[08:25] mhr3: that crash you pointed out in the tree-index branch is nasty... [08:26] kamstrup, which one? [08:26] clear or the other one? [08:26] the priv->reader_* being unset one [08:26] ah [08:26] the only way I can find to fix it is to include the reader fields in a gobject property [08:26] yea, i immediately didn't like the constructor after seeing the magic after g_object_new [08:27] and I don't want 3 props because that'll introduce all sorts of odd corner casesw [08:27] so ideally one prop [08:27] right, so make a construct (write only) gobject property and use set_data on it [08:27] but I don't want to have a public struct to hold that data either [08:28] setdata() - how'd you mean? [08:28] I was thinking just a construct only prop, that we hide in GI [08:28] standard g_object_set_data [08:28] set with what data? [08:28] :-) [08:28] and use it to hide the three props [08:29] or well... pass along [08:29] sorry, I don't how that can work [08:29] ok, here we go [08:29] we can't call set_data() until after we have the object instantiated and construct() called [08:30] ..index_new { var dummy = object_new(G_TYPE_OBJECT); dummy.set_data("reader-...", ...); ...; var index = object_new(TYPE_TREE_INDEX, "dummy", dummy, ...); } [08:31] i could have used enter :) [08:31] ah - you want to call set_data() on the *model* :-) [08:31] no, no, just some dummy [08:31] but model could work as well, actually [08:31] ah [08:32] I'd just go with a gpointer[3]... [08:32] also a way [08:32] ugly no matter what :-) [08:32] right :) [08:33] kamstrup, but making it a real prop would ease subclassing [08:34] mhr3: yeah, that's a good point [08:34] mhr3: since I have a DeeScanIndex lined up already [08:34] and HashIndex has the same problem i guess [08:34] indeed [08:50] kamstrup, btw i think the *_reader methods will need (skip) annotation, they won't be usable from python nor vala [08:51] I think I already tested them working in Python [08:51] no way [08:51] how did you pass the column in? [08:51] mhr3: but I think I'll refactor that part anyway [08:52] mhr3: just 0 I believe [08:52] probably working by chance because it's 32 bit... === MacSlow is now known as MacSlow|lunch === nuthinking_ is now known as nuthinking === yofel_ is now known as yofel [13:56] kamstrup, mhr3 ping [13:57] hey davidcalle [13:57] mhr3, when is libunity API supposed to break? [13:58] when we're done with all the changes, we dont want multiple breaks [13:59] Ok === MacSlow|lunch is now known as MacSlow [14:22] davidcalle: yeah, stuff is slowly starting to trickle into bzr as we speak [14:23] but as mhr3 said we aim for *one* break early in the cycle [14:23] and it'll only be a small break, so I don't expect any major refactoring needed [14:23] hmmm [14:23] :) [14:24] kamstrup, mhr3, I'm ok with refactoring as long as it means features :) [14:25] davidcalle: there will be features, faster methods, and an API designed to minimize the number of potential client side bugs [14:25] kamstrup, while we're at that do you any way to do the filtering/sorting of results from multiple scopes [14:25] so I hope you'll love it :-) [14:26] kamstrup, davidcalle can explain, although i think you've talked about it [14:26] mhr3: njpatel did all that, and I never really looked at it :-) [14:27] kamstrup, I'm sure I will. [14:28] kamstrup, I'm looking for a way to sort results from multiple scopes. I was thinking of something like mutliple scopes talking to the same dee model, and dee doing the sorting. [14:29] Use case being : someone wants to add a scope to the file lens (like a new datasource for games) and want the results to integrate nicely the existing game results. [14:30] hi all [14:30] the application lens* [14:30] davidcalle: right - there are a few subtleties here [14:30] davidcalle: I think it is very important that the results view don't "jump around" as slower results trickle in [14:31] of course, the one doing the merging could be clever enough to avoid that [14:31] kamstrup, I used to think the same thing, but I'm not so sure. [14:32] hehe, I am :-) [14:32] ok, there may be lenses that are special in this regard [14:32] but lenses like files and apps can not have that behaviour imho [14:33] kamstrup, I mean, either we have non-sorted results from multiple sources, either we have slow results from multiple sources. [14:33] fx a "Quake3 game servers lens" that lists servers sorted according to some clever combination of latency and #users ... :-) [14:34] So I guess the problems start when you want to merge results from a scope into an already populated category in the lens [14:40] kamstrup, indeed. With scopes with different latency. I have stuff appearing at top or bottom in an unreliable way. [14:44] btw, there is a bug I need to file: In a category : results from scope A appear, results from scope B appear, results from scope A jump after results from B. [14:45] But I believe it's related, somehow, to the way Unity loads icons. If I disable online thumbnail fetching for scope A, it doesn't happen. [15:04] anyone of you can tell me which one is the right XChat branch in Launchpad? There are many... I've prepared a very small fix and I don't know which one to branch from and which one to push into... I tried yesterday but probably I pushed to the wrong one and I got a lot of diff errors. Please note that my "fix" is only related to Ubuntu and must not be included in the upstream version of it. Thanks [15:16] and another question: which one is the preferred "mono" font to use in Ubuntu: "Ubuntu Mono" or "Monospace" ? thanks [15:32] Andy80, Ubuntu Mono is now default in Gnome Terminal. [15:36] davidcalle: my intention was modifing XChat to have the same font as default, to maintain the same UI experience, and I already modified the proper file... but now I've found another branch that already has "Monospace 9" as default font.... so I don't know what to do exactly :) [15:39] Andy80, you could ask Sladen in #ubuntu-design , he is the one dealing with the Ubuntu font integration. [15:40] davidcalle: ok, I ask him, thank you :) [15:40] Andy80, yw :) === davidcalle_ is now known as davidcalle [17:15] dbarth, one question, is lamalex on holidays? [17:15] btw, hi [17:15] ;) [17:53] API: Thanksgiving in the US, I'd say lamalex is probably off today :) [17:56] gema, well also asking because I was looking lamalex some days ago [17:56] but it is true that today is a bad day [17:56] to look after him ;) [17:57] gema, thanks, I forgot that USA turkey-related party ;) [17:57] API: yep, almost time for turkey here in the uk too! [17:59] API: where are you from out of curiosity? [18:05] gema, Spain, I didn't know about that uk Thanksgiving, I thought that it was just USA and Canada celebration [18:05] API: I meant Christmas , we don't do thanksgiving, I am also from Spain, encantada :D [18:06] gema, ah ok === JanC_ is now known as JanC === seiflotfy is now known as HuNannyYumYum === HuNannyYumYum is now known as seiflotfy