All times shown according to UTC.

Time Nick Message
00:20 ajonas left #miro-hackers
00:54 maggie_s left #miro-hackers
02:10 willkg mmm...  seems like my macbook problem is a problem with the charger.  irritating, but not too hard to fix.
02:10 anyhow...  time to jet.  later!
02:10 willkg left #miro-hackers
02:40 mattjohnson left #miro-hackers
02:41 paulproteus left #miro-hackers
02:43 paulproteus joined #miro-hackers
04:22 ajonas joined #miro-hackers
05:22 ajonas left #miro-hackers
06:04 Dawnrell joined #miro-hackers
06:32 Dawnrell left #miro-hackers
07:04 bmon left #miro-hackers
07:06 janetPCF joined #miro-hackers
07:18 lucsky joined #miro-hackers
08:20 janetPCF left #miro-hackers
08:21 janetPCF joined #miro-hackers
08:36 foocraft joined #miro-hackers
09:11 Dawnrell joined #miro-hackers
09:59 arpu joined #miro-hackers
10:03 Dawnrell left #miro-hackers
10:39 arpu left #miro-hackers
13:17 Dawnrell joined #miro-hackers
15:03 z3p morning
15:09 janetPCF1 joined #miro-hackers
15:09 z3p left #miro-hackers
15:10 z3p joined #miro-hackers
15:12 janetPCF left #miro-hackers
15:24 ajonas joined #miro-hackers
15:33 janetPCF1 left #miro-hackers
15:39 janetPCF joined #miro-hackers
16:43 paulproteus Morning, all.
16:49 ajonas: You wanted to chat later today. Did you end up finding Dean and/or picking a time?
16:50 z3p: Can I get a quick sanity check on http://bugzilla.pculture.org/s[…]_bug.cgi?id=14970 ?
16:52 ajonas: My day's priorities are: Land a fix for http://bugzilla.pculture.org/s[…]_bug.cgi?id=14970 ; Work more on tiers; http://bugzilla.pculture.org/s[…]_bug.cgi?id=15075
16:53 ajonas paulproteus: great! looks good.
16:53 z3p paulproteus: other than 'get_string' being a confusing name for that tag, it looks like it'll solve the issue
16:53 paulproteus Yeah, create_query_string might be a better name.
16:53 ajonas paulproteus: does the 1.2 milestone seem vaguely reasonable now? there were a couple things i was loathe to kick out, but if you see stuff you really think should wait, feel free to push it to 1.5.
16:54 z3p paulproteus: or 'create_replacement_query_string' or something, yeah
16:54 paulproteus z3p: b
16:56 ajonas: When are you hoping to have 1.2 by?
16:56 Ouch, the thumbnail bug kind of sucks.
16:57 ajonas paulproteus: i was thinking we'd see how tiers went and then set a date, maybe
16:57 paulproteus Yeah, I like that plan the most. (-:
16:57 Dawnrell left #miro-hackers
16:57 paulproteus I say we hold performance stuff like http://bugzilla.pculture.org/s[…]_bug.cgi?id=14355 off, and then do a performance-only release.
16:58 ajonas fwiw, i can't reproduce the thumbnail problem on a live test site
16:58 paulproteus Huh, okay.
16:58 ajonas paulproteus: that's fine w me
16:58 paulproteus Could have been fixed, then?
16:58 ajonas: I am personally extremely excited about a performance-only release right after 1.2.
16:59 ajonas paulproteus: i think that makes sense, esp if that's both performance for end users and for being more efficient on our resources
16:59 paulproteus Yeah, exactly. Conveniently, they're the same. (-:
17:00 Well, not literally the same, but they share a lot, anyway.
17:02 Can we similarly move http://bugzilla.pculture.org/s[…]_bug.cgi?id=13628 to the performance release?
17:03 (Do we call that 1.3? 1.5?)
17:04 z3p: What causes the emails to get sent, as per http://bugzilla.pculture.org/s[…]_bug.cgi?id=15075 ? Is it a cron job (AKA management command)?
17:04 Like, what triggers the once-daily action?
17:04 ajonas paulproteus: if we must, though we'd really like to minimize the load. but it's not unreasonable.
17:05 paulproteus For some reason, I have this feeling that I want to focus more on code-type stuff and less on sysadmin-y stuff for 1.2.
17:05 ajonas uh, let's make that 1.3, if that makes sense.
17:05 paulproteus I like 1.3 as the name.
17:05 ajonas what kind of things are sys-adminy?
17:05 z3p paulproteus: 'update_original_data'
17:05 paulproteus Site deactivation.
17:06 z3p paulproteus: and then the e-mails are sent from the notifications
17:06 paulproteus ajonas: Things that change the configuration on the web server or relate to our deployment of MC.
17:06 ajonas ok. i'm on board, as long as we tackle deactivation right after 1.2.
17:07 paulproteus Fine by me.
17:08 z3p: b
17:08 There's a cron job that calls this?
17:09 * paulproteus looks
17:10 z3p paulproteus: www-data's cron
17:10 paulproteus b
17:59 willkg joined #miro-hackers
18:16 willkg hi!
18:21 paulproteus Hi!
19:11 CarlFK joined #miro-hackers
19:18 Dawnrell joined #miro-hackers
20:15 Dawnrell left #miro-hackers
20:33 DGMurdockIII joined #miro-hackers
20:42 bendk joined #miro-hackers
20:44 bendk hi all
20:45 willkg: how would you feel about pushing off #14804
20:46 or, alternatively, can you think of a simple way to make it so we add extensions to youtube FLV's
20:46 willkg bendk: so...  this is funny.  i feel like i keep saying the same thing about this.
20:47 flashscraper is not set up to dictate what the file name should be.
20:47 DGMurdockIII left #miro-hackers
20:47 willkg so...  we can add that to whatever it is that figures out the filename.  your idea of looking at the content type and faking it based it on that if there's no extension is probably fine.
20:48 bendk the worry I have about that is now we're changing the behavior for all downloads, not just youtube ones
20:48 although, I guess that adding an extension wouldn't hurt
20:48 willkg i'm still skeptical that the problem is really a problem.  i think the test we're using to determine that is possibly wrong.  "open in miro" does different things.
20:48 bendk I'm almost positive it is
20:48 willkg what we should be doing is running qt_extractor on files that have no extension to see what it does.  that's a more appropriate test.
20:49 the open in miro thing has different code because we wanted miro to be able to open files immediately rather than wait for the moviedata script to run.
20:49 bendk I think it still goes through moviedata eventually
20:49 willkg lucsky: do you remember this?  i'm pretty sure you and i worked through this.
20:49 bendk I make qt_extractor printout it's output, and put that in the bug report
20:49 willkg right...  but it uses file extensions.
20:50 and we're positive this isn't a bug in qt_extractor?
20:50 bendk it might be possible to workaround it, but I think it's a quicktime bug/feature depending on how you want to look at it
20:50 willkg i'm just really surprised that quicktime expects the extension to be correct rather than just check the file for appropriate fingerprints indicating its media type.
20:50 bendk I know I've seen similar things in the past
20:50 willkg ok.
20:50 bendk yes, it's really stupid
20:51 willkg ok.  well, let's assume i'm wrong and go from there.
20:51 bendk when using final cut pro, I've had to rename movie files so that the had an extension to get it to be able to import the movies
20:51 willkg youtube still serves a lot of flv files.  so having this bug causes osx users to not be able to view flv files from youtube which i think is a big problem.
20:51 therefore it should get fixed somehow.
20:52 bendk I guess i'm not sure about the formats, but when I was testing the bug, I downloaded like 10 movies without extensions and they all played well
20:52 it was only when I went for that movie that geoff specifically pointed out that I saw the error
20:52 willkg pretty sure we only look at fmt codes for .flv and .mp4 files.  and i'm pretty sure i got flashscraper to send the right content type along.
20:52 bendk yes, I believe we do
20:52 okay, you've convinced me
20:53 willkg so...  whatever determines the file name should do a quick check.  hopefully it knows the content type at that point.
20:53 bendk in the downloader daemon, we check to see if we are saving something with no extension, then create an extension based on the content-type
20:53 willkg flashscraper sucks so much.  i'm really kicking myself that i didn't follow paroneayea's lead and ditch it when he wrote vidscraper.  bad decision on my part.
20:55 bendk is it worth it putting this code in master as well?  Seems like it could be useful for movies that don't go through flashscraper/vidscraper as well
20:59 willkg mmm...  yes.
21:00 if i had my druthers, i'd use the same code i wrote for video conversion filenames and change how we do filenames.
21:00 but babystepping in that direction is good in case my druthers don't pan out because i run out of time/energy.
21:00 or we hit some monster technical issue.
21:04 bendk: does that sound ok?
21:04 bendk sounds good to me
21:05 mattjohnson joined #miro-hackers
21:07 janetPCF left #miro-hackers
21:15 willkg the flashscraper -> vidscraper bug is 15190.  the filename one is 8913.
21:15 both are in the 4.0 target.  don't want to lose track of them.
21:29 bendk: i worked on the ffmpeg updates yesterday for osx.  i'm having problems with my macbook.  pretty sure i'm half done, but i'm bumping into conversion issues now.
21:29 bendk: so i'm still working on that.  i haven't even started on the windows side, yet.
21:29 bendk okay
21:30 willkg bendk: and i haven't gotten to investigating 14994.
21:30 bendk seems like an RC is at least few days off, but hopefully we can get it out in that timeframe
21:31 * willkg nods.
21:31 bendk hmm, that one seems nasty
21:31 do you think we need to fix it for 3.5.1?
21:31 willkg i don't know.  all i know is that there are a few different WindowsErrors users are running into.  needs more analysis in regards to how many users, how frequently, and what the specifics are.
21:32 and what the probably causes are.
21:32 seems like it's a small issue.  none of us can reproduce it.
21:32 that's why i stuck it in the p2 list.
21:35 bendk: one thing i'd love to do if i have enough time is to at least put more logging into those subprocess calls so we can better ascertain the context and what's going on.
21:36 bendk: like...  what path is it looking at?  is there a file there?  what are the permissions on the file?  ...
21:37 bendk that would be a huge step
21:37 if that's all we did, we could probably have a fix for 3.5.2/4.0
21:37 * willkg nods.
21:37 willkg i'll try to get to it.  i'll let you know if i can't.
21:38 anyhow...  off to ffmpeg stuff now.  afk a bit.
21:46 mattjohnson left #miro-hackers
21:52 maggie_s joined #miro-hackers
22:00 willkg ugh.  i can't believe it's 5:00pm already.
22:00 screw you time!  why can't you slow down so i can get stuff done!
22:06 lucsky left #miro-hackers
22:38 CarlFK left #miro-hackers
22:46 willkg bendk: i think i figured out some/most of the ffmpeg issues i had.  i'll check the osx code in soon.
22:46 bendk: but i think it causes breakage with the args, so i'll have to look at that.
22:47 afk for dinner....
23:03 maggie_s left #miro-hackers
23:14 bmon joined #miro-hackers
23:19 foocraft left #miro-hackers
23:27 paroneayea left #miro-hackers
23:32 foocraft joined #miro-hackers

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first