hosts to the db host.
14:36<gbee>yeah, don't like the ping stuff because it makes assumptions
14:37<sphery>I can't think of a better solution, though.
14:37<gbee>I had to change my firewall just to let it work
14:37<kormoc>it just pings? Why not do a connection attempt ourselves?
14:37<gbee>what was the purpose of using ping anyway?
14:37<sphery>Assuming mysqladmin is there (for a mysqladmin ping) is not good, either.
14:37<sphery>it's used to verify the db host is reachable before attempting the db connection.
14:37<gbee>I'd have said a connection attempt would be a lot more useful and reliable
14:38<jams>kormoc- i think the idea is that ping is faster then waiting for a timeout.
14:38<kormoc>well, if you just do a socket connection, you can set the timeout
14:39<kormoc>(unless they're running mysql over ssl or the like)
14:39<sphery>we can't control the connection timeout?
14:41<kormoc>should be able to
14:41<sphery>Hmmm. Seems there's also a DBHostPing=no in mysql.txt (don't know if it has an analog in config.xml)
14:42<gbee>yeah, which is fine as a default, but not something we can expect users to find easily during a first install
14:42<sphery> added the ping, added DBHostPing=no
14:43<gbee>the point of autodiscovery is to avoid the need for that configuration stuff
14:45<xris>gbee: except for people who don't/can't use it because of routing setup,etc
14:46<gbee>xris: yeah, in which case it falls back to the old methods, but I don't consider blocking ICMP pings on a network that unusual a setup - at the very least it's one case we can anticipate and work around by just not using ping
14:47<sphery>Yeah. Really, the only port on a MySQL server we know is open is the DB port. I like the idea of making a MySQL connection with a very short timeout (if that's possible--even if we throw it out). Wonder what mysqladmin's ping code looks like...
14:47<xris>gbee: mythbox in dmz, workstation in loc
14:47<kormoc>sphery, you don't even need to use mysql's connection stuff
14:48<kormoc>sphery, just open a raw socket to it and see if you get back a mysql blah string
14:49<sphery>That doesn't have negative consequences on the DB server? Sounds scary...
14:49<kormoc>Not at all
14:50<sphery>Though the mysqladmin ping code is all using the mysql libs, so useless to us.
14:50<sphery>would be too low level
14:51-!-nordenm [] has joined #mythtv
14:51<kormoc>sphery, when you connect to the server, it won't do anything until you send the proper auth credentials, so it'll just toss the connection after a sec or two.
14:51<sphery>That might go on my list of post-0.21 fixes. Unfortunately, my 3 weeks of business travel starting tomorrow is going to keep my away from heavy Myth coding/testing.
14:52*kormoc nods
14:54<gbee>anyone seen a weird scheduler bug like the following? I've three cards in the backend, two DVB-T and one PVR-150, three recordings are scheduled for the same time - 2 on the DVB-T source and a third on the PVR source - scheduler decided not to record the PVR programme because it conflicted with the two recordings on the DVB source!! Manually setting an override worked and it's now going to record all three ...
14:57<xris>so if any US people want invites to (I'd love to see this get myth integration someday), I have a few to give away. unfortunately, it doesn't work outside of the US.
14:57<kormoc>So here's a fun one as well. When we have a 'record this showing' schedule and say mythfilldatabase runs and changes the start time, we currently don't update the schedule and thus it no longer matches and doesn't record
14:58<kormoc>I sorta wonder if we should update the schedules as well when they change
15:08<xris>the problem is that there's no way to know that it's the same show
15:08<xris>unless we're using TMS data
15:08<xris>would have to add some hefty logic to get it to scan relative times with the same title/subtitle
15:09<xris>"within 15 minutes" or something like that.
15:14<kormoc>oh right, it's a wipe and re-insert, not a update
15:22-!-beata-- [] has quit [Read error: 110 (Connection timed out)]
15:23-!-aeha [] has joined #mythtv
15:26<sphery>gbee: I'd guess that's the "only one move for conflicts" thing. (see ). It's not really a bug just a corner case that could only be handled with more iterations, which were considered not worth the extra resource usage. More at (it was a broken thread).
15:30<gbee>sphery: sound similar, but refers to a situation where he has three cards but the recordings fall on the source covered by just two of them
15:32<gbee>in this case one recording was marked as conflicting with two schedules, even though it was on a completely different source and card
15:36-!-briand [] has quit [Connection timed out]
15:36-!-briand [] has joined #mythtv
15:41<janneg>bah, I don't get audio output on my notebook
15:56-!-okolsi [n=mythtv@] has joined #mythtv
15:56-!-jarle [] has quit ["Leaving"]
15:57-!-jarle [] has joined #mythtv
16:01-!-turbo [] has joined #mythtv
16:03-!-briand [] has quit [Read error: 110 (Connection timed out)]
16:13-!-cattelan_away [] has quit ["Terminated with extreme prejudice - dircproxy 1.2.0"]
16:25-!-foo8ar_ [] has joined #mythtv
16:25-!-foo8ar [] has quit [Read error: 104 (Connection reset by peer)]
16:27-!-joobie [] has joined #mythtv
16:29-!-czth_ [n=dbrobins@nat/microsoft/x-b4ceed433dc43d71] has joined #mythtv
16:31-!-mattwire_ [] has joined #mythtv
16:32<janneg>hmm, [15893] fixed it
16:37<danielk22>that's odd...
16:37<Chutt>danielk22, when do you want to make a branch for the release?
16:38<danielk22>I'm ready now..
16:38-!-onixian [n=xian@] has joined #mythtv
16:39<Chutt>anything else big to go in?
16:39<Chutt>aside from whatever fix makes both janneg and bruce least unhappy? =)
16:39*stuarta chuckles
16:39<danielk22>AFAIK the sound stuff was the last big thing other than the program flags..
16:40<danielk22>And I just committed that..
16:40<Chutt>bug's still open =)
16:41<danielk22>yeah, well I want to check out the patch for MythMusic, but it's only a few lines of code..
16:41-!-turbo [] has quit [Read error: 110 (Connection timed out)]
16:41<danielk22>plus, I'm sure some tweaks will be needed for mixing volumes..
16:42-!-turbo [] has joined #mythtv
16:42<Chutt>new bugs
16:42<Chutt>i'm updating right now
16:42<danielk22>new bugs.
16:42<Chutt>for new issues arising from the patch merge =)
16:42<stuarta>so we are branching for a feature freeze to bugfix for a few weeks?
16:42<Chutt>stuarta, that's what i'm thinking, yeah
16:43-!-mattwire [] has quit [Read error: 113 (No route to host)]
16:43<stuarta>works for me
16:43<Chutt>i'd really like to get the preview issues in mythweb worked out
16:43-!-jmpdelos [] has quit [Read error: 110 (Connection timed out)]
16:43-!-_gunni_ [] has joined #mythtv
16:43<Chutt>i can't load the recorded program screen without essentially killing mythbackend
16:43<stuarta>that was at the top of the "must be fixed" list
16:43<Chutt>danielk22, i'd _really_ like to see a reworked video settings screen
16:44<Chutt>the current stuff is really arcane
16:44<danielk22>Really? :)
16:44<Chutt>kinda unuseable unless you know what it's doing already
16:44<danielk22>Maybe some more default profiles?
16:44<Chutt>better names
16:45<danielk22>w/better names..
16:45<Chutt>and maybe a better default selection
16:45<Chutt>to match what the old stuff was closer
16:45<Chutt>so people don't email and say 'it _used_ to work'
16:45-!-czth__ [n=dbrobins@nat/microsoft/x-ebc888994a5edf56] has quit [Connection timed out]
16:46<danielk22>yeah, I'm not terribly proud of CPU++/CPU+/CPU--, I kind of thought people would create their own profiles and submit them. But that hasn't happened obviously..
16:46<Chutt>3gsm (mobile phone show in spain) starts tomorrow, so i'm actually finally going to have a bit of time
16:46-!-saxin [] has joined #mythtv
16:46<janneg>I would like to fix the software encoder bitrate bug before 0.21 but I've no idea what it caused
16:46<Chutt>audiooutputdigitalencoder.cpp:10:24: error: a52dec/a52.h: No such file or directory
16:46<Chutt>audiooutputdigitalencoder.cpp:10:24: error: a52dec/a52.h: No such file or directory
16:46<Chutt>doh, double paste
16:48<Chutt>config.h needed moved up
16:49<xris>Chutt: gbee has the most knowledge of the backend side of that stuff
16:49<xris>(pixmap stuff)
16:49<danielk22>I think gbee wrote here a couple days ago that he had it figured out..
16:52<gbee>I haven't unfortunately, been too busy to really look into it but I got as far as deciding that the problem lay somewhere in the http server code instead of the preview request handling function
16:52<Chutt>he had one failure figured out
16:52<Chutt>i thought =)
16:53<kormoc>While I've been poking at making mythweb validate as html 4.01 strict, does anyone mind if I update the http server to also validate to html 4.01 strict?
16:53<danielk22>I created #4631 for the video profiles..
16:53<Chutt>go for it
16:54<kormoc>Kk, nifty
16:54<gbee>Chutt: I've already fixed a couple of bugs in that area, but they were outstanding before the most recent one appeared, the 100% cpu utilisation
16:54<danielk22>ok, I gotta run. As far as doing the branch, the things I have remaining on my plate are all bugs, no features...
16:54<Chutt>i'll send an email to the list
16:54<Chutt>asking if anyone else has anything
16:55<gbee>we planning to branch already? I was though I might have another few days - mythweather still needs attention
16:55<Chutt>for non-bugfix stuff
16:56<Captain_Murdoch>I'd like to get sphery's DB backup and DB schema version check stuff in before 0.21. I can probably get those in tonight. I haven't had a chance to apply his patches to my tree and test yet.
16:56-!-leprechau [] has quit [Remote closed the connection]
16:56<Chutt>oh, the db schema stuff needs to go in
16:56<gbee>well if mythweather counts as a bug fix, then that's ok - all I've got left are bugs, I can push back the xmltv changes to 0.22 because I don't think I'll have time to work on that at all in the next 2 weeks
16:58<gbee>there is one contrib addition I'll try to get in tonight, can't review it, but I'll assume that it works (itunes playlist importer for mythmusic)
16:59<Chutt>that was weird
16:59<Chutt>make install just relinked all the programs/, before installing
17:02-!-|gunni| [] has quit [Read error: 110 (Connection timed out)]
17:09<sphery>Captain_Murdoch: Doing an update to the DB backup one--forgot to use --host, so it fails if DB host is not local...
17:10<superm1>i think that the linking done needs to be cleaned up a bit. There are a lot of libraries that are unnecessarily linked to binaries
17:10-!-joobie [] has quit ["This computer has gone to sleep"]
17:11-!-joobie [] has joined #mythtv
17:12<Chutt>superm1, i've never really trusted linking libraries to libraries
17:12<Chutt>doesn't work everywhere
17:12<Chutt>and that's all those are, basically
17:12<superm1>why is that?
17:13<kormoc>isn't that fixed by the --link-as-needed flag thing?
17:13<stuarta>didn't --link-as-needed get removed to fix those damn circular deps?
17:13-!-jmpdelos_ [] has joined #mythtv
17:13<Chutt>all those are dependencies of other libraries it _is_ necessary to link to
17:14<superm1>well additionally there are a few libraries that are linked unnecessarily to other libraries like line 603 and 604 of that same pastebin
17:14<kormoc>stuarta, I thought it went the other way, by removing the circular deps, it allows the use of --link-as-needed
17:15<stuarta>entirely possible
17:15<superm1>so then perhaps adding --link-as-needed would be a feasible solution by default?
17:16<Chutt>link-as-needed doesn't do anything for this, iirc
17:16<Chutt>cleaning up the link lines would
17:16<Chutt>but it's really not a big deal
17:16<stuarta>nn all
17:16<superm1>wouldn't there be an advantage in load time, with not having to load some of those libraries (if they weren't being used already)
17:17<Chutt>superm1, i'm pretty sure those are all going to be used already
17:17<Chutt>so, not really =)
17:17<Chutt>see, we just use the same link line for extra libraries everywhere
17:17<kormoc>superm1, prelinking might help if you don't already do that
17:17<Chutt>_could_ just add that to the appropriate libraries
17:17<Chutt>and only link the apps to the myth libraries
17:18<superm1>kormoc, we have some other solution that we use distro wide other than prelinking. i forget what it is
17:18<superm1>(those warnings are just generated at build time, so they are meant to be passed on upstream_
17:19<Chutt>i'm of the opinion that it's pretty innocuous
17:19<Chutt>since the other libs are going to pull those in
17:19<superm1>well the case that comes to mind is headless boxes
17:19<superm1>that are pulling in the X libraries unnecessarily
17:22<Chutt>libmythtv's linked against X
17:22<Chutt>all those apps are linked against libmythtv
17:24<Chutt>janneg, i'd consider that a go on the patch in 4615, then.
17:31-!-mattwire_ [] has quit [Read error: 113 (No route to host)]
17:33-!-mattwire [] has joined #mythtv
17:42<sphery>Captain_Murdoch: --host argument included in mythtv-database_backups_before_upgrade.3.patch . I'm pretty sure that's all I've forgotten (at least I can't remember anything else I've forgotten :).
17:45-!-ahbritto [] has joined #mythtv
17:45-!-ahbritto [] has quit [Remote closed the connection]
17:45-!-ahbritto [] has joined #mythtv
17:47-!-mattwire [] has quit ["Leaving"]
17:56-!-onixian [n=xian@] has quit [Success]
17:57<gbee>I can't see how it breaks anything for DD users to get rid of those columns, I might buy the power search arguments, though IMHO that just means we need to improve the UI there e.g. Display a combobox of possible values for those columns
17:58<gbee>anyhow, probably best to just leave it now, both sides are getting what they want, or close enough
17:58<janneg>that still won't fix existing rules
17:59<gbee>janneg: those can be fixed with a search/replace during a schema update
18:02-!-onixian [n=xian@] has joined #mythtv
18:05<gbee>in a way though, that's why giving users the ability to directly enter SQL was probably not the best decision, we get stuck with the current schema - it could have been done through itermediary 'tokens' which are easy for users to understand and mean we can change the underlying schema without breaking things
18:06-!-brot [] has joined #mythtv
18:07<kormoc>Well, given anyone who has the ability to write a power search likely has the ability to update it as needed as well, it's not that bad of thing is it?
18:07<brot>nice to see that there are many devs, i wanted to drop a line and say that you really help a great project. thanks everyone for contributing :)
18:09-!-_gunni_ [] has quit ["KVIrc 3.2.4 Anomalies"]
18:10<gbee>brot: thanks
18:12*justinh looks for the cheque. must still be in the post...
18:13*justinh gets his coat & heads home to bed
18:14<brot>goodnight :)
18:16<janneg>Chutt: yeah, I'll apply it tomorrow
18:22-!-MavT [] has joined #mythtv
18:25-!-beavis [] has quit ["Verlassend"]
18:26-!-turbo is now known as briand
18:27-!-robert__ [] has quit [Remote closed the connection]
18:32-!-reynaldo_ [] has quit [Read error: 110 (Connection timed out)]
18:33-!-onixian [n=xian@] has quit [Read error: 104 (Connection reset by peer)]
18:39-!-MaverickTech [] has quit [Read error: 113 (No route to host)]
18:43-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Success]
19:14-!-ahbritto [] has quit [Client Quit]
19:15-!-ahbritto [] has joined #mythtv
19:27-!-PointyPumper [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
19:33-!-joobie [] has quit ["Leaving"]
19:45-!-splat1 is now known as splAt1
19:46-!-moodboom [] has joined #mythtv
20:02-!-davilla [] has left #mythtv ["Leaving"]
20:14-!-beata [] has joined #mythtv
20:33-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
20:46-!-ahbritto [] has quit [Client Quit]
21:00-!-brot [] has quit ["KVIrc 3.2.6 Anomalies"]
21:28-!-PointyPumper [i=Pintlezz@] has joined #mythtv
21:35-!-Pumpernick [i=Pintlezz@] has joined #mythtv
21:37-!-Pumpernick [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
21:38-!-PointyPumper [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
21:38-!-PointyPumper [i=Pintlezz@] has joined #mythtv
21:43-!-PointyPumper [i=Pintlezz@] has quit [Read error: 104 (Connection reset by peer)]
21:44<Chutt>this is weird..
21:51-!-JoeBorn [] has joined #mythtv
21:56-!-ahbritto [] has joined #mythtv
21:58-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Read error: 104 (Connection reset by peer)]
21:59-!-PointyPumper [i=Pintlezz@] has joined #mythtv
22:22-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
22:24-!-jhulst [n=jhulst@unaffiliated/jhulst] has quit [Remote closed the connection]
22:33<Captain_Murdoch>sphery, you around?
22:36-!-PointyPumper [i=Pintlezz@] has quit [Read error: 110 (Connection timed out)]
22:36<sphery>Captain_Murdoch: Yeah.
22:37<Captain_Murdoch>RE: the password file for mysqldump. mysql will block out the password when you specify it on the command line, it isn't visible if you run a ps. is it visible somewhere else? I assume mysqldump does the same thing, but I haven't checked.
22:37<Captain_Murdoch>cpinkham 23608 3.0 0.1 7256 3052 pts/10 S 22:32 0:00 mysql -u mythtv -h mythdb -px xxxx mythdev
22:39<sphery>Seems so ... mysqldump -px xxxx
22:39<sphery>I mainly did the password file because of the MySQL docs...
22:39<Captain_Murdoch>ok, if they recommend it, then that's ok.
22:39<sphery>I'll get a ref... One sec.
22:41<sphery>Says that the blanking leaves a small period where it's visible and it's ineffective on some systems (SystemV Unix and perhaps others)...
22:42<sphery>I don't care either way. Mainly did it to keep the armchair quarterbacks from yelling. :)
22:42<Captain_Murdoch>yeah, makes sense. ok. I'm about to run it for the first time here. opening up a term with a long scrollback buffer since I can't redirect stdout to a file if I want the upgrade prompt.
22:43<sphery>Yeah. If your DB is on 1209, you can just set DBSchemaVer to 1208. The upgrade will fail without changing anything...
22:43<Captain_Murdoch>my dev DB gets copied from my production DB every night during my nightly backup script, so I'm about 15 revisions behind since I haven't upgraded production in a while.
22:44<sphery>Then I guess you don't have to fake it. ;)
22:46<Captain_Murdoch>ran this from the command line and it still didnt' ask me if I wanted to upgrade. I thought I was right the other day when I said I'd never seen that question.
22:46<Captain_Murdoch>backup file is only 2333 bytes also. hmmm
22:46<sphery>Any thoughts on the native SQL one. I could get that working post 0.21, but it will probably have issues--there's a lot of version differences in the SHOW commands needed to get lists of tables and columns and CREATE TABLES and stuff.
22:47<sphery>Do you have the "DBSchemaAutoUpgrade" setting specified?
22:47<sphery>1 or -1 would auto-upgrade.
22:48<Captain_Murdoch>to be honest, we (aka you) could spend a lot of time trying to get that right and be tweaking it forever to get a proper restore. we can be pretty sure that mysqldump will spit out a good file always. I'm not sure if I see the benefit in trying to do it native.
22:48<sphery>I have to admit that I've never gotten the prompt (until testing the backup and schema patches--and even then it was only for mythtv-setup and mythfrontend).
22:49<Captain_Murdoch>no setting here, it says I"m not in an interactive shell when it's just a regular xterm running directly on my dev box.
22:49<sphery>That's cool. The only down side is that it requires users wanting auto backup to install mysqldump. I guess when I write the code for putting it in housekeeper, I'll put in a check for the executable, and then we can leave out warnings if it's a housekeeping run.
22:50<sphery>but since I'm guessing package management systems have mysql client utilities separate from the server, it's probably not even a big package.
22:50<Captain_Murdoch>well, odds are if they don't have mysqldump already installed then they're not doing backups anyway. people doing remote backups are probably the exception not the rule.
22:51<sphery>Yeah. True. Looks like Nigel used: !isatty(fileno(stdin)) || !isatty(fileno(stdout) for the interactive check... Did you redirect to a file or something?
22:51-!-reynaldo [] has joined #mythtv
22:52<Captain_Murdoch>no, mythbackend --nosched --nohousekeeper --nojobqueue --noautoexpire
22:52<sphery>libs/libmyth/mythcontext.cpp +971
22:52<Captain_Murdoch>yeah, looking at that already.
22:55<Captain_Murdoch>my little test program says both of those return 1 for me. so it should have prompted me. I'll test with some debug code later. need to see why my backup file was only 2333 bytes.
22:56<sphery>That doesn't sound good... :(
22:56<Captain_Murdoch>moved the version check down below the backup and recompiling now.
22:56<Chutt>ah, hah
22:56<Chutt>old version of mythlcdserver never got killed off
22:57<sphery>I chose mysqldump params that should work on all 5.x versions of mysqldump...
22:58<Captain_Murdoch>ah, that's the issue. :) My mysqldump client is 3.23.58 on my dev box. DB is 5.0.22 though on the DB server.
22:58<sphery>Oh. That would probably do it...
22:58<Captain_Murdoch>--comments isn't valid for me. easy enough to test the rest of it.
22:58<sphery>OK. One other downside to using mysqldump vs native SQL... :)
22:59<sphery>comments isn't really required...
22:59<sphery>IIRC, it's default in 5.x. Docs say, "Write additional information."
23:00<sphery>Yep. Enabled by default. Let's take it out to make it more reliable. (Er, please, if you don't mind the extra work...)
23:00<Captain_Murdoch>had to take out --set-charset temporarily for me to test also. :) no biggie. I need another excuse to upgrade my dev system.
23:01<sphery>also enabled by default.
23:01<Captain_Murdoch>so take out --comments
23:01<sphery>and set-charset
23:01<Captain_Murdoch>will do.
23:01<sphery>If a user disables them in their my.cnf, they probably know what they're doing (fingers crossed... ;)
23:02<sphery> shows details not in --help
23:02-!-davilla [] has joined #mythtv
23:03<Captain_Murdoch>I'm going to add seconds to the last run times in the settings table, is that OK?
23:04<sphery>Definitely. I tried to keep it like mfdb (and, with SD, even for mfdb adding seconds makes sense, now).
23:05<sphery>Wonder if it's possible to do a "backwards-compatible" change to the mfdb times post 0.21... (On my TODO--or TOfindout)
23:05<Captain_Murdoch>mine finished in less than 1 minute, so at first I thougth that you used the same timestamp twice, then I saw it was currentdatetime without the seconds displayed.
23:05<sphery>Yeah. I did 2 tests in < 1 minute and it took me a moment to figure out why I didn't get another backup...
23:05<Captain_Murdoch>19 seconds for the backup.
23:06<sphery>Nice. 29 for me
23:06<sphery>Only 19s CPU time, but not bad (especially since my "exact" times were all done during recordings)
23:06<sphery>I still can't believe the difference between bzip2 and gzip on SQL compression.
23:10<janneg>sphery: try lzma
23:10<sphery>That's 7-zip's internal one, right?
23:12<Captain_Murdoch>sphery: I'm also wrapping the call to dbutil.BackupDB() in a #ifndef USING_MINGW for now to squelch any bug reports. after 0.21, we can remove the #ifdef to let the win32 guys test and provide any necessary patches.
23:13<sphery>Haven't tried it. I guess it must be fast, though. Fortunately, gzip is probably good enough for these files (i.e. my 180MB backups go to 31MB) and very likely to exist on (*nix) systems.
23:15<sphery>Captain_Murdoch: Sounds good. I don't know what the "approved" approach is now with the Windows port, and didn't even really consider that some of that would have to be modified for Windows. (I just hope the approved approach won't involve my learning a bunch of stuff I don't care to learn about Windows.)
23:15<janneg>sphery: no, it's not fast, but it compresses my database to 1/2 of the bzip -9 size
23:15<kormoc>janneg, I just spent 2 hours getting lzma working. Bloody hells can that be freaking messy
23:16<janneg>kormoc: emerge lzma-utils works fine
23:16<sphery>janneg: I see. That's cool. bzip2 made backups about 30% smaller than gzip. 1/2 that again is impressive.
23:16<sphery>We were going for fast and not too resource intensive.
23:17<kormoc>janneg, Sadly I had a old version of gcc. I'm using 4.1.2 in my profile, but I had 3.4.6 installed. lzma insisted in linking with 3.4.6 and then couldn't read my libstdc++ which was created with 4.1.2
23:17<kormoc>so sadly, emerge lzma-utils doesn't always work fine :P
23:19<sphery>Captain_Murdoch: Things going OK? I'm thinking about packing for my trip and going to bed so I can wake up in 4 hours for an early-morning trip to the airport, but don't want to leave you hanging.
23:19<Captain_Murdoch>sphery: yeah, about to commit, just did my last test and made a diff to copy over to my pristine tree so I can commmit.
23:20<sphery>Captain_Murdoch: Thanks for reviewing the patch. Though it's not nearly as complete as I wanted for 0.21, it's better than nothing. And, I apologize for making it harder on yo
23:20<sphery>u by submitting it at the last minute.
23:21<Captain_Murdoch>no problem... :) I've thought about something like this for a while, so am glad to see the work done on it. after this is in, I'm looking at the other one.
23:21<Captain_Murdoch>the schema version check
23:23<sphery>Hope that one (since it's much smaller) has fewer areas of concern/questions. I'll try to check scrollback whenever I get a chance tomorrow, though. Later.
23:25<Captain_Murdoch>no real concerns, I just like to test things before committing, no matter how simple they are. I don't like it when trunk isn't stable so I don't want to be the one that causes an issue by committing something. :)
23:25<kormoc>have a good trip Mr. sphery.
23:27-!-jhulst [n=jhulst@unaffiliated/jhulst] has joined #mythtv
---Logclosed Mon Feb 11 00:00:01 2008