All times shown according to UTC.

Time Nick Message
00:40 Odysimus joined #miro-hackers
02:17 CarlFK left #miro-hackers
02:35 bendk left #miro-hackers
04:12 CarlFK joined #miro-hackers
04:30 JasonWoof joined #miro-hackers
04:36 ajonas joined #miro-hackers
05:03 DGMurdockIII joined #miro-hackers
05:19 JasonWoof left #miro-hackers
05:22 CarlFK left #miro-hackers
05:43 ajonas left #miro-hackers
05:44 ajonas joined #miro-hackers
05:51 odenthaa joined #miro-hackers
05:52 CarlFK joined #miro-hackers
05:53 Odysimus left #miro-hackers
06:04 ajonas left #miro-hackers
07:57 DGMurdockIII left #miro-hackers
08:38 odenthaa left #miro-hackers
08:42 Odysimus joined #miro-hackers
08:44 Odysimus left #miro-hackers
12:42 uwe_ left #miro-hackers
12:52 arpu joined #miro-hackers
12:57 uwe_ joined #miro-hackers
14:57 z3p good morning
15:16 ajonas joined #miro-hackers
16:15 DGMurdockIII joined #miro-hackers
16:27 bendk joined #miro-hackers
16:32 bendk z3p: hey, you around?
16:42 z3p bendk: yep
16:43 bendk I just read your email about DeviceItem and I agree, but I think it means that we have to expand the scope of DataSource
16:43 z3p bendk: on second thought, I think that the search code can easily be done in terms of the datasource API
16:43 bendk: you just have one ItemSearcher per DataSource
16:44 bendk the issue is that I was thinking of DataSource as more like a view of the data, rather than the entire thing.  For DeviceItem, I guess it's basically the same thing, but for SQLite, it's slightly different
16:44 z3p bendk: ?
16:45 bendk I guess the distiction makes more sense for regular Item objects
16:45 does DataSource contain all Items?  Or just the Items "inside" a single tab?
16:48 maybe, the same API could work for both...  I'm trying to wrap my head around that now
16:48 z3p bendk: ah
16:52 bendk: I guess the question is which part of the code is the DataSource being inserted between
16:52 bendk yeah, my original idea was that it would go just below View/ViewTracker
16:54 z3p bendk: also, there are two classes called ViewTracker
16:54 bendk wow, that's not good
16:54 I meant the database ViewTracker
16:54 z3p bendk: tjere
16:54 bendk I think I named both too :)
16:55 z3p *there's also messagehandler:ViewTracker
16:57 bendk: so there's already an API that's close to the DataSource (maybe it should be ItemSource, for clarity?) API, ViewObjectFetcher
16:58 bendk yup, I just added that.  I think DataSource would replace that class and add a little more functionality
16:59 more importantly: a little more flexibility
16:59 yeah, ItemSource makes sense, since we are only talking about items
17:00 I guess we might want to extend it to feeds as well, but at this point, I don't see why
17:00 z3p maybe later, but that's not a current use case
17:01 so for search, the way I see it working is that instead of the current ItemSearcher class, we have an ItemSearcherManager, which searches over multiple DataSources
17:02 since the only item_info_cache part of ItemSearcher is when it's initialized with the set of items
17:03 er, ItemSources
17:05 or you could even have that be part of whatever class is managing the current ItemSources
17:11 bendk sounds good to me, here's what I'm thinking
17:11 we create a DataSource object for all Items, DeviceItems, and DAAPItems (not sure about the 3rd name)
17:12 z3p bendk: I think that's right, but glee is the person who'd deal with that
17:12 bendk so there's one thing that I'm still a bit confused about
17:12 we can have multiple devices connected right?
17:12 so would the DataSource cover all DeviceItems, or would we create a DataSource for each device?
17:13 seems like the latter to me, but I'm not sure
17:13 z3p the latter, yeah
17:14 and we'll just remove the ItemSource when the device is removed
17:14 bendk sure
17:15 so maybe we should have an ItemSourceManager class, that just holds all the current ItemSource objects?
17:16 z3p bendk: yeah, that sounds right
17:16 bendk Anyways, we also have code to create a DataSource for just a subset of Item objects, but those data sources don't get passed to ItemSearcher
17:17 z3p bendk: right, we just create a datasource from the View
17:17 bendk ItemSource, sorry :)
17:18 I was thinking the other way, we create a View from an ItemSource
17:18 then View would contain the common code to do things like search
17:19 that's in a class hierarchy sense
17:19 z3p bendk: hmm; but right now the view is doing things like joining tables and filtering in the DB
17:19 bendk in another sense, yeah, we would rip out that code from View and put it into an ItemSource subclass
17:20 z3p okay, so the hierarchy looks like [device/DB/item_info_cache] < ItemSource < View/ViewTracker/ItemSearcher
17:21 bendk yup
17:21 z3p and ItemSource returns ItemInfo objects, not Items
17:21 bendk I think so, does that make sense to you?
17:22 z3p I think the only other thing is that we change the ItemInfo class so that instead of taking an Item and doing stuff to it, it takes a dictionary, or kwargs, or something else
17:23 and then the responsibility for how to make an ItemInfo lives in the ItemSource class
17:24 bendk I agree with the part that ItemSource generates ItemInfo objects
17:25 I still think it's nice for ItemInfo to be able to input an Item though
17:27 what about we add 3 class methods, one creates itemInfo from Item, one from DeviceItem, and eventually one from DAAPItem
17:28 z3p bendk: I refactored it a bit like that earlier today (see http://git.participatorycultur[…]828bbc7121cd7baa)
17:28 bendk: and that works for now, but not for arbitrary ItemSources
17:28 bendk: because ItemInfo needs to know about all the sources
17:28 instead of just being a container
17:29 DGMurdockIII hay has there been anymore work on converting videos to work on zune HD
17:30 z3p DGMurdockIII: you should ping the bug for it
17:30 bendk okay, you've convinced me, let's make ItemInfo dumber and put the logic in Item/DeviceItem
17:31 z3p bendk: we should basically just be able to copy ItemInfo._from_item() into DataSource
17:31 DBItemSource
17:31 bendk sounds good to me
17:33 z3p bendk: okay; I'll start hacking on it then
17:34 bendk hacking to create ItemSource?
17:35 z3p bendk: yeah
17:35 bendk awesome
18:09 willkg joined #miro-hackers
18:09 willkg hi!
18:15 DGMurdockIII (z3p): just let me find it i map one
18:16 (z3p): i made one a wile ago
18:16 (z3p): i made one a while
18:16 z3p DGMurdockIII: yeah, it's in bugzilla if you search for 'zune'
18:20 DGMurdockIII http://bugzilla.pculture.org/s[…]_bug.cgi?id=14999
18:20 do you think i put enough info on the report
18:24 willkg DGMurdockIII: you should add another comment on that bug asking if anyone is doing anything with it.  i don't know what the MVC dev cycle is like or whether they're working on a new version now or not.
18:24 DGMurdockIII: but it's worth adding a comment to see if anyone is paying attention.
18:25 DGMurdockIII i now its built in to miro now
18:26 willkg no...  Miro Video Converter is a different product.  Miro also does conversions.  if you want to add Zune HD conversion support to Miro, then that bug is against the wrong product.
18:26 does that make sense?
18:27 z3p: sorry about that last peer review--i missed that last comment in the checkin.
18:28 z3p willkg: no worries; I made the other change you talked about, although after further discussion with bendk it's changing again :)
18:29 DGMurdockIII oh ok
18:29 willkg z3p: good.  i think the current situation is iffy because it doesn't easily allow for other item types.
18:29 afk a bit--baby crying.
18:29 z3p willkg: yes, you are correct
18:29 willkg: that is changing :)
18:30 DGMurdockIII i thnout that they stoped supported Miro Video Converter becouse the function was added to miro but that good to hear
18:32 it would be great if Miro Video Converter is still being worked on becse most of the video i need converted i dont get art ones i get from miro
18:40 glee joined #miro-hackers
18:50 JasonWoof joined #miro-hackers
18:50 JasonWoof left #miro-hackers
18:50 JasonWoof joined #miro-hackers
19:01 willkg z3p, bendk: i think i just bumped into a problem where changes were made to ItemInfo and i've got a stale/bad iteminfo cache.  how do i wipe my cache?
19:02 z3p willkg: what's happening?
19:02 willkg z3p: i have some items i downloaded ages ago.  i think you added a "device" attribute to the iteminfo in november.  so now when i go to delete items, i'm getting a traceback that the ItemInfo object has no "device" attribute.
19:03 z3p: i can delete newly downloaded items--just not items i've had for a while.
19:03 z3p willkg: huh; they should have gotten updated when ItemInfo.VERSION got updated
19:03 willkg z3p: i have vague memories of ben saying we had to wipe our iteminfo cache manually if it changes.
19:03 i'm a developer, so it's entirely possible that it didn't update right because i was in some funky state.
19:04 z3p willkg: if you change ItemInfo.VERSION, your cache should get dropped
19:05 willkg: or you can manually erase/drop the table
19:05 willkg z3p: i'm getting this: http://pastebin.com/f7hR4tLv
19:05 oh...  hrm.  seems to happen on newly downloaded items, too.  now i'm confused as to what i did earlier--i thought i successfully deleted an item.
19:07 z3p willkg: oh, I thing I broke it
19:07 willkg oh, hrm.  now that i'm looking at the iteminfo code, i think this is a bug.  it looks like it's entirely possible for an iteminfo to _not_ have a device attribute.
19:07 er, property.
19:07 so that "if iteminfo.device" should really be a hasattr or something like that.
19:07 z3p willkg: if you add self.device = None to _from_item() somewhere
19:08 willkg: or you could flip them all to hasattrs
19:08 willkg this is one of the things that's getting fixed when you re-do your device item stuff, right?
19:08 z3p but the former is probably easier
19:08 willkg: yeah; my working tree isn't in a great place to fix this right now :/
19:08 willkg that's fine.
19:09 i'll check in a hasattr fix so we don't get bombarded with bug reports from qa people deleting items.
19:09 or, actually, i'll just do what you suggested.
19:11 ok. all fixed and pushed.
19:14 kcw joined #miro-hackers
19:14 z3p willkg: thanks; sorry about breaking it
19:14 willkg no worries.  :)
19:28 Odysimus joined #miro-hackers
19:28 Odysimus left #miro-hackers
19:31 arpu left #miro-hackers
19:34 Odysimus joined #miro-hackers
19:35 odenthaa joined #miro-hackers
19:35 z3p bendk: is the limiter stuff used for anything other than items?
19:35 bendk: like, are feeds being limited?
19:50 nevermind, they're not
19:50 I'm going to move that support out of ViewTracker and into something else
19:58 mattjohnson joined #miro-hackers
20:03 kcw left #miro-hackers
20:04 kcw joined #miro-hackers
20:18 kcw left #miro-hackers
20:29 willkg mmm...  that's the second time I've been asked how we migrated from Trac to Mediawiki.  i should post the process and scripts I used.
20:29 glee: you on?
20:49 JasonWoof left #miro-hackers
21:00 DGMurdockIII left #miro-hackers
21:11 JasonWoof joined #miro-hackers
21:11 JasonWoof left #miro-hackers
21:11 JasonWoof joined #miro-hackers
22:10 Odysimus left #miro-hackers
22:12 Odysimus joined #miro-hackers
22:18 bendk z3p: hey, sorry I was away for a while
22:18 I think we need the limiter code in ViewTracker because that's how search is handled
22:19 that's basically the only piece of functionality that View/ViewTracker has that wouldn't be in ItemSource
22:40 jamessan` joined #miro-hackers
22:41 jamessan left #miro-hackers
22:42 willkg afk a bit....
23:01 DGMurdockIII joined #miro-hackers
23:26 JasonWoof left #miro-hackers

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