**** BEGIN LOGGING AT Fri Aug 18 04:33:51 2006 Aug 18 04:33:51 * Now talking on #gsv Aug 18 04:33:52 ok Aug 18 04:33:54 hi there Aug 18 04:33:55 soero non ci sia troppa coda Aug 18 04:34:01 hey muntyan Aug 18 04:34:03 ji guys Aug 18 04:34:04 hi Aug 18 04:34:12 muntyan: how are you? Aug 18 04:34:38 well, going to sleep actually Aug 18 04:34:43 was going to :) Aug 18 04:35:02 ops, we can meet later if you prefer Aug 18 04:35:37 no no Aug 18 04:35:43 better now Aug 18 04:35:47 ok Aug 18 04:35:57 (just not for four hours like last time :) Aug 18 04:36:13 hehe, ok Aug 18 04:36:35 I have just read you mail Aug 18 04:36:52 lemme see if I have a clear vision of the current state Aug 18 04:37:37 1. Syntax highlighting is working quite well but needs some tuning Aug 18 04:38:07 2. Before doing this it is better to try to test the code in some real condition Aug 18 04:39:24 3. For some reason you think we still have to wait for the printing stuff (can you explain me why?) Aug 18 04:39:39 4. You want to discuss about the style stuff and the undo manager Aug 18 04:40:10 5. You are worried about the interaction with the file chooser Aug 18 04:40:28 6. You have some question for me Aug 18 04:40:33 it should be all Aug 18 04:40:34 right? Aug 18 04:40:43 okay, let me go through these Aug 18 04:40:49 perfect Aug 18 04:41:35 1. Yes. it needs lot of tunings but hopefully really small tunings, i.e. normal bugs and small changes, i.e. not much in total Aug 18 04:42:23 2. absolutely. i'm just afraid to even say that something is fine (i do hope it's fine) Aug 18 04:44:54 sorry, slow typing Aug 18 04:45:28 no problem Aug 18 04:45:41 3. not sure we really need to wait, but i am frustrated enough with printing at the moment (that list of bugs), and it doesn't look like gtksourceview needs gtk printing right now Aug 18 04:46:12 i would really love to move printing to gtksourceview though Aug 18 04:46:34 if you say it makes sense or needed, i'll do it Aug 18 04:46:40 hmmm... are you frustrated by gtk+ printing stuff or what? Aug 18 04:46:51 yes, with gtk priniting stuff Aug 18 04:47:04 i am really frustrated with those bugs which don't get fixed Aug 18 04:47:09 a simple example: Aug 18 04:47:16 you can't cancel preview Aug 18 04:47:31 okay, i go to the code and going to write a cool patch Aug 18 04:48:09 but i can't, because it's made in some strange way, preview is some interface, and i can't understand what's going on in there, and i have other things to do, and so on Aug 18 04:48:13 another thing Aug 18 04:48:40 i thought gtk printing is about multi-platformness, but there's no win32 release yet Aug 18 04:48:42 I see Aug 18 04:48:47 i.e. i hate gtk-2.10 :) Aug 18 04:48:52 gtk+ printing stuff have serious problem Aug 18 04:49:04 hehe... ok Aug 18 04:49:32 and, e.g. gedit has *better* printing with gnomeprint Aug 18 04:49:44 so it really doesn't make sense to rush with gtk printing Aug 18 04:49:45 I think we can wait for printing (even if we will have to work on it sooner or later) Aug 18 04:49:59 of course i don't mind to share my work on it :) Aug 18 04:49:59 I agree Aug 18 04:50:10 okay, let me go farther Aug 18 04:52:29 4. one needs to say exactly what he wants from these two guys. not the api, but what's needed. Aug 18 04:52:46 4a. the styles thing - do you want live styles scheme editing, or it's fine to create a scheme file once, and then load it. Aug 18 04:53:34 4b. this one is harder: for example: you do replace-all, you don't want cursor to jump to last replaced word, it needs some interactive-non-interactive thing Aug 18 04:54:00 i.e. these two need special discussion dedicated to them Aug 18 04:54:26 normal work, but needs attention and can't be "just done" Aug 18 04:54:37 5. Aug 18 04:54:41 Yes, it's a bad bugf Aug 18 04:54:43 bug Aug 18 04:54:59 while it's a rare bug, but i do not think it's acceptable Aug 18 04:55:25 user can't cancel highlighting and can't open a file, it's really really lame Aug 18 04:55:34 About 5, please note that the async file chooser is going to be retired Aug 18 04:55:41 paolo: no it's not Aug 18 04:55:46 paolo: check gtk-devel :) Aug 18 04:55:58 oh Aug 18 04:56:03 * paolo reads the last mails Aug 18 04:56:07 and finally, 6: Aug 18 04:56:19 please please look at that highlighting thing Aug 18 04:56:35 i stuffed lot of lines of code with noone looked at it Aug 18 04:56:42 it needs eyes! Aug 18 04:57:44 ok Aug 18 04:58:14 um, 7. Aug 18 04:58:25 the important thing Aug 18 05:00:21 soc is going to be over, but not my gtksourceview thing. while i am not going to be a greatest gedit contributor, i do want to work on gtksourceview *at least* for things we are talking about Aug 18 05:00:37 e.g. i am not going to stop syntax highlighting after august, 21 :) Aug 18 05:01:03 this is great ;) Aug 18 05:01:11 um, not sure i should have said it, but i have. now time to smoke a bit :) Aug 18 05:01:20 Then, let us sketch some action plan Aug 18 05:02:11 - I will spend next days looking at the current code and thinking about the printing stuff Aug 18 05:02:38 - You will prepare a proposal about the undo manager stuff Aug 18 05:03:03 - We will meet to discuss about the requisites of the styles stuff Aug 18 05:05:23 1) ok Aug 18 05:06:31 2) okay, it will be different a bit from what i have now ("groups" needs to be better specified, to allow "Undo This Thing" in menus and such) Aug 18 05:06:39 * pbor returns... mission failed :( 10 people before me and the doctor didn't even show up yet! Aug 18 05:06:40 3) ok Aug 18 05:06:57 pbor: sounds like soviet union :) Aug 18 05:07:18 sounds like typical italy, even more so in august :) Aug 18 05:07:26 a note about printing stuff: Aug 18 05:08:15 i feel it must subclass GktPrintOperation, to allow connecting to its signals, the progress and cancel is the main thing Aug 18 05:09:39 Aug 18 05:10:14 pbor: i knew it was true that italians are good people :) Aug 18 05:12:51 muntyan: about 2) I'd like to have the possibility to add custom actions Aug 18 05:13:14 see also the undo manager related bugs in bugzilla Aug 18 05:14:20 well, the undo manager itself should not care (it does not) about actions at all Aug 18 05:14:37 for example I'd like to be able to add a custom action for "replace all" (so that I can store the related info in some more compact way, i.e. not as a sequence of insert-delete) Aug 18 05:14:53 it Aug 18 05:15:03 's perfectly fine Aug 18 05:15:10 ATM. it does not expose such details Aug 18 05:15:20 question is how to deal with this stuff in the buffer Aug 18 05:15:38 but it is probably worth exposing the undo manager as a public object Aug 18 05:15:48 paolo: well, i mean the new and great undo manager thing :) Aug 18 05:16:00 where users tell it to store actions Aug 18 05:16:07 okay, i' Aug 18 05:16:12 crappy enter key Aug 18 05:16:21 I have not seen the new undo manager Aug 18 05:16:29 so I don't know how it works Aug 18 05:16:30 i'll try to present better undo manager proposal, i guess Aug 18 05:16:41 ok, cool Aug 18 05:16:49 my only requirements are: Aug 18 05:17:01 - should have all the features of the current one Aug 18 05:17:08 - should support custom actions Aug 18 05:17:41 - should undo/redo markers (see bugzilla) Aug 18 05:18:00 well, that's part of custom actions :) Aug 18 05:18:30 along the same lines it should allow undo of tag-applying Aug 18 05:18:30 I'm not sure it is a custom action Aug 18 05:18:55 - should implements the features written on wiki Aug 18 05:19:40 "Store the list of actions on disk???" - crap Aug 18 05:19:59 it can wait, and may be we don't need it Aug 18 05:20:02 the rest looks fine Aug 18 05:20:25 (didn't look at gnumeric) Aug 18 05:20:42 but it would be very nice to have a way to serialize/de-serialize undo lists (for advance session management) Aug 18 05:21:04 but it is not a requirement now Aug 18 05:21:07 uhm... doesn't sound like a good idea to me Aug 18 05:21:08 well, i don't think it would be difficult, but does it make sense? Aug 18 05:21:24 you change a file, or better delete it, and what then? Aug 18 05:21:38 what if the file and the undo file go out of sync Aug 18 05:21:45 exactly Aug 18 05:21:52 the list will be invalidated Aug 18 05:22:13 but how can you know when to invalidate? md5 checksums? Aug 18 05:22:13 but what I mean is that we should expose an API to get the list Aug 18 05:22:24 well, that's better :) Aug 18 05:22:40 but you need sha1 or md5 or... we'll end up reinventing bzr/git etc :) Aug 18 05:22:54 then super-cool-text-editor-that-wants-to-be-like-emacs will use it as preferred Aug 18 05:23:05 i just object to making it make sense between different processes Aug 18 05:23:28 err, even different buffers Aug 18 05:23:33 BTW, as I said it is not a requirement for this first step Aug 18 05:23:56 well, as i said, the rest looks fine :) Aug 18 05:24:16 good Aug 18 05:24:25 I'd say you can go to sleep :) Aug 18 05:24:31 hehe Aug 18 05:24:47 * muntyan goes to sleep Aug 18 05:24:53 so, see you guys Aug 18 05:24:58 see you Aug 18 05:25:03 later Aug 18 05:25:13 i'm often here in the late night - early morning for you, so we can talk Aug 18 05:25:14 :) Aug 18 05:25:23 paolo: what do you think to add support for gsv2 in gedit cvs? (#ifdeffed and disabled, obviously) Aug 18 05:25:24 good night or whatever Aug 18 05:25:48 dunno Aug 18 05:26:25 I think we can Aug 18 05:26:29 - wait for RC1 Aug 18 05:26:35 - branch gedit and gtksourceview Aug 18 05:26:43 - working on HEAD for both Aug 18 05:27:50 yeah, I don't care in which branch it goes: my point was about having support for both gsv1 and 2 Aug 18 05:28:30 to allow normal use to casual jhbuilders but also allow easy testing for those interested Aug 18 05:28:50 without the hassle to keep two branches in sync Aug 18 05:31:07 um, while i'm here Aug 18 05:31:16 i have a cool plan for merge Aug 18 05:31:35 create a new branch, and commit bunch of patches to it Aug 18 05:31:56 one patch is to the contextengine, the rest as they are Aug 18 05:32:10 since the engine change is one big change anyway Aug 18 05:32:20 why not just join your branch? Aug 18 05:32:26 it will simplify merge Aug 18 05:32:32 changes on HEAD were minimal Aug 18 05:32:49 pbor: for example, because i removed printing :( Aug 18 05:33:07 now i think CVS is not as flexible for that Aug 18 05:33:24 ok, then if you go for 'collection of patches' I'd say to just commit them on HEAD after we branch Aug 18 05:33:32 maybe it's crap though, i never did such things with CVS Aug 18 05:33:48 pbor: right Aug 18 05:34:03 err, not right Aug 18 05:34:21 if it gets approved as "do it", then right Aug 18 05:34:32 but i'm not sure about it Aug 18 05:34:33 of course :) Aug 18 05:34:57 well, it all goes down to CVS details Aug 18 05:35:12 we'll see when it's time, I am not that concerned Aug 18 05:35:35 maybe we'll even have svn at some point Aug 18 05:35:39 * pbor dreams on Aug 18 05:35:44 i just wanted to say that the history in soc-2006 branch should not necessarily be preserved in HEAD Aug 18 05:35:53 yep, agreed Aug 18 05:36:25 okay, now i'm off. good night, etc. Aug 18 05:36:40 'night Aug 18 05:37:37 okay, so CVS can wait, i guess :) Aug 18 05:37:47 you have release soon, and everything Aug 18 05:37:58 then we can attempt hard surgery Aug 18 05:38:11 ok Aug 18 05:38:16 okay, i'm really off now :) Aug 18 05:38:20 :)