All times shown according to UTC.
| Time | Nick | Message |
|---|---|---|
| 00:08 | CarlFK joined #miro-hackers | |
| 00:19 | z3p joined #miro-hackers | |
| 00:24 | z3p | melinath: hey, I'm back |
| 00:25 | melinath: 1) we only reverse the video iter, so it's not relevant for searches | |
| 00:25 | er, feed iter | |
| 00:25 | melinath: we have a test that says it should be that way, hence the code | |
| 00:29 | melinath: re: next steps, work on moving feed/search imports into celery | |
| 00:29 | melinath: unless you've got something else | |
| 00:56 | melinath | z3p: in terms of next steps: sounds good to me |
| 00:58 | z3p: In terms of feed imports: people can take a search page, get a feed of it, and tell us it's a feed. Or more generally, we can't control the ordering of the feeds we need to process. | |
| 00:59 | z3p: We have a test that says it's reversed. I think that we should get rid of that test. Possibly replace it with one that makes sure it isn't reversed. | |
| 01:00 | z3p: Unless there's a really really good reason to use reversal. | |
| 01:17 | z3p joined #miro-hackers | |
| 01:27 | z3p | melinath: the existing behavior is that the default feed import puts the videos in 'feed' order, hence the test; I think that's the behavior people expect |
| 01:36 | trask | hey z3p |
| 01:36 | your last clue helped a lot | |
| 01:36 | found map of the python files on my system | |
| 01:42 | z3p | trask: hrm? |
| 01:43 | trask | in another channel, #miro |
| 01:43 | still havent found head or tails in the code, but thats gonna happen soon | |
| 03:28 | melinath | z3p: The thing is that it doesn't put them in feed order. It sets the submit dates to be in feed order, which is unexpected. |
| 03:29 | z3p: I think it makes sense to keep them in feed order. I just think we should actually track what feed order is. | |
| 03:29 | z3p: That being said, we don't have a way to do that right now. | |
| 03:29 | z3p: Luckily, it will be possible once we have that FeedImport class. | |
| 03:30 | z3p: In the meantime, I don't really care one way or the other about reversal, as long as we get rid of it eventually. | |
| 05:02 | uwe_mobile joined #miro-hackers | |
| 06:34 | trask | if i want to edit what gets displayed in the sidebar, wich file should i start looking at? /usr/share/pyshared/miro/folder.py seems to build them, but is it here i should hash out stuff or put in if statements? got to be some other place it gets assambled |
| 06:47 | maggie_s joined #miro-hackers | |
| 07:07 | Odysimus joined #miro-hackers | |
| 07:54 | Jarrhead joined #miro-hackers | |
| 09:19 | CarlFK joined #miro-hackers | |
| 10:48 | janet joined #miro-hackers | |
| 10:56 | janet1 joined #miro-hackers | |
| 12:34 | CarlFK joined #miro-hackers | |
| 12:49 | z3p | melinath: I don't see how FeedImport keeps track of what 'feed order' is |
| 12:50 | melinath: reversed() does, because the IDs are in the right order | |
| 12:50 | willkg joined #miro-hackers | |
| 12:51 | z3p | really, it's only a big issue for crawling large feeds, and we should limit that to the total # of videos available on the site anyways |
| 13:40 | arpu joined #miro-hackers | |
| 13:50 | willkg | hi! |
| 13:57 | paroneayea joined #miro-hackers | |
| 14:46 | Jarrhead joined #miro-hackers | |
| 15:33 | GitHub93 joined #miro-hackers | |
| 15:33 | GitHub93 | [miro] bendk pushed 2 new commits to master: https://github.com/pculture/mi[…]c3be7e4...fdc3027 |
| 15:33 | [miro/master] bz18375: Don't check Item.downloader attribute directly. - Geoffrey Lee | |
| 15:33 | [miro/master] Merge pull request #53 from geoffl/bz18375 - bendk | |
| 15:33 | GitHub93 left #miro-hackers | |
| 15:53 | thrillERboy joined #miro-hackers | |
| 15:54 | thrillERboy left #miro-hackers | |
| 16:34 | willkg left #miro-hackers | |
| 16:46 | Jarrhead joined #miro-hackers | |
| 17:26 | arpu joined #miro-hackers | |
| 18:23 | CarlFK joined #miro-hackers | |
| 19:00 | janet1 joined #miro-hackers | |
| 19:32 | janet2 joined #miro-hackers | |
| 19:36 | janet1 joined #miro-hackers | |
| 19:48 | paroneayea joined #miro-hackers | |
| 19:55 | melinath | z3p: tracking what order things were added by relying on pk order or on submit date order won't work in an asynchronous setting. |
| 19:56 | z3p: If a video has a relationship to a feed import, it can also have a field to track what its place in that import was. | |
| 19:58 | z3p: Even importing small feeds (say, 300 videos) is slow right now. | |
| 20:47 | maggie_s joined #miro-hackers | |
| 21:36 | z3p joined #miro-hackers | |
| 21:42 | z3p | melinath: sure, but that doesn't change the fact that we need to do the imports in reverse order; that's just using the FeedImport id instead of the primary key for ordering |
| 21:50 | mick_laptop joined #miro-hackers | |
| 22:05 | Odysimus joined #miro-hackers | |
| 22:06 | paroneay` joined #miro-hackers | |
| 22:17 | z3p | melinath: actually, nevermind; I figured out a way to have our cake & eat it too |
| 22:17 | melinath: but it'll have to wait for Tuesday | |
| 22:43 | melinath | z3p: The idea would be to have an additional field for the order. For example, Video.import for the import and Video.import_order for its position in that import. |
| 22:43 | z3p: not to use the feedimport id. | |
| 22:46 | z3p: I'm interested in the solution you've found, but I suppose I can wait until tuesday. |