| --- | Log | opened Fri Mar 07 00:00:14 2008 |
| --- | Day | changed Fri Mar 07 2008 |
| 00:00 | <xris> | I think a lot of students don't count it as a full time job. |
| 00:00 | <xris> | and take on other things at the same time. |
| 00:00 | <dagar> | I was thinking of trying it in place of a full time job |
| 00:00 | <dagar> | just barely enough money to justify not getting a job |
| 00:01 | <dagar> | besides, more money means more distractions for me |
| 00:01 | <dagar> | my myth boxes tend to multiply when I have cash... |
| 00:03 | <xris> | $4500 for two months not a full time job? that's right around $14/hour |
| 00:03 | <dagar> | I thought it was 3 or 4 months? |
| 00:04 | <xris> | I haven't checked the calendar for this year, but I think it was about 10 weeks in 2006 (which I guess puts it more near $12/hour) |
| 00:04 | <dagar> | not too shabby |
| 00:11 | -!- | ahbritto [n=guest@adsl-68-125-197-181.dsl.pltn13.pacbell.net] has quit [Client Quit] |
| 00:11 | -!- | reynaldo1 [n=rverdejo@190-82-33-234.adsl.cust.tie.cl] has quit ["Lost terminal"] |
| 00:11 | -!- | kormoc [n=kormoc@unaffiliated/kormoc] has quit [] |
| 00:16 | -!- | cattelan [n=cattelan@c-75-72-22-101.hsd1.mn.comcast.net] has quit ["This computer has gone to sleep"] |
| 00:17 | <Chutt> | xris, it's actually pretty low compared to real internships |
| 00:18 | <xris> | Chutt: depends on the industry, I guess. |
| 00:18 | <Chutt> | programming? =) |
| 00:18 | <xris> | I only paid my intern $12.50/hour |
| 00:18 | <xris> | and he was a college grad |
| 00:19 | <xris> | granted, I bumped him up to a decent salary after 3 months |
| 00:19 | <xris> | (although he's still not making as much as he's worth) |
| 00:22 | <xris> | I honestly don't know what pay ranges are valid. my wife has a degree from a pastry school and could only hope to *maybe* make $15/hour after working in the industry for a couple of years. |
| 00:22 | <Chutt> | MS pays a little of 60k/yr for interns (the equivalent, rather) |
| 00:22 | <Chutt> | iirc |
| 00:22 | <dagar> | where do you live xris? |
| 00:23 | <xris> | seattle |
| 00:23 | <xris> | Chutt: the general rule around here is that MS isn't "the industry".. smaller companies pay a lot less. |
| 00:23 | <Chutt> | :p |
| 00:23 | <dagar> | coop jobs at my university start around $15/hour |
| 00:23 | <xris> | heck, according to places like salary.com I'm bout $35k/year underpaid |
| 00:24 | <xris> | dagar: see, I don't know anything. I made like $7/hour in college (late 90s) working for the IT dept |
| 00:24 | <dagar> | is that based on other user input? |
| 00:24 | <xris> | salary.com supposedly is. I don't trust their numbers or their job descriptions, though. |
| 00:24 | <xris> | but have used them to get raises before |
| 00:24 | <Chutt> | xris, i've got no idea, either, just going by what my brother was paid as an intern |
| 00:25 | <Chutt> | the other places he was at weren't much less |
| 00:26 | <xris> | then google's stuff is low... |
| 00:26 | <xris> | I'm also going off of numbers from 2 years ago. |
| 00:26 | <Chutt> | well, for good stuff in the US |
| 00:26 | <Chutt> | it's lots o cash for international students |
| 00:26 | <xris> | inflation and cost of living in seattle alone is like 4.5%/year |
| 00:26 | <xris> | or rural US |
| 00:28 | <xris> | Chutt: there's also a lot to be said for working on something you enjoy, and getting good experience. |
| 00:28 | <xris> | but unfortunately, I don't think very many companies count FOSS work as good experience (it's sad) |
| 00:29 | <Chutt> | mine did |
| 00:29 | <Chutt> | hehe |
| 00:30 | <xris> | I do when I hire |
| 00:30 | <kdub> | salary.com shows pretty much exactly what mine is |
| 00:31 | <Chutt> | salary.com says $0 for me |
| 00:31 | <Chutt> | hrm |
| 00:32 | <dagar> | what did you enter? |
| 00:32 | <xris> | I never know what to put in for me. |
| 00:32 | <Chutt> | just software engineer 4 |
| 00:32 | <xris> | sr web developer, or just programmer 3, 4, or 5. |
| 00:33 | <Chutt> | ah, works in firefox |
| 00:37 | <Chutt> | heh, yah, that's way off |
| 00:37 | <xris> | too high or too low? |
| 00:37 | <Chutt> | low |
| 00:38 | <xris> | wow. mine's way higher than I make |
| 00:38 | <Chutt> | but then, it's probably because of the telecommuting |
| 00:38 | <xris> | yeah |
| 00:39 | <xris> | back in a bit |
| 00:39 | -!- | xris [n=xris@xris.forevermore.net] has quit [] |
| 00:39 | <clever> | 2008-03-07 01:32:18.635 MSqlQuery: DELETE FROM settings WHERE value = 'MusicAutoShowPlayer' AND hostname = 'theP4' ; |
| 00:39 | <clever> | 2008-03-07 01:32:18.637 MSqlQuery: INSERT INTO settings (value,data,hostname) VALUES ( 'MusicAutoShowPlayer', '1', 'theP4' ); |
| 00:39 | <clever> | where in the code is that delete/insert pair generated? |
| 00:41 | <clever> | MythContext::SaveSettingOnHost it seems |
| 00:46 | -!- | xris [n=xris@xris.forevermore.net] has joined #mythtv |
| 00:52 | -!- | davilla [n=davilla@nc-65-41-43-142.sta.embarqhsd.net] has quit ["Leaving"] |
| 00:55 | -!- | MrGandalv [i=mgandalf@cpe-72-225-32-214.rochester.res.rr.com] has joined #mythtv |
| 00:56 | -!- | _packetscan [i=Tsu@srv01.packetscan.com] has joined #mythtv |
| 00:58 | -!- | Netsplit anthony.freenode.net <-> irc.freenode.net quits: sphing, kurre2__, loops, MrGandalf, sphery, packetscan, mzb_d800, dlblog |
| 00:58 | -!- | Netsplit over, joins: loops, sphery, sphing, dlblog, kurre2__, mzb_d800 |
| 00:59 | <clever> | patch done, testing |
| 00:59 | -!- | cattelan [n=cattelan@c-75-72-22-101.hsd1.mn.comcast.net] has joined #mythtv |
| 01:01 | -!- | cattelan [n=cattelan@c-75-72-22-101.hsd1.mn.comcast.net] has quit [Client Quit] |
| 01:07 | <clever> | seems to work perfectly but half the if never ran |
| 01:08 | -!- | gnome42 [n=gnome42@76-10-151-103.dsl.teksavvy.com] has quit [Remote closed the connection] |
| 01:09 | <clever> | isnt detecting the failed updates... |
| 01:10 | -!- | Tanthrix [n=tanthrix@c-71-56-144-82.hsd1.or.comcast.net] has quit [Read error: 104 (Connection reset by peer)] |
| 01:10 | -!- | Tanthrix [n=tanthrix@c-71-56-144-82.hsd1.or.comcast.net] has joined #mythtv |
| 01:11 | <clever> | ahhh |
| 01:11 | <clever> | 0 rows matched isnt an error |
| 01:14 | -!- | _packetscan is now known as packetscan |
| 01:29 | <clever> | if i do an update that matches 1 row and updates 0 what will QSqlQuery::numRowsAffected return? |
| 01:29 | <clever> | (row allready set to value im setting) |
| 01:29 | <clever> | Rows matched: 1 Changed: 0 Warnings: 0 |
| 01:29 | <clever> | Rows matched: 0 Changed: 0 Warnings: 0 |
| 01:30 | <clever> | i need to be able to tell the diff between those from qt based code |
| 01:38 | <Deek> | should be zero, because no rows were changed |
| 01:51 | <clever> | yeah and thats the problem |
| 01:51 | <clever> | the docs dont seem to show a function for rows matched |
| 01:51 | <clever> | so if i try to update a setting which doesnt exist nothing goes wrong |
| 01:51 | <clever> | and it doesnt get saved |
| 01:52 | -!- | Tanthrix [n=tanthrix@c-71-56-144-82.hsd1.or.comcast.net] has quit [Read error: 110 (Connection timed out)] |
| 01:52 | <clever> | im changing SaveSettingOnHost to update then insert if it needs to |
| 01:53 | <clever> | the delete,insert seems wrong and feels like the cause of some past bugs ive had |
| 01:56 | <Deek> | how about select then insert/update? |
| 01:59 | <sphery> | why would delete, insert cause bugs (and how is it that update or insert wouldn't) |
| 02:00 | <Deek> | clever: no version of mysql_num_rows() ? |
| 02:00 | <clever> | in qt it seems to be .size() |
| 02:00 | <clever> | and it only returns the size for a select type query |
| 02:00 | <clever> | updates will have a .size() of -1 |
| 02:01 | <clever> | sphery: sometimes the fe/mysqld hang right between the delete and insert |
| 02:01 | <clever> | and i kill one of them |
| 02:01 | <clever> | then when it comes back up the setting is missing |
| 02:01 | <clever> | so far the only victim has been the language, causing the fe to ask on startup and give me a heartattack thinking everything is gone |
| 02:03 | <clever> | an update query would do atomicly within the server preventing that from happening |
| 02:03 | <sphery> | sounds like you need to run an optimize_mythdb.pl to update indices, ... |
| 02:03 | <clever> | most often its when the db server runs out of space |
| 02:03 | <clever> | and i restart the fe before discovering that |
| 02:04 | <Deek> | yeah, I'd almost never do delete/insert in a sitch like that; I'd select/(update|insert) |
| 02:04 | <clever> | my patch is currently trying to fetch the setting to see if it even needs a change |
| 02:04 | <clever> | then trying to update and insert if the update fails |
| 02:04 | <clever> | but it cant detect the update matching 0 rows(because it doesnt exist to begin with) |
| 02:05 | <clever> | so if the setting isnt in mysql it never gets saved |
| 02:06 | <clever> | 07 03:04:43 < tausq> clever: do you mean something like this? http://dev.mysql.com/doc/refman/5.0/en/insert-on-duplicate.html |
| 02:06 | <clever> | that query would fix my bug perfectly if there was a uniq key on value,hostname |
| 02:07 | <Deek> | so the Qt code is screwed (apparently written for mSQL) |
| 02:08 | <clever> | its just missing a function to access a mysql only feature |
| 02:08 | <clever> | and the qt guys gave a way to extract the original mysql object from the qt one and use real mysql lib api on it |
| 02:08 | <clever> | which would require including mysql headers |
| 02:09 | <clever> | which feels like more of a major change then my idea |
| 02:11 | -!- | mzuverink [n=mzuverin@unaffiliated/mzuverink] has quit ["Leaving"] |
| 02:12 | <stuarta> | gbee: i'd expect dvb-s initially |
| 02:13 | <stuarta> | (re: freesat) |
| 02:14 | -!- | PointyPumper [i=Pintlezz@190.19.128.73] has quit [Read error: 104 (Connection reset by peer)] |
| 02:17 | <clever> | testing the new code... |
| 02:22 | <stuarta> | the mysql sql code was written to work with even mysql 3.3 |
| 02:23 | <clever> | could wrap the whole thing in an if to use the old buggy method with that version |
| 02:24 | <stuarta> | there's a long term goal to move to embedded mysql which must be kept in mind |
| 02:24 | <clever> | the embeded mysql server probly wont be 3.3:P |
| 02:25 | <clever> | ive skimed over the mysql api before when using it in some of my c++ code |
| 02:25 | <stuarta> | no but embedded things generally aren't as featureful as the big brother |
| 02:25 | <clever> | other then the way you init&connect theres little change when you switch to an embeded i beleive |
| 02:25 | <clever> | ahh |
| 02:25 | <clever> | ALTER TABLE `settings` DROP INDEX `value` , |
| 02:25 | <clever> | ADD PRIMARY KEY ( `value` , `hostname` ) |
| 02:26 | <clever> | now the fe wont even start! |
| 02:27 | <hads> | Isn't 5+ required now anyway? |
| 02:27 | <xris> | should be for .21, yes |
| 02:27 | <clever> | undoing hasnt fixed it... |
| 02:27 | <clever> | im making this change in trunk |
| 02:28 | <clever> | which will become .21 if im correct |
| 02:28 | <hads> | Well, 21-fixes will become .21 :) |
| 02:28 | <stuarta> | i knew we upped the requirement, but i couldn't remember the new version |
| 02:28 | <clever> | ahh |
| 02:28 | <clever> | i need to track down how i killed the frontend now... |
| 02:28 | <clever> | 2008-03-07 03:27:43.062 This version of MythTV requires an updated database schema. Please run mythtv-setup or mythbackend to update your database. |
| 02:28 | <stuarta> | but still develop against trunk, as we are in feature freeze. |
| 02:29 | <sphery> | gbee: The frontend was able to upgrade my schema without issue using command line: mythfrontend -O ThemePainter=qt -geometry 800x600 -u (SSH X redisplay). Even tested once with the backup failing and once with it succeeding. Both times, the upgrade happened without issue. |
| 02:30 | <sphery> | gbee: I even tried moving the -u around in the command line and couldn't get it to fail. Your 1213 upgrade was great because I could keep resetting DBSchemaVer to 1212 and the upgrade would be able to succeed without other changes. Therefore, if you could try it again and let me know what you find... Thanks. |
| 02:31 | <clever> | i see the problem |
| 02:31 | <sphery> | clever: Why all the concern over losing a single value when your MySQL server ran out space? That's a pretty critical error that--had it not occured when changing the setting--would have happened soon enough (and perhaps at a much more critical location). |
| 02:32 | <clever> | DBSchemaVer is owned by the '' hostname |
| 02:32 | <sphery> | MythTV cannot be expected to cope with a MySQL server that doesn't have space to write MythTV critical data. |
| 02:32 | * | stuarta goes to work |
| 02:32 | <clever> | sphery: it could also crash for other reasons right at that point |
| 02:32 | <sphery> | clever: That means it's a global setting, not a host-based setting. |
| 02:32 | <clever> | another thread may happen to segfault then |
| 02:32 | <clever> | yeah but its not under the NULL hostname |
| 02:32 | <clever> | thats the problem |
| 02:32 | <clever> | somehow it got moved to the '' hostname |
| 02:33 | <clever> | so now the frontend is failing to find DBSchemaVer |
| 02:33 | <clever> | and is refusing to start |
| 02:34 | <clever> | wtf |
| 02:34 | <clever> | no settings are owned by NULL anymore |
| 02:34 | <clever> | i screwed something up:P |
| 02:34 | <sphery> | your re-indexing, probably |
| 02:35 | <clever> | mysql> update settings set hostname=NULL where hostname=''; |
| 02:35 | <hads> | Or a stray UPDATE |
| 02:35 | <clever> | that didnt fix it |
| 02:35 | <clever> | theres values belonging to other hosts |
| 02:35 | <clever> | just none to NULL |
| 02:36 | <clever> | ahh |
| 02:36 | <clever> | hostname cant be null |
| 02:36 | <clever> | when did that happen... |
| 02:36 | <clever> | Rows matched: 127 Changed: 127 Warnings: 0 |
| 02:36 | <Deek> | Primary keys can't be null. |
| 02:37 | <clever> | that was my guess |
| 02:37 | <clever> | what about uniq's? |
| 02:39 | -!- | czth__ [n=dbrobins@nat/microsoft/x-8a414d7ecd6f446e] has joined #mythtv |
| 02:41 | <clever> | ALTER TABLE `settings` DROP INDEX `value` , |
| 02:41 | <clever> | ADD UNIQUE `value` ( `value` , `hostname` ) |
| 02:41 | <clever> | and hostname can still be null! |
| 02:42 | <sphery> | CDev: I updated to trunk r16421, and was unable to get the 100% CPU issue to resurface. Tried 1) deleting all png's in the recordings directory, 2) deleting all png's in the recordings directories and in the MythWeb cache, 3) deleting all png's in the MythWeb cache and never caused the backends to hit 100% (Athlon XP 2400+ and 2000+ doing previews for 442 HDTV shows). I'm pretty sure xris was using trunk, too, and he was only 27 ... |
| 02:42 | <clever> | seems to be working without any bugs now |
| 02:43 | <sphery> | ... changesets behind me when he saw the issue resurface. Guess he'll have to help you track it down. |
| 02:43 | <clever> | sphery: i can reproduce that bug on my end |
| 02:47 | <sphery> | What bug? The 100% CPU thing? |
| 02:47 | <clever> | yeah |
| 02:47 | <sphery> | trunk, right? what rev? |
| 02:47 | <clever> | MythTV Version : 16319M |
| 02:47 | <clever> | ive updated recently but ive mostly just been avoiding triggering it |
| 02:47 | <clever> | but i beleive its still there |
| 02:48 | <sphery> | I just updated from r16316, where I couldn't reproduce the issue. I'm now on r16421, and I still can't reproduce it. |
| 02:48 | -!- | _splat1 [n=splat1@69.10.135.137] has joined #mythtv |
| 02:49 | -!- | splat1 [n=splat1@rf1.splat1.com] has quit [Read error: 104 (Connection reset by peer)] |
| 02:49 | <clever> | ive got a healthy mix of framegrabber recordings, nuv's from transcode, and mpg's from a pvr150 |
| 02:49 | <clever> | it could be a certain random header value in one of them causing it |
| 02:49 | <clever> | like the nuv files that segfault mythtranscode yet play just fine |
| 02:49 | <sphery> | Can you test with r16319? Try rm *.png in the recordings and MythWeb cache (mythweb/data/cache) directories, then use MythWeb to call the Recorded Programs page (when your backend is not scheduled to record for a while). |
| 02:50 | <clever> | allready in the process of testing it without the rm |
| 02:50 | <clever> | recording list is still loading |
| 02:50 | <clever> | db and mythweb server is 400mhz so its a bit slow |
| 02:50 | <sphery> | OK. If that doesn't do it, might need the rm to really strestt test it. |
| 02:51 | <clever> | the probly isnt exactly 100% while making the images |
| 02:51 | <clever> | but 100% that stays there long after it finishes |
| 02:51 | <sphery> | Yeah, it's the 100% after the fact that's the problem. |
| 02:51 | -!- | ahbritto [n=guest@adsl-68-125-197-181.dsl.pltn13.pacbell.net] has joined #mythtv |
| 02:51 | <clever> | same bug then:) |
| 02:51 | <sphery> | (Though mine never hits 100%--while making previews, or after--since the problem disappeared (unexpectedly). |
| 02:51 | <clever> | it also seems to be generating the preview images in a sync method(not async) |
| 02:52 | <clever> | add in the network delays between mythweb and the master and 300 recordings and it gets damn slow |
| 02:52 | <clever> | i also think theres something horidly wrong with my P4 box |
| 02:52 | <sphery> | Yeah, there's a new thread for each request, and the thread can't do anything else until it's got an image to return int he response... |
| 02:53 | <clever> | 400mhz ubuntu 7.10 has a relatively fast phpmyadmin |
| 02:53 | <clever> | 1.6ghz ubuntu 6.06 has a painfully slow phpmyadmin |
| 02:53 | <sphery> | BTW, thanks for testing. CDev couldn't reproduce it this afternoon, and I told him I'd try (since I was previously affected). |
| 02:53 | <clever> | and alot of other things are also oddly slow |
| 02:53 | <clever> | that just doesnt seem right |
| 02:53 | <Chutt> | sphery, it's still fixed here as well |
| 02:53 | <sphery> | Yeah. Strange |
| 02:54 | <clever> | it was even more strange before i put linux on it |
| 02:54 | <clever> | it was runing xp just fine |
| 02:54 | <clever> | i found a stick of ram that looked right |
| 02:54 | <clever> | shutdown and shoved it in |
| 02:54 | <clever> | bios yelled at me for bad ram and cpu freq settings |
| 02:54 | <sphery> | Wonder why xris is seeing it, again? If it's really there, it's a very timing-sensitive bug. |
| 02:54 | <clever> | (it can play some wav files thru the onboard sound card) |
| 02:55 | <clever> | i removed the ram and because the bios made me think of the cpu settings, overclocked it by maybe 5% |
| 02:55 | <Chutt> | it was happening before within the first couple previews created |
| 02:55 | <Chutt> | without fail |
| 02:55 | <clever> | booted xp back up |
| 02:55 | <clever> | it turned into a 5mhz box! |
| 02:55 | <clever> | horidly slow |
| 02:55 | <clever> | dads fix all answer of reboot went into play |
| 02:56 | <clever> | it locked up during shutdown and never booted that drive since |
| 02:56 | <clever> | and the drive was extremely hot |
| 02:56 | <clever> | somehow a stick of ram damaged the hdd |
| 02:56 | -!- | czth_ [n=dbrobins@nat/microsoft/x-ed76fd6821ebcf0a] has quit [Read error: 110 (Connection timed out)] |
| 02:57 | <clever> | and may also be to blame for why its so slow still |
| 02:58 | <clever> | if i do power the hdd up it gives off an odd sound and works fine for the first hour |
| 02:58 | <clever> | and once it heats back up it just stops functioning |
| 02:58 | <clever> | any1 got some dry ice so i can finaly backup the whole thing?:P |
| 03:00 | <sphery> | Anduin: Looks like paulh independently arrived at the same conclusion as you regarding the RefCounted.Release() causing #4816. |
| 03:00 | <Chutt> | it's a race to see who fixes it first |
| 03:01 | <Chutt> | winner gets, um |
| 03:01 | <sphery> | a commit message with their name on it? |
| 03:01 | <clever> | and this lirc bug is damn anoying in config pages |
| 03:01 | <clever> | when i 'press' a button it gets stuck pushed in 60% of the time |
| 03:02 | <sphery> | clever: I saw you mention a failed automatic DB backup the other day. Do you have any info on what caused it (i.e. logs or anything)? |
| 03:02 | <clever> | sphery: a change to the db schema i manualy did caused hostname to become 'not null' causing all global settings to get shoved over to the '' hostname |
| 03:02 | <clever> | o wait no that was something else |
| 03:03 | <clever> | not shure what caused the db backup to fail |
| 03:03 | <sphery> | yeah, this was from a couple of days ago. |
| 03:03 | <Anduin> | sphery: Yeah, I still have an occasional crash on exit I still need time to track down, I really should get better at updating tickets rather than assuming everyone who needs to know is on irc. |
| 03:03 | <clever> | i didnt look that closely |
| 03:03 | <clever> | it was from when i did the svn update |
| 03:03 | <clever> | most i can do is update to a new db schema and see if it fails again |
| 03:03 | <Anduin> | (crash on exist is in roughly the same area) |
| 03:03 | <sphery> | OK. Just wondered. If you see it again, please let me know (and I'd appreciate your keeping a copy of logs for me). THanks. |
| 03:03 | <clever> | the log is long gone |
| 03:04 | <clever> | screen is keeping the last few 1000 lines in scrollback |
| 03:04 | <clever> | i just hit the 100% bug |
| 03:04 | <clever> | 50% system cpu usage |
| 03:04 | <clever> | and the recording list isnt even done yet |
| 03:05 | <Dibblah> | clever: Current SVN? |
| 03:05 | <sphery> | clever: If you update to head, you'll get 1213, which can be re-run over and over (great for testing) by just changing the value of DBSchemaVer. |
| 03:05 | <clever> | MythTV Version : 16319M |
| 03:05 | <clever> | 16319 still has the 100% cpu bug |
| 03:06 | <sphery> | Dibblah: it cleared up for me by the time I upgraded to trunk r16316, and I'm not seeing it now at r16421, so he's in between those two... |
| 03:06 | <clever> | its one of the recordings between dec 1st and today |
| 03:07 | <clever> | 81% cpu to mythbackend and Cpu(s): 38.0% us, 57.7% sy, 1.6% ni, 0.0% id, 0.0% wa, 0.0% hi, 2.7% si |
| 03:07 | <Dibblah> | Could it be the seeking bug? |
| 03:07 | <sphery> | It shouldn't be the actual recording causing it (or the preview generation). All backtraces we had seemed to indicate it was the http server. |
| 03:07 | <Dibblah> | Meaning the seektable is offset, therefore the I-frame isn't aligned? |
| 03:08 | <clever> | i'll update all my systems to head and see if it happens again |
| 03:08 | <sphery> | Anduin: perhaps you could compare notes with paulh on the -dev list as he's still checking his "hypothesis" |
| 03:08 | <clever> | Updated to revision 16421. |
| 03:09 | <sphery> | let me know how your DB backup goes, please. :) |
| 03:09 | <Chutt> | guh |
| 03:09 | <Chutt> | i'm up to 30 bugs at work now |
| 03:09 | <xris> | only? |
| 03:10 | <Chutt> | i prefer to keep it around 10 |
| 03:11 | <clever> | ive also tracked down a DST bug i had |
| 03:12 | <clever> | most of my frontends where unable to cancle recordings past the DST change |
| 03:12 | <clever> | the master is forced into TZ=AST4 without any dst rules |
| 03:12 | <xris> | Chutt: small dev team leads to lots of small ignored bugs |
| 03:12 | <clever> | all the slaves are properly set for the local zone(currently gmt-4) and will shift by an hour after the change |
| 03:13 | <sphery> | clever: using MythWeb or using mythfrontend to cancel? |
| 03:13 | <clever> | so my slaves are addjusting for dst and the master with its poor TZ isnt |
| 03:13 | <Chutt> | xris, my team is 3 people |
| 03:13 | <clever> | sphery: frontend to cancle |
| 03:13 | <xris> | Chutt: there are two of us working on an ERP for a $100M company |
| 03:13 | <clever> | after forcing the slave frontend into TZ=AST4 without any dst rules i 'fixed' it |
| 03:13 | <xris> | used to be just one.. I'll let you guess what state the code is in... and it's perl. :) |
| 03:13 | <clever> | which prooved that it was the improperly set tz |
| 03:14 | <clever> | i probly need something like TZ=America/Halifax |
| 03:14 | <sphery> | Oh. Well, TTBOMK, Myth requires all frontends/backends to be in the same time zone (with same DST settings). MythWeb still may have an issue with DST even if all have same TZ/DST. |
| 03:14 | <clever> | which will bring the dst rules into all myth systems |
| 03:14 | <xris> | the mythweb issue is with recordings made in/out of DST |
| 03:14 | <clever> | i was unable to set TZ in mythweb |
| 03:15 | <xris> | clever: it's in php.ini |
| 03:15 | <clever> | so i just gave up and moved it to a box with the proper /etc/logcheck/ |
| 03:15 | <clever> | i mean /etc/localtime |
| 03:15 | <clever> | i was using set_env within the mythweb part of apache.conf |
| 03:15 | <clever> | and doing it in php.ini would break all my other php scripts which depend on the global tz being something else |
| 03:16 | <clever> | sticking it on a box with /etc/localtime set right fixed it and should also reduce the mythweb<->db delays(db is now local to mythweb) |
| 03:19 | <clever> | sphery: when i had done the TZ=AST4 change i didnt think about the dst rules it would cause to break, but it now seems simple to fix |
| 03:21 | * | clever uploads a ticket with the setting change |
| 03:22 | <Chutt> | could someone merge 16419 to the branch, please? |
| 03:22 | <Chutt> | i've gotta get some sleep, but want it in there for testing.. |
| 03:25 | <janneg> | I'll do it |
| 03:25 | <clever> | http://svn.mythtv.org/trac/ticket/4884 |
| 03:25 | -!- | xris [n=xris@xris.forevermore.net] has quit [] |
| 03:26 | <clever> | ive tested a few things with it on my end and it seems bug free |
| 03:26 | -!- | Tanthrix [n=tanthrix@c-71-56-144-229.hsd1.or.comcast.net] has joined #mythtv |
| 03:27 | <janneg> | done |
| 03:28 | <Chutt> | thanks |
| 03:31 | <clever> | who wants to test out my patch? |
| 04:00 | -!- | rn114 [n=rn114@robthebob.plus.com] has joined #mythtv |
| 04:07 | -!- | mykeul [n=mykeul@81.255.197.125] has joined #mythtv |
| 04:13 | <clever> | sphery: nearly updated to head |
| 04:14 | <clever> | yay |
| 04:14 | <clever> | WARNING: MythTV was unable to backup your database. |
| 04:14 | <clever> | http://pastebin.ca/931643 |
| 04:18 | <sphery> | clever: is mysqldump in the PATH of the user running mythbackend (or mythtv-setup or whatever you ran)? |
| 04:18 | <clever> | mythtv@theP4:~$ type mysqldump |
| 04:18 | <clever> | mysqldump is /usr/bin/mysqldump |
| 04:18 | <clever> | bash can find it:) |
| 04:18 | -!- | rn114 [n=rn114@robthebob.plus.com] has quit [Read error: 113 (No route to host)] |
| 04:18 | -!- | PointyPumper [i=Pintlezz@190.19.128.73] has joined #mythtv |
| 04:19 | <sphery> | was mythbackend/mythtv-setup started from the same environment (i.e. in the same shell that could find mysqldump)? |
| 04:20 | <clever> | i allways run mythbackend and friends from the shell |
| 04:20 | <clever> | on the master |
| 04:20 | <clever> | and some slaves |
| 04:20 | <clever> | and i ran the mythbackend in a shell WITHOUT the & this time |
| 04:21 | <sphery> | Hmmm. 512 looks like it should be a MySQL connection error. Do you have a password (that's not the password for user mythtv@192.168.1.103) specified in a mysql options file (~/.my.cnf)? |
| 04:21 | <clever> | because it seems to detect the lack of stdin and not ask(causing it to use stupid defaults) |
| 04:21 | <clever> | ls: /home/mythtv/.my.cnf: No such file or directory |
| 04:22 | <clever> | removing the 2>/dev/null may help:P |
| 04:22 | <sphery> | what do you get when you run: mysqldump -p --host='192.168.1.103' --user='mythtv' --add-drop-table --add-locks --allow-keywords --complete-insert --extended-insert --lock-tables --no-create-db --quick 'mythconverg' > '/media/mainlv/mythtv/mythconverg-1212-20080307051409.sql' |
| 04:23 | <clever> | mysqldump: Got error: 1146: Table 'mythconverg.weatherdatalayout' doesn't exist when using LOCK TABLES |
| 04:23 | <clever> | a similar problem was killing mythweb |
| 04:23 | <clever> | ive moved my mysql server between several hosts |
| 04:23 | <clever> | and the inodb files got left behind |
| 04:24 | <sphery> | you don't have InnoDB storage enabled, do you? |
| 04:24 | <clever> | so the inodb tables are screwed up |
| 04:24 | <clever> | root@media:/media/videos/media/mythconverg# rm weatherdatalayout.frm |
| 04:25 | <clever> | now the dump is doing something... |
| 04:25 | <sphery> | Hmmm. Don't think it's worth the code to detect that--would require passing a list of all the tables I want mysqldump to backup. |
| 04:25 | <clever> | | 87437 | mythtv | theP4:42103 | mythconverg | Query | 2 | Sending data | SELECT /*!40001 SQL_NO_CACHE */ * FROM `recordedseek` | |
| 04:25 | <clever> | a better way is to just check for the error ahead of time |
| 04:25 | <clever> | and maybe try and fix it |
| 04:26 | <clever> | but i havent found a way to fix it from mysql |
| 04:26 | <clever> | half the querys claim the table exists |
| 04:26 | <clever> | yet drop table and others claim it doesnt exist |
| 04:26 | <clever> | no way to 'fix' it from mysql querys |
| 04:26 | <sphery> | Yeah. The mythweather reliance on foreign keys (and, therefore, InnoDB storage) may catch quite a few people off guard. |
| 04:27 | <clever> | i couldnt even get it to work when i last tried |
| 04:27 | <sphery> | MySQL seems to do stupid things if the storage engine isn't supported (i.e. creates "ghost" tables). |
| 04:27 | <clever> | so i just rm'ed the .frm file |
| 04:27 | <clever> | but now that i think of it THIS may be exactly why mythweather didnt work! |
| 04:27 | <clever> | id add pages to the layout |
| 04:27 | <clever> | save |
| 04:27 | <clever> | view |
| 04:27 | <clever> | error: no pages in layout |
| 04:28 | <clever> | arg! |
| 04:28 | <clever> | repeat 50 times |
| 04:28 | <clever> | arg! arg! arg! |
| 04:28 | * | clever turns green |
| 04:28 | * | clever hurls monitor thru wall |
| 04:29 | <sphery> | I have a patch that will move all the backup commands to a script to give the user much more control (and allow --debug output). Finishing up the final script (it's now a Perl rewrite of one that was originally a shell script). |
| 04:29 | <sphery> | Yeah. MythWeather 0.21/trunk requires InnoDB. |
| 04:29 | <clever> | and playing musical chairs with the datadir and 3 mysql servers may have screwed the tables up before i tried to actualy use it |
| 04:30 | <clever> | so that just leaves the 100% cpu bug and the lirc bug |
| 04:30 | <clever> | done upgrading to 16421 |
| 04:30 | <clever> | Cpu(s): 62.1% us, 10.8% sy, 23.2% ni, 0.0% id, 0.0% wa, 0.5% hi, 3.4% si |
| 04:31 | * | clever pulls the trigger |
| 04:31 | <clever> | whoa |
| 04:32 | <clever> | the recording list raced thru up to dec area |
| 04:32 | <clever> | that took 20mins to load last time |
| 04:32 | <clever> | some cache must finaly be doing its job:P |
| 04:34 | <sphery> | You may need to rm *.png in the recordings directories and in the mythweb/data/cache directory |
| 04:34 | <sphery> | (to test the 100% CPU issue) |
| 04:34 | <clever> | its only got ~ 40% of the images cached |
| 04:34 | <clever> | its still got several months to generate |
| 04:35 | <clever> | its down to oct now |
| 04:35 | <clever> | | 2007-05-19 20:00:00 | |
| 04:35 | <clever> | oldest recording |
| 04:37 | <clever> | still crunching on oct... |
| 04:40 | <clever> | aug |
| 04:41 | <clever> | jul |
| 04:42 | <clever> | finaly done loading the recording list |
| 04:42 | <clever> | no bug |
| 04:42 | <clever> | rm *.png away |
| 04:43 | <clever> | this is taking awhile to simply remove some png files... |
| 04:44 | <clever> | reloading:) |
| 04:47 | <clever> | thats odd |
| 04:47 | <clever> | the --generate-preview has a nice value of 9 |
| 04:47 | <clever> | that cant be helping performance:P |
| 04:54 | <clever> | nothing yet and ive reached feb |
| 04:58 | -!- | mzb_d800 [n=mzb@ppp108-88.static.internode.on.net] has quit [Read error: 113 (No route to host)] |
| 05:03 | <sphery> | clever: cool. sounds good. Now we just have to figure out why xris is affected by the 100% CPU thing. |
| 05:03 | <clever> | still crunching feb |
| 05:04 | <clever> | and unaffected |
| 05:04 | <clever> | it hit me during dec last time |
| 05:04 | <clever> | 604 recordings! |
| 05:08 | <clever> | boinc still has 26% of the cpu |
| 05:08 | <clever> | and it reniced itself to 19 |
| 05:10 | <clever> | jan... |
| 05:27 | -!- | camelreef [n=nico@user-5af0e527.wfd96.dsl.pol.co.uk] has joined #mythtv |
| 05:27 | <camelreef> | good morning from Scotland |
| 05:27 | <camelreef> | oh, wrong channel, I'm a user |
| 05:28 | -!- | camelreef [n=nico@user-5af0e527.wfd96.dsl.pol.co.uk] has left #mythtv [""oh god a "l33t3r" (c) 2007 leftovers"] |
| 05:30 | <clever> | oct and still no crash... |
| 05:41 | <clever> | done loading and no crash! |
| 05:41 | <clever> | the bug is either lucky or dead |
| 05:42 | <clever> | 2008-03-07 06:41:44.774 Preview Error: Run() file not local: '/GetPlaybackURL/UNABLE/TO/FIND/LOCAL/FILE/ON/theP4/1045_20071205125900.nuv' |
| 05:43 | <clever> | what if i was to make such a file just to screw with mythtv:P |
| 05:52 | <clever> | sphery: a 2nd problem ive noticed |
| 05:52 | <clever> | i thought i commented out the line to generate preview images to work arround the 100% cpu bug |
| 05:52 | <clever> | yet somehow its come back on its own:S |
| 05:53 | -!- | mntmst [n=root@pool-71-117-252-159.ptldor.fios.verizon.net] has quit [Read error: 110 (Connection timed out)] |
| 05:55 | <gbee> | if someone supplies a patch to replace the reliance on foreign keys in mythweather then I'll include it for 0.21, but I don't have time to do all the work myself |
| 06:06 | -!- | mzb [n=ubernut@ppp108-88.static.internode.on.net] has quit [Read error: 113 (No route to host)] |
| 06:22 | -!- | clever_ [n=clever@fctnnbsc16w-156034209103.nb.aliant.net] has joined #mythtv |
| 06:36 | -!- | clever [n=clever@fctnnbsc16w-156034210056.nb.aliant.net] has quit [Read error: 110 (Connection timed out)] |
| 06:44 | -!- | mntmst [n=root@pool-71-182-79-65.ptldor.fios.verizon.net] has joined #mythtv |
| 06:58 | -!- | MrGandalv [i=mgandalf@cpe-72-225-32-214.rochester.res.rr.com] has quit ["work"] |
| 07:30 | -!- | GBee1 [n=gbee@cpc2-derb9-0-0-cust34.leic.cable.ntl.com] has joined #mythtv |
| 07:40 | <teprrr> | btw, is there no menu editor available? |
| 07:40 | <teprrr> | just wondering whether I should start creating one.. |
| 07:41 | -!- | dagar [n=dagar@206-248-137-66.dsl.teksavvy.com] has quit [Read error: 110 (Connection timed out)] |
| 07:42 | -!- | dagar [n=dagar@206-248-137-66.dsl.teksavvy.com] has joined #mythtv |
| 07:46 | -!- | gbee [n=gbee@cpc2-derb9-0-0-cust34.leic.cable.ntl.com] has quit [Read error: 110 (Connection timed out)] |
| 07:48 | -!- | briand [n=brian@c-68-35-254-149.hsd1.fl.comcast.net] has quit [Read error: 110 (Connection timed out)] |
| 08:01 | -!- | GBee1 is now known as gbee |
| 08:02 | <gbee> | teprrr: just edit the xml, not sure it's really worthwhile working on an editor |
| 08:03 | -!- | clever_ is now known as clever |
| 08:04 | -!- | reynaldo [n=rverdejo@190-82-33-234.adsl.cust.tie.cl] has joined #mythtv |
| 08:04 | <gbee> | there are plenty of ways to help out MythTV (there are 300+ outstanding tickets) but if a menu editor is what you really want to work on, then don't let me stop you ;) |
| 08:04 | <teprrr> | gbee, imho it'd be cool if there was support for dynamic menus based on what some plugins provide+core functionality.. and which would be configurable easily :P |
| 08:04 | <teprrr> | yeah, sure, perhaps I'll take a look more soon |
| 08:42 | -!- | ToadP [n=jboyle@bas8-ottawa23-1177762653.dsl.bell.ca] has quit [Remote closed the connection] |
| 08:42 | -!- | ToadP [n=jboyle@bas8-ottawa23-1177762653.dsl.bell.ca] has joined #mythtv |
| 09:01 | -!- | briand [n=brian@c-68-35-254-149.hsd1.fl.comcast.net] has joined #mythtv |
| 09:26 | -!- | Cardoe [n=Cardoe@gentoo/developer/Cardoe] has joined #mythtv |
| 09:36 | <gbee> | http://www.theregister.co.uk/2008/03/07/patent_crackdown_at_cebit/ |
| 09:37 | <gbee> | leave mythmusic out of the demo at linuxtag? :P |
| 09:45 | <GreyFoxx> | Hmmm looks like I had several recordings not end at the proper time this week |
| 09:45 | <GreyFoxx> | 2 half hour shows, 1 recorded for 5.5 hours (12G file) and another recorded for 10hours (23G file) |
| 09:46 | <GreyFoxx> | both are on my master backend with pvr cards |
| 09:46 | <danielk22> | Wow! I was going to say check to see if ntpd is still working, but that usually accounts for at most a few minutes.. |
| 09:46 | <GreyFoxx> | yeah |
| 09:47 | <GreyFoxx> | the 12G one was last night, and the 23G was Monday |
| 09:47 | <GreyFoxx> | in fact now that I think about it I had a firewire HDTV recording a couple days ago that was reporting 4 hours long even though the show was 1 hour. I didn't actually verify if it was longer than an hour |
| 09:47 | <jamesd> | sounds like bug, perhaps a conspiracy by the harddrive makers to sell 1TB drives to the myth community ;-) |
| 09:53 | <GreyFoxx> | http://www.pastebin.ca/932120 Between 4:30am and 10am nothing was logged but regular upnp stuff |
| 09:54 | <GreyFoxx> | recording starts at 4:30am and then another starts at 10am on another tuner. At the same time as the new one starts the first one ends |
| 09:57 | -!- | packetscan [i=Tsu@srv01.packetscan.com] has quit [Remote closed the connection] |
| 09:58 | <gbee> | GreyFoxx: similar problems with livetv got me thinking about adding checks and maximum recording size/length safeguards - http://svn.mythtv.org/trac/ticket/3618 |
| 09:58 | -!- | packetscan [i=Tsu@srv01.packetscan.com] has joined #mythtv |
| 09:59 | <GreyFoxx> | hmmmm each of the two I still have on my system that did this, both recorded from the same tuner, on the same channel |
| 10:00 | <GreyFoxx> | and both ended within 1-2 minutes of a logged "Reschedule requested for id 0." |
| 10:00 | <GreyFoxx> | I'm sure it happened on afirewire recording as well, but I deleted that already |
| 10:01 | <gbee> | odd |
| 10:01 | <gbee> | Chutt: this is the fix I'm thinking about for #4756 - http://pastebin.ca/932135 |
| 10:02 | <Chutt> | does it actually work? |
| 10:03 | <gbee> | Chutt: not a clue, I can't reproduce the crash so I'm just guessing at the likely cause from the backtrace |
| 10:03 | <Chutt> | heh |
| 10:04 | <Chutt> | have you used both qt and ogl painters to repro? |
| 10:04 | <gbee> | calling popscreen directly is a little too dangerous as we potentially delete the wrong screen, at least if DeleteScreen is called it should only ever delete itself (well that's the theory) |
| 10:04 | <Chutt> | right |
| 10:05 | <gbee> | Chutt: no, I should probably try QT as it bypasses the fade ... |
| 10:05 | <Chutt> | PopScreen should probably take the screen as parameter |
| 10:06 | <Chutt> | but still, i still don't see how that'd get called twice |
| 10:06 | <gbee> | yeah, that's the other patch I wrote, on balance I liked the first one as it was simpler and seemed less likely to trip up |
| 10:06 | <Chutt> | ooh |
| 10:06 | <Chutt> | one sec |
| 10:07 | <gbee> | Chutt: QApplication::postEvent(mainwindow, new ExitToMainMenuEvent()); |
| 10:07 | <Chutt> | yes |
| 10:08 | <Chutt> | and? |
| 10:08 | <Chutt> | postEvent's fine |
| 10:09 | <gbee> | heh, well I don't fully see how it's happening, but the jumpoint calls popscreen -> popscreen sends an Escape keypress -> Escape triggers a second popscreen in mythweather |
| 10:10 | <gbee> | by the time the second popscreen is handled top (mythweather screen) has been deleted and so it deletes the main menu instead |
| 10:10 | <Chutt> | where does the jumppoint call popscreen? |
| 10:10 | <Chutt> | (it shouldn't) |
| 10:19 | <gbee> | jumppoint exits to mainmenu first by sending the ExitToMainMenuEvent() event, which sends the Escape keypress, which triggers a call to popscreen |
| 10:19 | <danielk22> | ugh, popups in Qt.. that's why the channel scanner is so crash prone :( |
| 10:19 | <gbee> | sorry it took so long, kdevelop decided to crash while I was tracing it through |
| 10:21 | <gbee> | JumpTo -> Escape Keypress -> MythAppear calls popscreen -> Popscreen -> Escape Keypress (why?) -> Popscreen -> Segfault |
| 10:22 | <Chutt> | i still don't see how that'd happen |
| 10:22 | <Chutt> | JumpTo sends an event |
| 10:22 | <Chutt> | next Qt loop iteration, it processes event, calls ExitToMainMenu |
| 10:22 | <Chutt> | that checks the top screen, and it's not 'mainmenu', so it sends an escape to exit the top screen |
| 10:23 | <Chutt> | next qt loop iteration, mythappear process escape, calls popscreen |
| 10:23 | <Chutt> | popscreen deletes top screen (mainwindow->isexitingtomain is true), postevent's new exittomainmenu event |
| 10:24 | <Chutt> | next qt loop iteration, postEvent process |
| 10:24 | <Chutt> | calls ExitToMainMenu |
| 10:24 | <Chutt> | sees top screen is now 'mainmenu', stops |
| 10:27 | <gbee> | you're right, shouldn't happen as I've stated it - missed the "screen->name() != QString("mainmenu")" in ExitToMainMenu |
| 10:28 | <gbee> | ok, I'm stumped |
| 10:30 | <gbee> | can't reproduce it with the QT painter either |
| 10:31 | -!- | catinpan is now known as catinpan_away |
| 10:31 | -!- | catinpan_away is now known as catinpan |
| 10:31 | -!- | superm1 [n=superm1@ubuntu/member/superm1] has quit ["Leaving"] |
| 10:33 | <gbee> | I guess the Exit event in popscreen is to take us completely back to the main menu if we're more than one screen deep |
| 10:33 | <Chutt> | right |
| 10:34 | <Chutt> | it's meant to go one more level up |
| 10:34 | <Chutt> | then the eventhandler sees we're at mainmenu, and stops the jumpping, and calls the jumppoint |
| 10:34 | <gbee> | yeah |
| 10:38 | <gbee> | ok, I give up |
| 10:38 | -!- | lcase [n=l-case@p5B0EBA7B.dip0.t-ipconnect.de] has joined #mythtv |
| 10:38 | <gbee> | if I could reproduce it I'd stand a chance of seeing what is going wrong |
| 10:39 | <gbee> | don't think I can expect the user to step through it in gdb for me |
| 10:41 | <Chutt> | bump it to 22? |
| 10:42 | -!- | aevil [n=aevil@ip-78-94-247-37.reserved.ish.de] has joined #mythtv |
| 10:42 | <gbee> | yeah, unless anyone else wants to have a go? |
| 10:43 | <Chutt> | naw |
| 10:49 | <gbee> | well I'm running out of reasons to avoid the mythweather job |
| 10:49 | <Chutt> | =) |
| 11:17 | <GreyFoxx> | What are the major bugs.blockers left ? |
| 11:20 | -!- | gnome42 [n=gnome42@76-10-151-103.dsl.teksavvy.com] has joined #mythtv |
| 11:21 | <sphery> | There's a report from xris of the 100% CPU on XML preview request thing returning, but I know of at least 4 people who were previously affected who can not reproduce with current head. |
| 11:23 | <gbee> | utf8 fixes for upnp, but I'm not sure I'd call it a blocker just something that shouldn't be too hard to fix |
| 11:24 | <gbee> | probably just needs local8bit conversions on the file and directory names read directly from the filesystem |
| 11:24 | <gbee> | http://svn.mythtv.org/trac/ticket/4372 < If anyone wants to help out |
| 11:26 | <gbee> | channel scanner segfaults, video profile naming, mythweb not handling recordings created during DST and the segfault in SSDP code are the main ones |
| 11:27 | <sphery> | Sounds like Anduin/paulh have a good start on the segfault in SSDP code. |
| 11:27 | <danielk22> | gbee: i'm about to look at video profiles |
| 11:28 | <sphery> | xris/kormoc don't even have a plan for fixing the MythWeb DST issue, right? |
| 11:29 | <gbee> | sphery: there was a plan, but it involved backend changes iirc |
| 11:29 | <sphery> | Oh, yeah. Change the programinfo to use unix timestamps, right? |
| 11:29 | <sphery> | sounds a bit dangerous for right before a release... |
| 11:30 | <gbee> | yeah, but it's a big enough bug that it might be better to fix it now |
| 11:30 | <sphery> | unless, perhaps, we have a new protocol command to specifically request a "special" stringlist (so the rest of the code using programinfo is unaffected), perhaps. |
| 11:30 | <danielk22> | yep.. I thought that change was already applied? |
| 11:31 | <danielk22> | the yep, was for 'sounds a bit dangerous' |
| 11:34 | <gbee> | I've already forgotten the important details of the bug |
| 11:36 | <sphery> | me too... What do you think of a new command QUERY_RECORDING*_UNIXDATE or something? Ugly for the long term, but might be an acceptable transitional approach to support MythWeb 0.21 (and allows testing changes to QUERY_RECORDING* in trunk before 0.22). |
| 11:39 | -!- | beavis [n=beavis@drms-590d5fba.pool.einsundeins.de] has joined #mythtv |
| 11:40 | <janneg> | danielk22: you probabbly mix it up with the QUERY_PIXMAP_LASTMODIFIED change which was applied in http://svn.mythtv.org/trac/changeset/16090 |
| 11:41 | -!- | mykeul [n=mykeul@81.255.197.125] has left #mythtv [] |
| 11:41 | <danielk22> | janne: yep, your right |
| 11:44 | -!- | xris [n=xris@66.236.8.178.ptr.us.xo.net] has joined #mythtv |
| 11:45 | <sphery> | There's the guy who knows what needs modified in the backend to fix MythWeb's DST issue... :) |
| 11:45 | <gbee> | sphery: there may be a purely mythweb/php fix, but as I can't remember exactly what was going wrong I'm not sure |
| 11:46 | -!- | _splat1 is now known as splAt1 |
| 11:46 | -!- | splAt1 is now known as splat1 |
| 11:46 | <sphery> | xris: Did you say you needed the QUERY_RECORDING* commands to send Unix dates to fix the DST thing? |
| 11:48 | <xris> | sphery: not sure how/if that would help. |
| 11:49 | <gbee> | one of PHPs date functions should know when a given time falls into a DST period in the current locale, so it may be enough to +1/-1 an hour to the the requested starttime |
| 11:49 | <xris> | I actually don't really know what the issue is with the DST thing.. whether mythweb's converting the date wrong, or what |
| 11:49 | <xris> | gbee: you'd think that. but I don't trust it. I do trust mysql, though |
| 11:50 | <gbee> | let me do some digging and add details to the ticket which I should have done in the first place |
| 11:51 | <gbee> | does no-one else have recordings made during a DST/BST period? I don't know if we've confirmed that this is a genuine bug and not some problem with my install |
| 11:52 | -!- | aevil [n=aevil@ip-78-94-247-37.reserved.ish.de] has quit ["Ex-Chat"] |
| 11:58 | <sphery> | gbee: Do you remember where the issue presents itself? I have recordings from DST that show up fine in my Recorded Programs page (and I've recently regenerated the previews through MythWeb many times without issues). |
| 11:58 | <gbee> | it's the Record Programmes page of mythweb |
| 11:59 | <gbee> | hang on, I'm just posting an example to the ticket |
| 12:00 | <sphery> | cool (I guess previews would have been fixed with QUERY_PIXMAP_LASTMODIFIED, anyway)... |
| 12:05 | <gbee> | added an example of starttime/progstart pulled from the database for "Thirteen Days" and then a screenshot of the time as shown by mythweb |
| 12:06 | <gbee> | no preview image because it's requesting the wrong starttime |
| 12:06 | <gbee> | clicking on the recording gives the error "Unknown programme" |
| 12:08 | <gbee> | that applies to every recording on the backend created during BST (UK equivalent of DST) |
| 12:08 | <gbee> | oh and I've just managed to reproduce the 100%cpu bug :( |
| 12:09 | <danielk22> | gbee: is there a ticket for translation strings for playback profiles? |
| 12:09 | -!- | dekarl [n=deKarl@dslb-084-058-057-177.pools.arcor-ip.net] has joined #mythtv |
| 12:10 | <gbee> | danielk22: don't know of one |
| 12:10 | <danielk22> | k |
| 12:10 | <sphery> | Hmmm. I'm not seeing that. I have recordings from before, during and after the DST change from last year, but they're all showing previews and I can click them to see program details. (BTW, DST starts here in 2 days, so I'm not in it now, but the shows recorded during last years' DST were.) |
| 12:11 | <gbee> | danielk22: the main ticket does mention translations - http://svn.mythtv.org/trac/ticket/4631 |
| 12:11 | <sphery> | gbee: Is your MythWeb on a different host from your backend? If so, do you have the same timezone specified in the php.ini as on the backend? |
| 12:12 | <gbee> | sphery: different host yeah, let me check php.ini (system is on same timezone though) |
| 12:24 | <gbee> | sphery: think you've hit the nail - there was no timezone given in php.ini, when I check phpinfo() it thought the timezone was UTC instead of GMT/BST (I blame the French) |
| 12:25 | -!- | MrGandalf [i=mgandalf@cpe-72-225-42-190.rochester.res.rr.com] has joined #mythtv |
| 12:25 | <sphery> | Actually, xris hit the nail on the head last night when talking to someone else with the same issue. :) |
| 12:25 | -!- | dekar1 [n=deKarl@dslb-084-058-068-230.pools.arcor-ip.net] has quit [Read error: 110 (Connection timed out)] |
| 12:25 | <xris> | yeah, he definitely has a weird setup... runs his server in UTC time |
| 12:26 | <xris> | but php is f'd up by not inheriting timezone info |
| 12:27 | <gbee> | running anything in UTC is f'd up as it's not a proper timezone :p |
| 12:28 | <sphery> | Was looking at the Errors/Exceptions part of http://us2.php.net/date/t but it seems like that wouldn't help us determine if we should warn the user, right? |
| 12:28 | <gbee> | no idea why php wasn't using the system timezone |
| 12:28 | <sphery> | Yeah. It sounds like PHP will warn you if it /is/ using the system timezone. ??? |
| 12:30 | <gbee> | that's just a warning because in many cases your server may not be in the same timezone as your customers etc |
| 12:31 | <gbee> | so it can safely be ignored if that isn't the case |
| 12:31 | <gbee> | but it should be using the system time by default and I've no idea why my install isn't doing that |
| 12:32 | <sphery> | Wonder whether a new protocol command QUERY_TIMEZONE would be useful for detecting this issue... |
| 12:33 | <gbee> | I think we should probably include TZ info and then use date_timezone_set() to force php into using the same |
| 12:35 | <sphery> | include it where? |
| 12:36 | <xris> | gbee: that'd be a nice addition |
| 12:37 | <gbee> | sphery: as part of the time string sent by the backend maybe, or as a seperate protocol command as you've suggested |
| 12:38 | <gbee> | seperate command makes a little more sense, but only if we cache the result somehow |
| 12:39 | <gbee> | part of the session or something |
| 12:39 | <sphery> | Agreed. Might be difficult to match C/PHP TZ abbreviations up all the time... |
| 12:44 | <sphery> | How would the backend reliably determine the timezone it's using? |
| 12:46 | <gbee> | well there I think we have to assume that the system timezone is accurate |
| 12:47 | <gbee> | we can extend the same idea to frontends though, so the frontend doesn't have to be in the same timezone as the backend any more (not that it's a good idea to use a frontend over a remote connection) |
| 12:48 | <sphery> | Yeah. I was looking for more of an implementation... All POSIX stuff |
