Beirdobot, irc.freenode.net

Thursday, July 21st, 2011, 00:00 UTC

[00:00:21]

Beirdo:

heh

[00:00:26]

wagnerrp:

so for mythbackend/frontend/filldatabase, that is a broken configuration

[00:00:41]

wagnerrp:

and possibly mythfillmetadata

[00:00:44]

Beirdo:

anyways, it would be smarter to default to packaging with /var/log/mythtv/

[00:00:54]

Beirdo:

and then log everything in there

[00:01:18]

Beirdo:

and make that dir owned by the user who runs mythtv

[00:01:27]

Beirdo:

or use group perms, etc

[00:02:09]

sphery:

though, IMHO, if distros want logs in /var/log, they should configure it to use syslog with appropriate syslog configuration

[00:02:17]

Beirdo:

now, something will want to clean that directory out... killing off or archiving old files > 1 month old or whatever

[00:02:34]

Beirdo:

yeah, or just use syslog, but that doesn't bring all the debug details

[00:02:49]

Beirdo:

it could if we want

[00:02:57]

sphery:

heh, I'd like it to :)

[00:03:22]

Beirdo:

and db logging as well, not all the debug details, but could be extended to save that too

[00:03:44]

Beirdo:

and a log viewer... :)

[00:04:05]

sphery:

yeah... I'm not so concerned about db logging getting the details since that's more of a temporary storage location, but it wouldn't hurt. And, yeah, I need to get that done. :)

[00:04:32]

Beirdo:

if you want more details so it's more complete, the data's available, just not put into the db

[00:04:40]

Beirdo:

it's a simple change to extend

[00:09:02]

cattelan (cattelan!~cattelan@c-66-41-26-220.hsd1.mn.comcast.net) has quit (Ping timeout: 252 seconds)

[00:23:28]

wagnerrp:

danielk22: that may be our single most efficient commit ever

[00:23:40]

wagnerrp:

deletion of a single character took two bugs off trac

[00:25:05]

iamlindoro:

That's nothing, I can create ten by adding one

[00:25:17]

iamlindoro:

That's the same, right?

[00:28:41]

pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has quit (Quit: Leaving.)

[00:31:52]

kormoc is now known as kormoc_afk

[00:38:33]

jams_ (jams_!~jams@cpe-184-58-217-97.wi.res.rr.com) has joined #mythtv

[00:42:10]

jams (jams!~jams@cpe-184-58-217-97.wi.res.rr.com) has quit (Ping timeout: 264 seconds)

[00:50:28]

tgm4883:

wagnerrp, so is my understanding of this that you are removing --logfile and you will just write to /var/log/ then?

[00:50:51]

brfransen (brfransen!~brfransen@216.254.250.47) has quit (Ping timeout: 258 seconds)

[00:51:12]

tgm4883:

I agree with Beirdo, /var/log/mythtv/ is nicer

[00:54:06]

wagnerrp:

thats what we are considering

[00:55:16]

wagnerrp:

not that we would be saving to /var/log/mythtv, but rather we would only accept a directory on the command line, not a filename

[00:56:03]

wagnerrp:

if nothing is defined, it just dumps to stdout as is currently the case

[00:57:01]

tgm4883:

ah

[00:57:23]

tgm4883:

I see, so instead of setting it to --filename=/var/log/mythtv/mythbackend.log we would just set /var/log/mythtv/

[00:57:42]

wagnerrp:

well, it would be --logpath=/var/log/mythtv

[00:57:51]

wagnerrp:

and then mythbackend would decide where it wanted to log to

[00:57:58]

wagnerrp:

based off the date and PID apparently

[00:58:34]

wagnerrp:

that might cause some difficulty with whatever application you have set up to attach the logs to bug tickets

[00:58:50]

tgm4883:

hmm

[00:59:04]

tgm4883:

mrand, ^^

[00:59:19]

tgm4883:

I'm not entirely sure how apport does that

[00:59:31]

tgm4883:

would definitly need to fix up -log-grabber though

[00:59:49]

tgm4883:

wagnerrp, and that is for 0.25?

[00:59:59]

wagnerrp:

thats for... nothing yet

[01:00:05]

wagnerrp:

the decision has not been made

[01:00:15]

wagnerrp:

figured you guys might have an opinion

[01:01:01]

wagnerrp:

the issue is that with the new logging, child processes no longer log to the parent's log file

[01:01:38]

wagnerrp:

so if mythbackend fires off mythcommflag, you dont have any notice of what its doing

[01:02:04]

wagnerrp:

if you give it a path, it will log to its own file within that path

[01:02:15]

tgm4883:

hmm

[01:02:21]

wagnerrp:

its a significant change from the old behavior that will throw people off

[01:02:37]

tgm4883:

I think it's rather nice to have a single log file for backend type processes vs frontend type processes

[01:02:39]

wagnerrp:

so if we make a clean break, it will force people to learn the new behavior

[01:03:48]

davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)

[01:04:14]

davide_ (davide_!~david@host103.16.intrusion.com) has joined #mythtv

[01:04:14]

davide_ (davide_!~david@host103.16.intrusion.com) has quit (Changing host)

[01:04:14]

davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv

[01:19:20]

brfransen (brfransen!~brfransen@216.254.250.47) has joined #mythtv

[01:26:57]

sphery:

tgm4883: we also have support for syslog in master--so with an appropriate syslog config, you could log it wherever you like

[01:33:22]

sphery:

but we do need to separate out backend log output from child process log output so we don't get users thinking that "backend deadlocks up with Protocol version check failure"

[01:34:12]

sphery:

though with syslog, you could log it all to one file...

[01:43:45]

superm1:

Beirdo, yeah that's what we do in *buntu, make a /var/log/mythtv and put all myth related items there

[01:49:04]

wagnerrp:

superm1: this would just mean mythtv decides where in that folder to store stuff, rather than you

[02:13:41]

kormoc_afk is now known as kormoc

[02:20:05]

iamlindoro:

You know you've completely lost perspective when you're waiting for "this silly recording to finish" so that you can test your latest changes

[02:48:30]

xris (xris!~xris@mythtv/developer/xris) has quit (Quit: Terminated with extreme prejudice - dircproxy 1.2.0)

[02:50:47]

xris (xris!~xris@mythtv/developer/xris) has joined #mythtv

[03:12:03]

superm1:

wagnerrp, well i would vote /var/log/mythtv so we don't have to change anything else then here :P

[03:12:53]

superm1:

this new usr/bin/mythccextractor, back, common or front?

[03:13:15]

wagnerrp:

superm1: well there wouldnt be a static folder defined, and there wouldnt be a folder even used unless the user specified it on the command line

[03:13:26]

wagnerrp:

your scripts would have to call --logpath instead of --logfile

[03:13:34]

superm1:

oh that's fine, OK

[03:13:48]

superm1:

so we would just need to add --logpath /var/log/mythtv and things would be essentially the same then

[03:13:52]

wagnerrp:

but the bigger consequence is one i mentioned a while back, the applications would decide what to name their log files on their own

[03:14:05]

wagnerrp:

i.e. mythbackend.date.pid.log

[03:14:16]

superm1:

which would really mess up log rotation

[03:14:26]

wagnerrp:

and your log grabber

[03:14:30]

superm1:

Yeah

[03:18:28]

superm1:

wagnerrp, any idea on this mythccextractor placement?

[03:19:37]

wagnerrp:

anywhere a user might want it

[03:19:45]

wagnerrp:

mythtv is not using it internally for anything at this time

[03:20:56]

wagnerrp:

likely, anywhere you have a jobqueue

[03:22:02]

kormoc is now known as kormoc_afk

[03:22:15]

superm1:

jobqueue would run on back only right? so i'll put it in back for now then

[03:22:29]

wagnerrp:

jobqueue would be mythbackend and mythjobqueue

[03:22:42]

wagnerrp:

typically, mythjobqueue would exist with a frontend only package

[03:22:45]

superm1:

oh

[03:22:48]

superm1:

okay then common it is

[03:22:51]

kormoc_afk is now known as kormoc

[03:26:19]

zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection)

[04:30:52]

Goga777 (Goga777!~Goga777@shpd-92-101-146-255.vologda.ru) has joined #mythtv

[04:42:20]

abqjp (abqjp!~abqjp@174-28-184-228.albq.qwest.net) has joined #mythtv

[05:00:49]

abqjp (abqjp!~abqjp@174-28-184-228.albq.qwest.net) has quit (Quit: abqjp)

[05:01:16]

PointyPumper (PointyPumper!~pintlezz@190.244.73.13) has quit (Ping timeout: 252 seconds)

[05:08:35]

MythBuild:

build #1791 of master-linux-64bit is complete: Exception [exception interrupted] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1791 blamelist: Robert McNamara <rmcnamara@mythtv.org >

[05:09:32]

PointyPumper (PointyPumper!~pintlezz@190.244.73.13) has joined #mythtv

[05:09:41]

iamlindoro:

whaaaat?

[05:09:46]

iamlindoro:

Wait, that's not a failure

[05:09:53]

iamlindoro:

don't blame me, you stupid robot

[05:10:50]

wagnerrp:

hahaha

[05:11:57]

jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has joined #mythtv

[05:27:52]

Beirdo:

hehe

[05:28:06]

Beirdo:

the machine is down, I probably shut it down mid build, sorry

[05:28:27]

jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has quit (Quit: Leaving.)

[05:39:15]

Beirdo:

MythBuild: force build master-linux-64-bit now

[05:39:15]

MythBuild:

no such builder 'master-linux-64-bit'

[05:39:21]

Beirdo:

MythBuild: force build master-linux-64bit now

[05:39:22]

MythBuild:

build forced [ETA 6m37s]

[05:39:22]

MythBuild:

I'll give a shout when the build finishes

[05:39:47]

Beirdo:

it should be up for at least a few hours while I transfer disk contents

[05:44:15]

jmartens (jmartens!~jmartens@109.232.42.33) has joined #mythtv

[05:46:31]

MythBuild:

Hey! build master-linux-64bit #1792 is complete: Success [build successful]

[05:46:31]

MythBuild:

Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1792

[05:57:27]

Beirdo:

there ya go, iamlindoro. You're vindicated :)

[05:57:44]

iamlindoro:

It praises you for triggering the rebuild :P

[05:57:51]

iamlindoro:

It blames me for your computer crashing

[05:57:54]

iamlindoro:

just doesn't seem fair

[05:58:21]

Beirdo:

heh, I know

[06:15:23]

Goga777 (Goga777!~Goga777@shpd-92-101-146-255.vologda.ru) has quit (Remote host closed the connection)

[06:21:28]

kormoc is now known as kormoc_afk

[06:58:26]

kth (kth!~kth@unaffiliated/kth) has joined #mythtv

[08:10:11]

mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Ping timeout: 240 seconds)

[08:10:41]

jwhite_ (jwhite_!~jwhite@75-146-153-89-minnesota.hfc.comcastbusiness.net) has quit (Ping timeout: 240 seconds)

[08:16:11]

jwhite_ (jwhite_!~jwhite@75-146-153-89-minnesota.hfc.comcastbusiness.net) has joined #mythtv

[08:16:34]

mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv

[08:36:45]

superm1 (superm1!~superm1@ubuntu/member/superm1) has quit (Read error: Operation timed out)

[08:38:46]

superm1 (superm1!~superm1@204.8.45.13) has joined #mythtv

[08:38:46]

superm1 (superm1!~superm1@ubuntu/member/superm1) has joined #mythtv

[08:38:46]

superm1 (superm1!~superm1@204.8.45.13) has quit (Changing host)

[08:44:45]

stuartm:

iamlindoro: heh, the latest xkcd is rather appropriate given what we were discussing for a standardised naming scheme for the UI – http://xkcd.com/927/

[08:48:58]

mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Ping timeout: 258 seconds)

[08:49:42]

jwhite_ (jwhite_!~jwhite@75-146-153-89-minnesota.hfc.comcastbusiness.net) has quit (Ping timeout: 255 seconds)

[08:51:45]

mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv

[08:52:31]

kth (kth!~kth@unaffiliated/kth) has quit (Ping timeout: 260 seconds)

[08:53:38]

kth (kth!~kth@unaffiliated/kth) has joined #mythtv

[09:03:13]

jwhite_ (jwhite_!~jwhite@75-146-153-89-minnesota.hfc.comcastbusiness.net) has joined #mythtv

[09:08:48]

tris (tris!tristan@2001:1868:a00a::4) has quit (Excess Flood)

[09:13:36]

tris (tris!tristan@2001:1868:a00a::4) has joined #mythtv

[09:35:41]

mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Ping timeout: 240 seconds)

[09:48:42]

mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv

[10:05:01]

mike|3 (mike|3!~mike@c-24-21-63-118.hsd1.or.comcast.net) has quit (Remote host closed the connection)

[10:05:48]

mike (mike!~mike@c-24-21-63-118.hsd1.or.comcast.net) has joined #mythtv

[10:06:14]

mike is now known as Guest90681

[10:12:32]

mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Ping timeout: 258 seconds)

[10:13:53]

stuartm:

Beirdo: http://pastebin.com/QfWe7JSP

[10:18:10]

stuartm:

I'm none the wiser

[10:24:18]

stuartm:

ah, I think I see it, deadlock in sockListLock

[10:24:48]

stuartm:

we send a response before removing the read lock

[10:40:29]

Goga777 (Goga777!~Goga777@shpd-95-53-185-111.vologda.ru) has joined #mythtv

[10:45:43]

mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv

[10:51:25]

kth (kth!~kth@unaffiliated/kth) has quit (Quit: Leaving.)

[11:11:10]

stuartm:

ok, fix committed

[11:40:53]

stuartm:

I wonder if sockListLock shouldn't be recursive, it seems less easy to screw up the read locks that way

[12:04:06]

davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)

[12:04:32]

davide_ (davide_!~david@host103.16.intrusion.com) has joined #mythtv

[12:04:40]

davide_ (davide_!~david@host103.16.intrusion.com) has quit (Changing host)

[12:04:40]

davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv

[12:07:33]

zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv

[12:25:19]

kth (kth!~kth@unaffiliated/kth) has joined #mythtv

[12:30:26]

danielk22:

stuartm: is there a wait on it? you can't wait on recursive locks in qt..

[12:33:49]

stuartm:

not on sockListLock, no

[12:35:56]

Goga777 (Goga777!~Goga777@shpd-95-53-185-111.vologda.ru) has quit (Remote host closed the connection)

[12:49:57]

mrand:

Beirdo, wagnerrp: it seems like mythbackend.date.pid.log would be treading new ground – are there many other applications that do something like that? Unless I'm missing something, automatically collecting the most recent log(s) using a script could be a challenge because there is no fixed filename(s) to go to. Why not mythbackend.main.log, mythbackend-%threadname%.log, etc?

[12:49:57]

mrand:

I don't know python well at all, so I could easily be missing how the apport hooks could handle random filenames (https://wiki.ubuntu.com/Apport/DeveloperHowTo).

[12:49:57]

mrand:

Or maybe log it all to one file, and just have the threadname on each line... that would allow approximate sequences to be tracked rather than having to track things between files.

[12:55:00]

Goga777 (Goga777!~Goga777@shpd-95-53-185-111.vologda.ru) has joined #mythtv

[12:57:11]

danielk22:

iamlindoro: Yeah, Digital Nirvana is using MythTV but it doesn't use it as a recorder for all variations of the box. DN wrote it's own DVR frontend that can operate with different recording backends so long as they produce mpeg files.

[12:58:36]

stuartm:

how would log rotation work with lots of uniquely named files?

[12:59:13]

danielk22:

iamlindoro: But DN isn't using mythccextrator yet.

[13:07:06]

PointyPumper (PointyPumper!~pintlezz@190.244.73.13) has quit (Ping timeout: 252 seconds)

[13:10:29]

Ricky-Bobby (Ricky-Bobby!~jordack@h69-131-44-221.plmomi.dedicated.static.tds.net) has joined #mythtv

[13:12:02]

Goga777 (Goga777!~Goga777@shpd-95-53-185-111.vologda.ru) has quit (Remote host closed the connection)

[13:12:30]

Ricky-Bobby is now known as Jordack

[13:14:05]

PointyPumper (PointyPumper!~pintlezz@190.244.73.13) has joined #mythtv

[13:25:29]

stuartm:

bug squashing events don't work, but we do need to organise some sort of organised ticket sifting, we actually received a patch for the bug I just fixed 8 months ago

[13:26:42]

stuartm:

a one line fix which was triaged by no less than two experienced developers but never actually reviewed/committed

[13:44:13]

iamlindoro:

I agree about the BSPs, I think they've caused more bugs than they've solved because of the pressure to work hastily. Regarding that ticket, and it's no excuse for it going that long, but I often triage bugs whose cause and solution I don't understand, and steer clear of for fear of breaking things badly

[13:44:13]

kth1 (kth1!~kth@2001:638:708:10:222:5fff:fee7:6f0f) has joined #mythtv

[13:44:23]

kth (kth!~kth@unaffiliated/kth) has quit (Quit: Leaving.)

[13:48:33]

iamlindoro:

As much as people weren't interested in the idea, it's one of the reasons I suggested a mailing list ffmpeg-style review for patches, that way people could get an upfront evaluation when they submit patches. I know there wasn't much interest, but I'd always be open to some mechanism by which we can give prompt feedback to patch submitters

[13:49:58]

wagnerrp:

stuartm: was that the cause of our strange socket errors?

[13:50:45]

wagnerrp:

where the backend stops responding to queries, and clients just get a timeout when attempting to check protocol version

[13:52:35]

wagnerrp:

mrand: honestly, i would be fine sticking to %APPNAME%.log

[13:52:57]

stuartm:

wagnerrp: one cause, from looking at open tickets there is probably another involving a different lock

[13:52:58]

wagnerrp:

my interest in the change was specific limited to the jobqueue

[13:53:32]

wagnerrp:

and the desire to have output from jobs go to their own log file

[13:54:32]

stuartm:

the other deadlock looks to be much more complicated

[13:59:48]

stuartm:

it would be nice to have more options at our disposal to catch locking problems, the problem I found could have been picked up by a decent static analysis tool

[14:00:31]

stuartm:

even just lock some lock debugging output would make it easier than interpreting a stacktrace

[14:01:35]

j-rod|afk is now known as j-rod

[14:05:56]

danielk22:

I've toyed with the idea of creating a MMutex that checks if it's being re-locked when it's not recursive and knows it's locking order in relation to other locks so it can detect out of order locking. But I have other more pressing things to tackle.

[14:07:38]

wagnerrp:

if nothing else, a quick wrapper around QMutex with a bit of logging should at least indicate where those problem areas are

[14:09:19]

stuartm:

locks do seem to cause us lots of problems, some of which are long-standing complaints

[14:09:40]

stuartm:

the email just sent to the -dev list being a timely example

[14:10:24]

danielk22:

There are 382 instances of the text QMutex in our header files...

[14:11:27]

stuartm:

that doesn't count the QReadWriteLock and other types in use

[14:11:58]

danielk22:

16 R/W locks and 4 QSemaphores

[14:13:04]

stuartm:

hmm, that's not so bad against the QMutex figure

[14:17:22]

danielk22:

stuartm: what is the other deadlock ticket? is there a backtrace?

[14:17:30]

sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Ping timeout: 240 seconds)

[14:18:54]

sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv

[14:20:25]

danielk22:

sphery: it appears my db connection problems are not gone with the latest code. :|

[14:21:49]

stuartm:

danielk22: http://code.mythtv.org/trac/ticket/9704

[14:22:03]

danielk22:

sphery: nm.. the code didn

[14:22:12]

danielk22:

't update do to an uncommitted change..

[14:23:13]

stuartm:

there are others including – http://code.mythtv.org/trac/ticket/9885 and http://code.mythtv.org/trac/ticket/9773

[14:24:00]

stuartm:

and http://code.mythtv.org/trac/ticket/9792

[14:25:06]

jmartens (jmartens!~jmartens@109.232.42.33) has quit (Quit: Leaving.)

[14:30:24]

stuartm:

huh, someone tried to login as me on Freenode last week – 173.81.156.72

[14:30:43]

stuartm:

"-NickServ- Last failed attempt from: mythtv!~mythtv@173.81.156.72 on Jul 14 18:06:23 2011"

[14:31:54]

stuartm:

any reason I shouldn't just ban them for trying to hi-jack a channel operator's account?

[14:34:16]

danielk22:

It looks like mysql deadlocked in 9704.

[14:34:42]

danielk22:

#9885 should be fixed as of this week in master.

[14:35:59]

danielk22:

It looks like 9792 & 9773 may also be due to mysql deadlocking.

[14:40:32]

danielk22:

hmm, kormoc asked for a "SHOW FULL PROCESSLIST" on one of them and it looks like mysqld doesn't know of anything going on but we're deep in the mysql lib with a prepare statement.

[14:40:43]

danielk22:

Ticket #9773: BEtrace2_and_MySqlProcList.txt

[14:41:24]

kth1 (kth1!~kth@2001:638:708:10:222:5fff:fee7:6f0f) has quit (Quit: Leaving.)

[14:42:45]

jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has joined #mythtv

[14:44:18]

jwhite_ is now known as jwhite

[14:47:27]

stuartm:

I guess we need to know whether there are any common factors, same version of mysql or whether they are all using the same distro

[14:55:06]

paul-h (paul-h!~paulh@mythtv/developer/paul-h) has quit (Remote host closed the connection)

[14:55:50]

paul-h (paul-h!~paulh@5adce259.bb.sky.com) has joined #mythtv

[14:55:50]

paul-h (paul-h!~paulh@5adce259.bb.sky.com) has quit (Changing host)

[14:55:50]

paul-h (paul-h!~paulh@mythtv/developer/paul-h) has joined #mythtv

[14:55:58]

sphery:

mrand: if you guys use the MythTV support for syslog, you can put all the logging in a single file by any name you like... And, if you have a decent modern syslog implementation, you can break things out by program name and other criteria, too--to put the logging for any program or child program in any file you like with any name you want. :)

[14:56:09]

stuartm is now known as MythTV

[14:56:55]

MythTV (MythTV!~stuartm@mythtv/developer/stuartm) has left #mythtv ()

[14:57:02]

stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has joined #mythtv

[15:06:46]

iamlindoro:

paul-h: Nice cleanup on the menu themes, I didn't realize that so much had been missed, and two of those three are my fault

[15:09:40]

j-rod is now known as j-rod|afk

[15:09:43]

abqjp (abqjp!~abqjp@71-38-208-26.albq.qwest.net) has joined #mythtv

[15:18:02]

jarle (jarle!~jarle@70.84-234-133.customer.lyse.net) has quit (Remote host closed the connection)

[15:20:07]

jarle (jarle!~jarle@70.84-234-133.customer.lyse.net) has joined #mythtv

[15:20:23]

j-rod|afk is now known as j-rod

[15:21:07]

srk9 (srk9!~richard@unaffiliated/srk9) has joined #mythtv

[16:07:27]

stuartm:

well so much for that theory

[16:10:43]

paul-h:

iamlindoro: pressing escape or using the Previous button in the "Setup Wizard" doesn't seem to be working properly. I just get a blank screen, pressing escape again does returns to the menu though

[16:12:10]

paul-h:

also every time I try to test the video it fails to download the videos

[16:19:22]

iamlindoro:

paul-h: That's a MythUI/Theme issue that I can't replicate in Arclight, but I've seen it in MythCenter

[16:20:42]

iamlindoro:

Regarding the videos, it's possible something has changed on the server side, worked last time I tested it

[16:23:02]

iamlindoro:

I suppose it's also possible that it's related to MythDownloadManager issues-- all the screen does is tell MythDownloadManager to download the file to the MBE's ~/.mythtv/tmp folder, backend and frontend logs with -v file might be enlightening

[16:23:49]

iamlindoro:

paul-h: http://services.mythtv.org/samples/video/?sample=HD Server broken

[16:24:01]

iamlindoro:

Captain_Murdoch: ^^^ Video sample service we set up seems no longer to work

[16:24:15]

iamlindoro:

Looks like the samples are missing from OSUOSL

[16:29:26]

iamlindoro:

paul-h: Can you please confirm that the issue does not exist when you try Arclight?

[16:34:09]

paul-h:

yeah works OK in Arclight

[16:34:25]

paul-h:

So what's different?

[16:34:35]

iamlindoro:

obviously something, but I have no idea what

[16:34:48]

iamlindoro:

ergo my feeling that it's a fundamental MythUI issue

[16:35:57]

iamlindoro:

Hmm... wonder if I might be using the wrong stack?

[16:36:41]

iamlindoro:

could you try changing the setup screens to use the main stack instead? That might be it

[16:38:36]

Beirdo:

mrand: I think you are misunderstanding the contents of the logfiles.

[16:38:54]

Beirdo:

the pid is the pid of the main thread, and all threads log in the same file

[16:39:29]

mrand:

Beirdo: quite possible: it's hard to follow fragmented discussions on channel (insert ranting about mailing lists being the proper place for technical discussions here ;-)

[16:40:07]

Beirdo:

The file boundary is per run of mythbackend (or whichever program), and determining the newest file is dead simple because you can sort on the date

[16:40:55]

Beirdo:

I still plan to make a cleanup script to run from cron to remove logs older than say 4 weeks

[16:41:13]

Beirdo:

or archive them

[16:42:06]

mrand:

The main point though was that having unpredictable filenames impacts automated bug submission (apport hook)

[16:43:05]

mrand:

(bug = crash)

[16:43:49]

Beirdo:

actually, come to think of it, logrotate can be used to clean up the dir

[16:44:07]

Beirdo:

well, we don't support apport.

[16:44:30]

mrand:

The Mythbuntu group has it working just fine.

[16:44:37]

Beirdo:

and it shouldn't be hard, you know the pid and the name of the crashed program, no?

[16:45:02]

sphery:

mrand: which syslog implementation does ubuntu use? I use rsyslog, and if that's what *buntu uses, I'll do up a config today that should break it out nicely and allow you to control filenames, etc.

[16:45:59]

mrand:

We're all short handed here, so I'm just flagging that apport hooks currently work, and I'm not aware that they can use the .pid in a hook to autosubmit a crash. But I could just as easily be wrong. Maybe superm1 knows.

[16:49:40]

kormoc_afk is now known as kormoc

[16:50:27]

bill6502 (bill6502!~bill6502@2002:d850:4778:7cd3:9ef4:32ea:c6ea:200) has joined #mythtv

[16:51:50]

tgm4883:

sphery, mrand IIRC ubuntu uses rsyslog

[16:52:59]

sphery:

great... I'll do up a starter config and post it to the wiki for review/additions/...

[16:55:51]

Beirdo:

sweet

[17:04:51]

davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)

[17:05:20]

davide_ (davide_!~david@host103.16.intrusion.com) has joined #mythtv

[17:05:20]

davide_ (davide_!~david@host103.16.intrusion.com) has quit (Changing host)

[17:05:20]

davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv

[17:06:45]

andreax (andreax!~andreaz@p57B9407C.dip.t-dialin.net) has joined #mythtv

[17:12:28]

paul-h (paul-h!~paulh@mythtv/developer/paul-h) has quit (Remote host closed the connection)

[17:20:07]

paul-h (paul-h!~paulh@mythtv/developer/paul-h) has joined #mythtv

[17:24:16]

Jordack is now known as AAAAA-Jordack

[17:26:32]

superm1:

yeah rsyslog

[17:31:25]

Captain_Murdoch:

iamlindoro, I don't know what happened to those files, it may have been before we cutover the rsync mirror source and they weren't on the new server possibly. I just put them in the rsync mirror source dir, so they should show up again after the rsync tonight.

[17:31:45]

knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Ping timeout: 276 seconds)

[17:32:27]

knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv

[17:37:43]

iamlindoro:

Captain_Murdoch: Cool, thanks!

[17:38:59]

AAAAA-Jordack is now known as Jordack

[17:43:34]

stuartm:

oh sweet

[17:44:57]

stuartm:

iplayer directly through the red button, utterly pointless but pretty cool all the same

[17:47:01]

wagnerrp:

do you know if they broadcast any sort of time limit on those things?

[17:47:05]

stuartm:

what would be really great is a 'press Red to record this' feature during programme trailers

[17:47:33]

wagnerrp:

would be amusing to click on the button on a 3yr old recording, only to get a 404

[17:48:14]

stuartm:

yeah, which is why the mheg engine currently only runs for livetv

[18:08:50]

paul-h:

stuartm: I'd rather have our player playing the iPlayer videos rather than using the flash player any day

[18:10:03]

paul-h:

I did see some green button stuff in the patch not sure if it's actually hooked up to anything

[18:10:08]

stuartm:

paul-h: indeed, is that what this patch allows because I had thought it just started up the browser

[18:10:17]

stuartm:

?

[18:10:49]

paul-h:

We would use our player not a web browser

[18:11:18]

stuartm:

oh, well that's even better then

[18:11:41]

stuartm:

I didn't think the BBC would be sending unencrypted video

[18:11:49]

paul-h:

the green button stuff is in reference to being able to press red to record a trailer

[18:12:48]

stuartm:

paul-h: yeah, I figured that's what you were talking about

[18:13:54]

kormoc is now known as kormoc_afk

[18:14:24]

paul-h:

FreeSat players don't use a web browser and we are only using the same interface as they use

[18:14:32]

kormoc_afk is now known as kormoc

[18:16:28]

paul-h:

iamlindoro: I don't see where the wizard is using the popup stack except for menus and busy dialogs?

[18:17:35]

stuartm:

paul-h: I've never seen a FreeSat STB up close

[18:19:32]

paul-h:

iamlindoro: the difference between Arclight and the default versions is you are using a non fullscreen window but the default is full screen

[18:19:52]

paul-h:

if I make the default non full screen it works fine?

[18:21:53]

stuartm:

gigem: can you sign off on the following patch since it's touching the scheduler? http://code.mythtv.org/trac/attachment/ticket . . . r.cpp.2.diff

[18:22:25]

paul-h:

iamlindoro: It also works ok if I remove all the Show() and Hide() calls except you can see the previous screens in Arclight because the background is transparent

[18:23:06]

stuartm:

it sounds like it's in the wrong stack

[18:23:32]

paul-h:

if it is I can't see where

[18:29:07]

paul-h:

Maybe it would be better to use AddScreen() and PopScreen() rather that Show() and Hide()?

[18:30:24]

kth (kth!~kth@unaffiliated/kth) has joined #mythtv

[18:52:24]

abqjp (abqjp!~abqjp@71-38-208-26.albq.qwest.net) has quit (Remote host closed the connection)

[19:01:45]

taylorr (taylorr!~taylorr@cpe-173-095-144-220.nc.res.rr.com) has joined #mythtv

[19:01:45]

taylorr (taylorr!~taylorr@unaffiliated/elmojo) has joined #mythtv

[19:01:45]

taylorr (taylorr!~taylorr@cpe-173-095-144-220.nc.res.rr.com) has quit (Changing host)

[19:04:05]

Mousey (Mousey!~wtfisme@ross154.net) has joined #mythtv

[19:22:19]

taylorr:

fyi, I unassigned my tickets because I'm not very active right now and didn't want anyone to think I was working on them

[19:28:04]

sphery:

ah, good--we don't want to lose you :)

[19:33:59]

taylorr:

one of those tickets is quite old and I'm still not sure we should allow seek tables for AVI or really any container that doesn't support timestamp discontinuities

[19:34:27]

taylorr:

I have a hunch what might fix it but haven't been motivated since anything AVI makes me want to run and hide

[19:43:14]

simonckenyon (simonckenyon!~simoncken@195.7.61.12) has quit (Read error: No route to host)

[19:57:21]

stuartm:

paul-h: have you see the email to the -dev list from warpme?

[20:01:40]

abqjp (abqjp!~abqjp@71-38-208-26.albq.qwest.net) has joined #mythtv

[20:22:14]

kormoc is now known as kormoc_afk

[20:22:44]

morxe (morxe!~martin@c-2a7370d5.024-74-736b7610.cust.bredbandsbolaget.se) has joined #mythtv

[20:25:21]

kormoc_afk is now known as kormoc

[21:08:02]

sphery:

stuartm: Are you still around? Just wondering if it's OK to assume that in https://github.com/MythTV/mythtv/blob/master/ . . . enu.cpp#L219 the only callbacks we'll have for the exit situation is the one in TVMenuCallback() for exiting_app that calls handleExit() ( https://github.com/MythTV/mythtv/blob/master/ . . . in.cpp#L1018 ). In other words, since handleExit() just does a qApp->quit(); ...

[21:08:08]

sphery:

... if prompts are off, can I just do a QCoreApplication::exit() in myththemedmenu for the no-prompt key binding action?

[21:17:03]

jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has quit (Quit: Leaving.)

[21:20:41]

stuartm:

if I understand your question, then no, there is a different callback for mythtv-setup which performs some setting sanity checks before allowing the user to exit, it's not safe to assume that we can just exit in myththemedmenu

[21:21:15]

sphery:

yeah, this would be only for mythfrontend

[21:22:55]

sphery:

the other option is to add another string for the callback... haven't decided which I like better

[21:22:59]

stuartm:

although not pretty myththemedmenu uses the callbacks because it may be used in any application which requires a different exit behaviour

[21:23:41]

sphery:

I think the new string is better, anyway--lets the app decide how to handle it

[21:23:46]

stuartm:

I don't know of any third party apps using mythui, but they could theoretically exist, so it's relevant even if mythtv-setup goes away

[21:24:04]

stuartm:

sphery: I think you're right

[21:24:18]

sphery:

Thanks. I'll go that way, then.

[21:26:39]

pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has joined #mythtv

[21:26:44]

Jordack (Jordack!~jordack@h69-131-44-221.plmomi.dedicated.static.tds.net) has quit (Quit: Todays Why Did I Watch This Moment: http://www.youtube.com/watch?v=7y-x2fWKbmo)

[21:28:17]

sphery:

stuartm: so, that way, I need to know what m_killable is used for. Is that only for starting up a plugin directly (like mythfrontend mythvideo )?

[21:30:46]

stuartm:

heh, I knew you'd ask that because I really don't know

[21:31:46]

stuartm:

I figured it out as much as I needed to know when re-factoring myththemedmenu but ...

[21:32:32]

stuartm:

I think it's a reasonable conclusion that it's there for something like that, it's certainly only used by plugins (about half of them)

[21:33:39]

sphery:

ok, I'll try to track down a way to test it--thanks for confirming I'm not way off base on it

[21:34:02]

stuartm:

I'm not sure it's really needed and I may just remove it

[21:34:29]

sphery:

heh, that works, too

[21:34:52]

sphery:

heh, and btw, mythfrontend mythvideo doesn't work anymore--there is no mythvideo, anymore :)

[21:37:01]

morxe (morxe!~martin@c-2a7370d5.024-74-736b7610.cust.bredbandsbolaget.se) has quit (Remote host closed the connection)

[21:39:45]

superm1 (superm1!~superm1@ubuntu/member/superm1) has quit (Quit: Coyote finally caught me)

[21:39:45]

rhpot1991 (rhpot1991!~rhpot1991@ubuntu/member/rhpot1991) has quit (Quit: Coyote finally caught me)

[21:39:45]

tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has quit (Quit: Coyote finally caught me)

[21:40:21]

j-rod is now known as j-rod|afk

[21:42:38]

Gibby (Gibby!~Gibby@204.118.10.244) has quit (Ping timeout: 252 seconds)

[21:43:49]

wagnerrp:

taylorr: i wouldnt feel at all bad about closing AVI tickets, and telling people to remux to a worthwhile container

[21:44:52]

wagnerrp:

in fact...

[21:45:07]

wagnerrp:

Beirdo: will mythffmpeg remux to mp4 or mkv?

[21:46:46]

taylorr:

wagnerrp: yes, I've converted them to mkv using avidemux2 and they seek perfectly then

[21:47:40]

wagnerrp:

i mean, if we ship with something that could do it locally, and not require avidemux2, i say we close them

[21:47:55]

wagnerrp:

i.e. 'mythffmpeg -i somefile.avi -o somefile.mp4'

[21:48:18]

wagnerrp:

mkv could be a bit iffy if hes using external hardware players

[21:48:29]

wagnerrp:

but i doubt there is anything that supports avi and not mp4

[21:49:01]

Beirdo:

I think that it will, you can check with the --formats or --codecs (I think) parameters

[21:49:27]

wagnerrp:

especially considering avi only technically supports one audio and one video stream

[21:49:44]

wagnerrp:

so there is no issue with making sure theyre all transferred

[21:50:19]

Gibby (Gibby!~Gibby@204.118.10.244) has joined #mythtv

[21:52:10]

taylorr:

wagnerrp: .mp4 isn't much different than .avi – they both are horrible

[21:53:12]

taylorr:

at any rate, I don't think we should force users to not use .avi – I'm just not going to put a lot of effort into an old container format that seems to be primarly used by pirates

[21:53:48]

wagnerrp:

no, im just saying IF users come up with strange errors, and we have an easy method of converting to something more modern and better supported

[21:53:58]

rhpot1991 (rhpot1991!~rhpot1991@204.8.45.13) has joined #mythtv

[21:53:58]

rhpot1991 (rhpot1991!~rhpot1991@204.8.45.13) has quit (Changing host)

[21:53:58]

rhpot1991 (rhpot1991!~rhpot1991@ubuntu/member/rhpot1991) has joined #mythtv

[21:54:00]

stuartm:

there might be an issue of portability, I'm guessing that at least some have chosen avi because that's the container their portable device supports

[21:54:03]

sphery:

forcing mythcommflag --rebuild and mythtranscode --buildindex to exit without building a seektable for avi makes sense, though

[21:54:03]

taylorr:

right

[21:54:04]

wagnerrp:

that should be the solution, rather than updating the player

[21:54:44]

wagnerrp:

wait, so playback was only broken because the user had built a seektable for the video?

[21:55:02]

taylorr:

sphery: I'm going to see if I can make seektables work with .avi one last try... if not they we can remove support for seektable generation

[21:55:23]

sphery:

wagnerrp: was basing that conclusion on: 07.21 15:33:59 < taylorr> one of those tickets is quite old and I'm still not sure we should allow seek tables for AVI or really any container that doesn't support timestamp discontinuities

[21:55:25]

taylorr:

wagnerrp: yes, the user built a seektable and when seeking the video was corrupted

[21:56:21]

sphery:

I think we have it so that they won't create seektables for mkv, right now, right? something about not supporting the bitstream format or something like that that was way over my head

[22:06:21]

kormoc is now known as kormoc_afk

[22:24:21]

kormoc_afk is now known as kormoc

[22:35:49]

andreax (andreax!~andreaz@p57B9407C.dip.t-dialin.net) has quit (Read error: Connection reset by peer)

[22:49:35]

davide_:

stuartm: i just saw your ping regarding #3597. i've been very busy lately and almost blew off readying the irc backlog today. i'll take a look, but it will likely be tomorrow before i can get to it.

[22:49:56]

davide_:

s/readying/reading/

[22:53:25]

bill6502 (bill6502!~bill6502@2002:d850:4778:7cd3:9ef4:32ea:c6ea:200) has left #mythtv ()

[22:53:45]

stuartm:

davide_: no problem, there's no real hurry, I just wanted to close a four year old ticket

[22:54:28]

tgm4883 (tgm4883!~tgm4883@204.8.45.13) has joined #mythtv

[22:54:29]

tgm4883 (tgm4883!~tgm4883@204.8.45.13) has quit (Changing host)

[22:54:29]

tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has joined #mythtv

[22:56:25]

stuartm:

there are only four or five tickets older than that still in the system, this one had the simplest patch ;)

[22:58:00]

superm1 (superm1!~superm1@ubuntu/member/superm1) has joined #mythtv

[23:10:57]

danielk22:

stuartm: how the heck did #3001 survive so long??

[23:14:20]

stuartm:

you're asking me? :)

[23:15:36]

abqjp (abqjp!~abqjp@71-38-208-26.albq.qwest.net) has quit (Quit: abqjp)

[23:15:41]

stuartm:

I think there was some desire to make the effort because bjm was the reporter, but it's clearly going nowhere and bjm hasn't been around for years now

[23:20:54]

zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Ping timeout: 255 seconds)

[23:27:43]

zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv

[23:30:21]

stuartm:

danielk22: since I'm looking at the old tickets, this one has been open for 5 years and is assigned to you – http://code.mythtv.org/trac/ticket/2650

[23:35:56]

stuartm:

danielk22: so there seems to be no obvious correlation between those DB related deadlocks and mysql version

[23:36:58]

stuartm:

danielk22: but warpme is reporting that switching to innodb fixed the issue for him, if that does turn out to work for the others then we might as well go ahead and switch to innodb

[23:37:31]

stuartm:

we've been talking about doing it for months and there are no downsides, only benefits

[23:38:16]

stuartm:

innodb is much more resilient to corruption, more stable and has many more useful features

[23:38:46]

stuartm:

I believe kormoc has been running innodb for months now

[23:40:12]

stuartm:

sphery: is there an established procedure, or set of instructions for switching which we can give to the users who are suffering these deadlocks?

[23:40:34]

kormoc:

stuartm, aye, I have been

[23:40:41]

sphery:

stuartm: none that I know of... kormoc may have something

[23:41:04]

kormoc:

Not really. You just alter each table one at a time

[23:41:06]

wagnerrp:

stuartm: the biggest blocker to switching to innodb is that many packagers ship without it enabled

[23:41:11]

sphery:

if we do switch, we'll need to alert distros--since many of them were disabling innodb engine at one point

[23:41:33]

kormoc:

plus a non-configured innodb can be way slower then myisam :(

[23:41:34]

sphery:

I'm guessing that--at least those who have switched to MySQL 5.5--have re-enabled it since it's default in 5.5

[23:41:36]

wagnerrp:

availability of the innodb engine is something we track in the updated smolt jams is working on

[23:41:43]

sphery:

(though they may well disable MyISAM?))

[23:41:47]

kormoc:

nah

[23:41:53]

kormoc:

myisam is still required for the system tables

[23:42:00]

sphery:

ah, didn't realize that

[23:42:09]

wagnerrp:

stuartm: it might be worth waiting one more version to see what those numbers are

[23:42:15]

wagnerrp:

to see how difficult it will be to switch over

[23:42:52]

pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has quit (Quit: Leaving.)

[23:43:40]

stuartm:

wagnerrp: if it fixes these deadlocks then I don't see that we have anything to use, I'd sooner that we tell some users they'll have to wait for their distro before upgrading than leave others with unusable backends

[23:43:47]

stuartm:

s/use/lose/

[23:45:22]

sphery:

I would guess that it didn't actually fix the issue--as it would imply that MyISAM is broken on some versions of MySQL and a) I'd expect some MySQL dev to notice that and b) I'm using a version of MySQL in between versions reported by these users and haven't had any issues with my MyISAM engine

[23:46:26]

sphery:

I know warpme/warped has been modifying a bunch of MySQL settings to try to narrow it down/fix it, and it's possible something else he did fixed it (he was modifying some query cache settings and stuff--and he had run mysqltuner.pl on his system before getting the deadlocks and changed things it said he might want to consider changing)

[23:47:02]

sphery:

I'd guess it's more likely some Qt-MySQL/MySQL C API interaction issue--but that's pure speculation

[23:47:27]

sphery:

that said, I'm ambivalent about whether we switch or not--as long as we notify packagers, it should be fine

[23:48:42]

stuartm:

I suppose we need to get hard figures on the speed difference of a stock MyISAM configuration vs stock InnoDB

[23:49:29]

kormoc:

query cache can cause issues

[23:49:30]

stuartm:

it's only going to make a difference to the scheduler, nothing else we do comes close to being hard on the database

[23:49:34]

kormoc:

it's a single mutex

[23:50:23]

stuartm:

anyway, maybe we're getting ahead of ourselves, let's find out if it fixes the deadlock and if not there isn't any urgency

[23:51:36]

sphery:

his discussion of query cache was on http://code.mythtv.org/trac/ticket/9792 and http://www.gossamer-threads.com/lists/mythtv/dev/481868#481868

[23:52:10]

stuartm:

I tend to agree with sphery that it would be unusual for such a bug to go unnoticed for so long in mysql, but we don't have a better explanation right now

[23:53:11]

sphery:

yeah--but then again, he said he switched to InnoDB 2 weeks ago and it's working... but in #9792, he said he turned off query cache 2 weeks ago and it was working (and both seem to be the same mysql deadlock)

[23:53:28]

sphery:

so I have a feeling he'll find later that it's again not working even with innodb

[23:53:43]

sphery:

seems it's a slow-to-reproduce issue

[23:54:27]

kormoc:

a SHOW ENGINE INNODB STATUS\G when it's deadlocked would help

[23:56:11]

stuartm:

I'd actually like to see one or all of the reporters produce two or more stacktraces from the same instance of the wedged backend just to be certain that it's appears stuck in the same place within libmysql

[23:57:18]

stuartm:

it's entirely possible that the database is a red-herring and that it's not that thread which has deadlocked

[23:57:27]

sphery:

kormoc: his ticket is http://code.mythtv.org/trac/ticket/9773 and #9704 and #9792 have other possibly related/same issues if you'd like to post about the engine status stuff (I'm assuming there's an equivalent for MYISAM status?)

[23:57:47]

Mousey (Mousey!~wtfisme@ross154.net) has quit (Ping timeout: 258 seconds)

[23:57:55]

kormoc:

sphery, there is not :(

[23:58:08]

sphery:

stuartm: yeah, warped says he has a backtrace of a deadlocked scheduler that's not a mysql thing--we should probably ask for that

MythLogBot@irc.freenode.net :: #mythtv

Current users (82):