All times shown according to UTC.

Time Nick Message
00:13 Sergeiqa left #miro-hackers
00:53 maggie_s paulproteus: Hi!
01:10 paulproteus maggie_s: Hi!
01:11 maggie_s paulproteus: Just wanted to double-check maggie-dot settings - you said you switched this site to no-enforce mode.
01:13 pauproteus: But it actually enforces everything except the number of approved videos - prohibits uploading new themes, adding custom pages and CSS, does not support multiple admins
01:15 paulproteus maggie_s: Huh... I'll double-check
01:17 maggie_s: The new way that no-enforce works is that as soon as you do a paypal transaction, tiers start being enforced.
01:17 FWIW
01:17 That's a change from before, I realize.
01:18 maggie_s pauproteus: yes, but I am speking of the state BEFORE the first PayPal transaction
01:18 *speaking
01:18 paulproteus Weird.
01:19 Let me check one thing
01:19 Then I have to go, I'm afraid
01:19 maggie_s paulproteus: OK
01:20 maggie_s left #miro-hackers
01:21 maggie_s joined #miro-hackers
01:21 paulproteus Hey
01:21 maggie_s pauproteus: yes?
01:21 paulproteus So maggie_s, I guess it's true that there's a bug where we enforce CSS and some other restrictions even when tiers enforcement is "off"
01:21 That's the story.
01:21 I have to run out for now, I'm afraid!
01:22 maggie_s paulproteus: OK, then I'll file it and go on with switching to paid
01:22 paulproteus Thanks!
01:22 maggie_s pauproteus: see you tomorrow!
01:44 DGMurdockIII joined #miro-hackers
02:02 DGMurdockIII left #miro-hackers
02:44 arpu left #miro-hackers
02:48 maggie_s left #miro-hackers
02:49 maggie_s joined #miro-hackers
02:52 maggie_s left #miro-hackers
04:24 Odysimus joined #miro-hackers
05:04 Odysimus left #miro-hackers
05:30 CarlFK joined #miro-hackers
05:45 CarlFK left #miro-hackers
06:17 Odysimus joined #miro-hackers
07:24 Odysimus left #miro-hackers
07:36 Odysimus joined #miro-hackers
08:22 janetPCF joined #miro-hackers
08:34 Odysimus left #miro-hackers
08:34 Odysimus joined #miro-hackers
08:37 janetPCF left #miro-hackers
08:49 janetPCF joined #miro-hackers
08:51 Odysimus left #miro-hackers
08:57 maggie_s joined #miro-hackers
09:47 maggie_s left #miro-hackers
09:55 glee joined #miro-hackers
11:56 janetPCF left #miro-hackers
12:31 z3p left #miro-hackers
12:42 janetPCF joined #miro-hackers
13:20 z3p joined #miro-hackers
13:39 z3p deardiary: working on misc. P1s
13:42 CarlFK joined #miro-hackers
13:53 willkg joined #miro-hackers
13:54 janetPCF left #miro-hackers
13:55 janetPCF joined #miro-hackers
14:01 willkg hi!
14:30 CarlFK left #miro-hackers
14:34 ajonas joined #miro-hackers
15:14 willkg deardiary: i'm back.  planning to work on P1 bugs i grabbed over the last week or so.
15:24 CarlFK joined #miro-hackers
15:31 CarlFK left #miro-hackers
15:36 z3p left #miro-hackers
15:38 z3p joined #miro-hackers
16:17 willkg left #miro-hackers
16:25 bendk1 joined #miro-hackers
16:25 bendk left #miro-hackers
16:32 willkg joined #miro-hackers
16:38 zanoi willkg, ping
16:39 willkg zanoi: hi!
16:39 zanoi hey
16:39 would you have time to talk about the ogg stuff?
16:40 willkg sure!  what's up with ogg stuff?
16:40 and pretend i only vaguely know what "ogg stuff" means because i've been out of the loop for the last week.
16:40 zanoi it's about this: http://bugzilla.pculture.org/s[…]g.cgi?id=16436#c6
16:41 do i understand it correctly that we currently decide what an item is purely based on extension?
16:42 willkg no.
16:42 zanoi ok
16:42 how do we decide then?
16:42 willkg that's a good question for kcw who can walk you through the gory details.
16:42 zanoi well except for if it's a feed and thus has mimetype metainfo
16:42 then we use that i think
16:42 willkg it depends.
16:43 zanoi and in general there is lots of guessing when downloading, but i'm wondering mostly about normal local files
16:43 willkg if you do "open item" from the "file" menu, i think miro does a good guess based on the file extension and then the renderer will adjust accordingly.  (the last part might not be true anymore.)
16:43 zanoi kcw, ^
16:44 willkg otherwise, the metadata extractor gets run on the item and we use whatever it gives us.
16:44 zanoi what do you mean by metadata extractor?
16:45 is that a class or do you just mean we extract the metadata
16:45 willkg it's not a class.  there's an extractor for every renderer.  it's a thing that runs in its own process to figure out what kind of thing a file is and extracts a duration and a thumbnail if appropriate.
16:46 kcw overhauled that to use mutagen as a first pass, i think.  i haven't really looked at his changes, yet.
16:46 zanoi so why do we associate mimetypes with extensions in filetypes.py?
16:47 willkg there are a few possibilities:
16:47 1) it's dead code.
16:47 2) it's used when someone does "open item".
16:47 3) it's used in a place i've forgotten.
16:47 er, forgotten and/or don't know about.
16:47 best to trace the code.
16:49 zanoi i'll look into it then
16:50 so do you know what has to be fixed in order to support all the ogg extensions/mimetypes on all platforms?
16:50 willkg kcw might know more.  he's been playing in that code a lot over the last few months.
16:50 no, i have no idea.  i don't even know what works currently.
16:50 zanoi i'll ask him then
16:50 willkg miro plays oggs just fine.  so "support" in this case is more association-related.
16:51 zanoi well until recently .oga, .ogx didn
16:52 *didn
16:52 *didn't seem to be supported
16:52 willkg when you say "didn't seem to be supported", what does that entail behavior-wise?
16:52 zanoi sorry, can't type, somehow my keyboard layout magically switched
16:54 i didn't test it, but they weren't mentioned anywhere in the code. I guess the renderer we delegate to might still recognize them but they would at the very least be interpreted wrongly in the audio/video sorting
16:54 * willkg nods.
16:55 zanoi anyway, i'll ask kcw
16:56 z3p left #miro-hackers
16:56 zanoi thanks for the help
16:56 * willkg nods.
16:57 zanoi willkg, oh, one more thing. Should I look into handling uninstalling Miro & magnet link association?
16:58 so that we "disassociate" when we uninstall
16:59 willkg i read the status report you wrote, but haven't looked at what you did.  in windows, we associate and disassociate in the installer.  but it looks like you're doing the association for magnet stuff in the code.
16:59 is that right?
16:59 zanoi yes
16:59 z3p joined #miro-hackers
17:00 willkg if someone uninstalls miro, what happens?
17:00 zanoi because together with the magnet.exe there was quite a bit more logic required
17:00 well miro would still be associated, but not installed
17:01 willkg ok.  so that's bad.
17:02 my vote is you talk with z3p who has (to my knowledge) done most of the associate/disassociate work with the windows installer.
17:02 zanoi sure
17:02 willkg i bet he has a good feel for what we've done in the past and what's good/bad going forward.
17:03 and if it turns out that there's a convention that we use, it'd be good to codify it in the wiki.
17:03 z3p: ^^^
17:04 i really like this conversation.  i feel like i'm made of the teflon of "oo--talk to someone else who knows more!".  nothing sticks to me!
17:06 zanoi At least I find out who to talk to :-)
17:07 willkg i'm just being silly.  knowing who to talk to helps a ton.
17:12 zanoi willkg, also, do you have any P1 bugs I should work on?
17:13 or bendk1?
17:15 anyway, gotta go
17:18 arpu joined #miro-hackers
17:19 willkg zanoi: i don't know...  i've been gone for a week so i don't have a good feel of what there is to do.
17:20 zanoi: my suggestion is to read through the P1 bugs assigned to admin and pick ones you want to work on.
17:21 zanoi yeah, will do. Was just wondering if there was anything you'd like me to do.
17:34 kcw zanoi: re:filetype, are you interested in the file's format, or just whether it's audio/video?
17:56 z3p willkg: zanoi: actually, I think nassar did that work, but I'm probably the next best person
17:57 willkg z3p: i had thought you fiddled with it at some point.
17:59 z3p willkg: I've fiddled with some of the installer stuff, but I don't remember doing anything with the file handling
17:59 looks like I missed something while I was at lunch
17:59 * willkg nods.
17:59 z3p zanoi: what's the question?
18:00 willkg z3p: i think zanoi is gone right now.
18:00 z3p willkg: right you are
18:00 willkg z3p: zanoi added some magnet association code to miro.  so we need to do something in the uninstall step so miro doesn't stay associated with magnet things.
18:00 z3p: i'm really fuzzy on the details like what a magnet thing is.
18:01 z3p willkg: it's not an extension though, right?
18:01 willkg i have no idea.
18:04 z3p anyways, it looks like we're doing our file association kinda right, in that it works
18:05 but we don't disassociate ourselves on uninstall, or even remember what the old setting was
18:05 * willkg nods.
18:23 arpu left #miro-hackers
18:27 z3p left #miro-hackers
18:37 mattjohnson left #miro-hackers
18:44 mattjohnson joined #miro-hackers
19:22 morganknutson joined #miro-hackers
19:46 Odysimus joined #miro-hackers
20:24 arpu joined #miro-hackers
20:44 Odysimus left #miro-hackers
21:32 zanoi sorry, had an appointment at other end of town. Back now
21:36 kcw, re: filetype, i'm just interested in having miro behave the right way, and  for ogg that is not to use the filetype to determine mimetype or whether it is audio/video
21:37 kcw, but I don't know what we currently do
21:39 kcw zanoi: we currently use extensions for everything because that seems the most reliable way for most files. We can use the mime type from mutagen for .oggs; there's one issue though - files can be played before they have any metadata in several ways
21:41 if we require a mutagen pass before playing any file, that would solve this and #16405
21:46 zanoi files can be played before they have any metadata in several ways <-- I don't quite understand that sentence
21:46 kcw File -> Open can play files before they're added to the library
21:46 zanoi well what the ogg people told me is that ogg extension isn't reliable
21:47 kcw that's true. it isn't. we don't handle .ogg videos properly right now.
21:47 zanoi right
21:48 but I agree that assumptions from the extensions works most of the time
21:50 but technically it's not the right solution and it will produce wrong results for .ogg, .ogm, .ogx, from time to time
21:51 so I guess we should do as you say, either for 4.0 or later depending on how much work it is.
22:03 kcw, reading the bug report i'm not sure I understand what you intend Miro to do
22:04 do you want to check files as soon as possible with mutagen and use that to determine whether it's audio or video?
22:04 bendk1 left #miro-hackers
22:04 zanoi or do you only want to do a mutagen check right before playing a file, and determine audio/video before that according to extension
22:05 bendk joined #miro-hackers
22:06 kcw when a file is requested to be played, and it doesn't have metadata: wait until extracting the metadata before playing
22:08 (and of course, if the reason it doesn't have metadata yet is that it's queued to be checked but hasn't been reached yet, prioritize it)
22:09 zanoi ok, does that mean there's no need to know whether a file is audio/video earlier than that?
22:10 kcw no, because we need to put things in the Music or Videos tab; this is just how to handle the user playing a file that we don't know anything about yet
22:11 zanoi ah, ok. And for files in the music/video tab we're already making a mutagen check?
22:12 kcw yeah
22:13 zanoi ah ok.
22:20 kcw,  actually I don't think a mutagen check for files we don't know about yet is really necessary. It seems to work fine when testing it.
22:20 although it probably is more technically correct
22:20 even when i have a theora video with a .oga it recognizes that it's a video
22:21 ....as in: it shows the video.
22:22 willkg kcw, zanoi: more importantly, we need to know whether to show the audio renderer or the video renderer.
22:23 z3p joined #miro-hackers
22:23 kcw zanoi: it isn't working in all cases, though. I have a .ogg video that doesn't.
22:23 zanoi kcw, ah ok
22:23 mutagen it is
22:25 kcw willkg: yeah - I think that happens based on which tab it's played from, and would need to be addressed in #16743 (adding note.)
22:27 zanoi kcw, another question: .oga and .ogx weren't listed with the other extensions in filetypes.py, so I added them. Do I need to add them anywhere else to fully support them in Miro?
22:28 or does it not matter because mutagen figures it out anyway?
22:28 kcw zanoi: I think adding them to that map does everything necessary
22:30 willkg kcw: it didn't used to be.  it used to be based on the type in the iteminfo.
22:31 kcw ok, that's probably what it does. the two should always be the same, but as causal relationships go that makes more sense
22:32 bendk left #miro-hackers
22:32 bendk joined #miro-hackers
22:32 * willkg nods.
22:32 willkg bendk: i met ralph bean.  he's awesome!
22:32 bendk hey will
22:32 yeah RJ is great
22:32 willkg hi!
22:33 bendk going to dinner though, otherwise I would go on for hours :)
22:33 willkg heh.  :)
22:34 sorry i haven't gotten much done today.  i've been tending the baby.
22:34 hoping to get more done tonight.
22:35 zanoi willkg, do you mean file_type of ItemInfo?
22:36 willkg zanoi: i forget what it is.  the values are something like video/audio/misc.
22:36 zanoi yes that's it
22:36 willkg zanoi: there are a bunch of things in iteminfo that have terrible names.
22:36 probably my fault.
22:36 zanoi I added a database upgrade that should set the item.file_type to audio if they filename is .oga and video if it's a .ogx
22:36 when I did the changes in filetypes.py
22:37 willkg oh...  don't do that.
22:37 take that out.
22:37 zanoi because that's how other database upgrades did it
22:37 willkg the file_type is set by a bunch of other stuff other than "what's the extension".
22:37 zanoi why?
22:37 willkg so we should assume whatever is in the db is ok now.
22:38 zanoi hmm I'll remove it then, but why do other upgrades do it that way?
22:38 willkg what other upgrades?
22:38 we've had upgrades for 6 years.  so they span a _very_ long period of time.
22:39 6 years might not be the right number.  but they've been around for ages and ages.
22:39 zanoi upgrade 93 for example
22:39 kcw zanoi: that's when file_type was introduced
22:40 willkg ok.  so that's from 2 years ago.
22:40 zanoi so back then you did everything file_type on extension but not anymore?
22:40 willkg please back it out asap.
22:40 then we can chat about it.
22:40 right now you're hosing anyone who does a git fetch and then a db upgrade.
22:41 i'm really glad you said something about it.  otherwise i wouldn't have known until i looked at changesets.
22:41 zanoi who has a .ogx that is an audio or a .oga that is a video.....
22:41 willkg please back it out asap.
22:41 zanoi sure
22:41 willkg then we can chat about it.
22:43 zanoi fixed
22:45 willkg we've been refactoring large parts of miro over the last few years to fix lots of problems and also to get to a better architecture.
22:45 ajonas left #miro-hackers
22:45 willkg the iteminfo stuff is new as of the ui overhaul that bendk did.
22:46 honestly, i don't remember how miro used to work in a lot of cases.
22:46 and in a lot of cases, that's probably a good thing.
22:46 i try not to do things that make me cry.  ;)
22:46 i have to go clean the kitchen...  afk a few.
22:47 zanoi so according to the ogg guys .ogx isn't in practical use, so the only risk would be that someone renamed his movie to .oga (ogg audio) but that wouldn't make sense either
22:47 willkg zanoi: if you're curious, it's probably worth doing a git blame, finding the checkin that upgrade93 was a part of, and looking at checkins nearby.
22:47 zanoi you might want to remove the comment in filetypes.py that says one should write a databaseupgrade when changing the map
22:49 or do you want me to remove it?
23:21 willkg zanoi: i don't know.  probably best to find out who wrote it and why.
23:22 zanoi seems like ben wrote it, and i'm pretty sure I know why.
23:23 it says: NOTE: if you change VIDEO_EXTENSIONS or AUDIO_EXTENSIONS, consider writing a  database update so that the file_type attribute of the item table gets fixed
23:23 so it's quite clear. But it seems like you don't want people to do it
23:23 willkg ...
23:23 zanoi but I can ask bendk
23:23 willkg ok, so it was written in 2009.
23:23 zanoi does that mean i should remove it?
23:24 willkg like i said before, it's best to find out why he wrote it.  amongst other things, you might find out if that should be removed because it's not true anymore.
23:24 i'm pretty sure it's bad, but bendk would know for sure.
23:24 zanoi bendk, ^

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