[00:26:38] *** lpereira has quit IRC
[07:19:17] *** wks has joined #eclipse-e4
[08:56:58] *** tschindl has joined #eclipse-e4
[09:47:32] *** tschindl has quit IRC
[10:44:21] *** tschindl has joined #eclipse-e4
[11:00:35] *** wks has quit IRC
[13:47:35] *** rcjsuen_ is now known as rcjsuen
[14:02:07] <magnet> hi rcjsuen
[14:02:44] <magnet> are there bugs opened about delta problems?
[14:03:13] <rcjsuen> magnet: Some, but depends what you mean by "problems".
[14:04:21] <magnet> rcjsuen, main problem i experience is that delta contains a "selectedElement" for the partstack, so I get an IllegalArgumentException stracktrace in stdout
[14:04:22] <rcjsuen> Though I suppose if you don't really care you could always just save the whole moel.
[14:04:51] <rcjsuen> I guess it is the new code we added where the selected element has to be an actual child of the parent?
[14:04:58] <magnet> yes
[14:05:00] <magnet> and that's cool
[14:05:07] <magnet> but what's not cool is throwing an IAE in a saferunner
[14:05:25] <magnet> because you can't prevent it from printing the stacktrace on stdout
[14:05:39] <rcjsuen> I thought we log the error.
[14:05:58] <magnet> thing is, the setSelectedElement() is used at different places, and we need a way to report when it fails
[14:06:17] <magnet> it's logged as well, but SafeRunner prints it first
[14:06:29] <magnet> 		if (newSelectedElement != null && newSelectedElement.getParent() != this) {
[14:06:29] <magnet> 			throw new IllegalArgumentException("The selected element "
[14:06:29] <magnet> 					+ newSelectedElement + " is not a child of this container");
[14:06:29] <magnet> 		}
[14:06:32] <magnet> that's the code
[14:07:31] <rcjsuen> magnet: So you want the error to not be caught?
[14:07:55] <rcjsuen> I feel like I'm not understanding something :)
[14:07:58] <magnet> I want it to be caught and a possibility to make it fail silently, because it's not a showstopper :)
[14:08:14] <magnet> in fact I have this exception but the whole thing works as it should
[14:08:16] <rcjsuen> So you want a way to suppresss delta errors?
[14:08:25] <magnet> yeah
[14:08:36] <rcjsuen> Okay maybe we could change it to be a trtacing option instead.
[14:08:36] <magnet> ignore them at least :)
[14:08:47] <magnet> the problem is tricker
[14:08:51] <magnet> oh
[14:08:57] <magnet> except it's written such as:
[14:09:12] <magnet> (2 mins)
[14:09:42] <magnet> ModelReconcilingService:57
[14:09:44] <magnet> 	SafeRunner.run(new ISafeRunnable() {
[14:09:45] <magnet> 				public void run() throws Exception {
[14:09:45] <magnet> 					status[0] = delta.apply();
[14:09:45] <magnet> 				}
[14:10:28] <magnet> I think a new SafeRunner should be used instead
[14:10:47] <rcjsuen> Actually I wonder if there's actually a bug in the persistence or if the selected element is correct but the children list hasn't been merged yet
[14:10:54] <rcjsuen> maybe the selectedElement deltas needs to be a delayed delta too
[14:10:59] <rcjsuen> or maybe it is alrdy
[14:11:08] <magnet> one that doesn't consider everything except OperationCanceledException an "irreversible" problem
[14:11:24] <magnet> yes rcjsuen the main problem is a bug in delta serialization
[14:11:34] <magnet> i can reproduce it on my app:
[14:11:46] <magnet> 1) I select a part in my part stack (not the default one)
[14:11:57] <magnet> 2) I switch to a perspective where that partstack doesnt exist
[14:12:35] <magnet> 3) I start my app again and it complains about the part I selected not being in the a partstack of the new perspective
[14:12:47] <rcjsuen> how odd
[14:13:33] <rcjsuen> magnet: Okay I see what you mean in that SafeRunner line
[14:13:45] <rcjsuen> it's cause the default SafeRunner from Equinox logs the err
[14:13:48] <rcjsuen> if not an OCE
[14:13:52] <magnet> rcjsuen, yep
[14:13:57] <rcjsuen> So it's not somethign we (Platform UI) controls
[14:14:01] <magnet> it's an ugly SafeRunner to use here :)
[14:14:13] <magnet> Maybe an E4SafeRunner we control would be better
[14:14:58] <magnet> I don't really expect model reconciliation problems to be logged in equinox logs anyway
[14:15:47] <rcjsuen> magnet: Yes I guess they should be non-fatal anyway.
[14:36:09] *** yarthorn has joined #eclipse-e4
[15:06:08] <yarthorn> emoffatt: don't worry about a new screen shot for N&N
[15:06:11] <yarthorn> http://www.youtube.com/watch?v=07kESXY5m2Q
[15:06:31] <rcjsuen> the suggested videos are very weird
[15:07:09] <yarthorn> rcjsuen: yeah, I don't know how that works
[15:23:48] <emoffatt> yarthorn: thanks dude, looks great !!
[15:24:28] <emoffatt> ya gotta show me how to do that sometime :)
[15:24:59] <emoffatt> be good for the (coming) 'ghosted' Drag and Drop
[15:25:25] <yarthorn> emoffatt: Markus asked for a bugzilla link to give feedback for the transition effects
[15:25:38] <emoffatt> kk, coming
[15:25:42] <yarthorn> emoffatt: a screen shot of transition effects just wasn't working!
[15:29:55] <emoffatt> yarthorn: https://bugs.eclipse.org/bugs/show_bug.cgi?id=357939
[15:30:10] <emoffatt> yarthorn: lol
[15:58:25] <magnet> :)
[15:58:35] <magnet> rcjsuen, where does the deltas serialization take place?
[15:58:45] <rcjsuen> magnet: on shutdown
[15:59:02] <rcjsuen> magnet: org.eclipse.e4.ui.internal.workbench.ResourceHandler.save()
[15:59:22] <magnet> thanks
[15:59:47] <magnet> is there any way (like a tag) to prevent serialization (or part of it)?
[15:59:53] <magnet> I don't want my labels to be saved :(-
[16:00:06] <rcjsuen> Hm, no, we don't have that option.
[16:01:03] <magnet> mhh.
[16:01:53] <magnet> any way to plug code to the reconcilier so i can tell it to ignore some features?
[16:03:39] <rcjsuen> magnet: only if you want to roll your own bundle :)
[16:04:02] <rcjsuen> where's that damn code
[16:04:38] <rcjsuen> magnet: line 262 						modelReconcilingService.applyDeltas(deltas);
[16:04:47] <magnet> rcjsuen, i'm already rolling a few patched e4 bundles but I hope to get back to mainline by 4.2, so I can't diverge too much :)
[16:04:54] <rcjsuen> you can use the (Collection, String[]) method
[16:05:01] <rcjsuen> try doing new String[] { "label "}
[16:05:31] <magnet> yeah that looks good
[16:05:48] <rcjsuen> magnet: But it will (if working ;)) ignoer _all_ labels
[16:05:56] <magnet> from ResourceHandler.java:loadMostRecentModel
[16:07:10] <magnet> wow :)
[16:07:15] <magnet> that method must not be tested ;)
[16:07:30] <magnet> oh no
[16:07:34] <magnet> log contains old entries
[16:08:18] <magnet> need to rebuild cleanly first
[16:09:20] <magnet> 						IModelReconcilingService modelReconcilingService = new ModelReconcilingService();
[16:09:26] <magnet> injection would be nice here :)
[16:09:50] <magnet> if it were an OSGi service I could plug my own with the filtering I like
[16:10:27] <rcjsuen> magnet: So is it working or not? :o
[16:10:27] <magnet> (ResourceHandler:256)
[16:10:54] <magnet> rcjsuen, I have to rebuild first, I had forgotten that I renamed 30 classes 10 mins ago :)
[16:10:59] <rcjsuen> o
[16:11:05] <rcjsuen> magnet: Let me open a bug about using @I for IJMRS
[16:11:06] <rcjsuen> IMRS
[16:11:12] <magnet> recurring spelling mistake by a coworker :)
[16:11:53] <magnet> rcjsuen, yes that would be nice. OSGi service would be ideal so we can provide one with a higher service ranking.
[16:13:03] <magnet> rcjsuen, works perfectly
[16:13:08] <rcjsuen> ok
[16:13:13] <rcjsuen> magnet: It is now officially tested :)
[16:13:25] <rcjsuen> I have junit tests but that doesn't always prove production ;)
[16:13:27] <magnet> 						modelReconcilingService.applyDeltas(deltas, new String[] { "label" }); from ResHandler
[16:14:41] <rcjsuen> ~tell magnet about 357947
[16:14:41] <Arbalest> magnet: Bug 357947 - https://bugs.eclipse.org/bugs/show_bug.cgi?id=357947 - Platform / UI / 4.2 - All / All - NEW / normal / - Assignee: Platform-UI-Inbox - ResourceHandler should inject its instance of IModelReconcilingService
[16:14:48] <magnet> btw our app on 4.1 feels a lot more reactive / lightweight, it's a very positive update for us
[16:14:57] <magnet> thanks
[16:15:00] <rcjsuen> I wonder why :o
[16:15:08] <rcjsuen> I guess it means we did something right last yr :)
[16:15:21] <magnet> yeah, that or the SWT guys did ;)
[16:16:07] <magnet> the new DnD / etc addons work much better also. And deltas aren't the mess they were, even if it's not perfect because of that ugly trace I get on stdout
[16:21:01] <magnet> rcjsuen, will you make the IMRS a context service or an OSGi service?
[16:21:48] <rcjsuen> magnet: you can choose if you want to do the work :)
[16:22:23] <magnet> rcjsuen, I can do it now, it's a quick one, but i just looked and all the DS defined in that bundle are context functions
[16:22:48] <magnet> I'm a stock OSGi service kind of guy and I fear IMRS would feel alone among those context services ;)
[16:23:19] <magnet> well i'll give it a try, OSGi style
[16:23:25] *** tschindl has quit IRC
[16:24:22] <magnet> rcjsuen, btw, didn't follow the events / Git migration / platform migration /etc, is e4.ui.wb in platform now?
[16:24:33] <rcjsuen> magnet: yes it's in Git
[16:24:38] <rcjsuen> get the R4_development branch
[16:24:42] *** tschindl has joined #eclipse-e4
[16:24:53] <magnet> oh, new repositories :)
[16:25:06] <magnet> good
[16:25:21] <magnet> Unnamed repository; edit this file 'description' to name the repository.
[16:25:24] <magnet> for anyone with commit rights
[16:26:02] *** tschindl has quit IRC
[16:27:10] <magnet> see http://git.eclipse.org/c/
[16:27:42] <rcjsuen> that's e4, you want platform
[16:28:02] <rcjsuen> magnet: http://git.eclipse.org/c/platform/eclipse.platform.ui.git/
[16:29:53] <magnet> cloning
[16:31:53] <magnet> rcjsuen, is there a gerrit instance?
[16:31:56] <magnet> for that repository
[16:50:10] <magnet> well it's not cloned yet. Slow network.
[17:09:22] <rcjsuen> magnet: No, we don't have Gerrit up
[17:09:24] <rcjsuen> only jgit/egit guys have it
[19:15:35] *** emoffatt_ has joined #eclipse-e4
[19:16:09] <emoffatt_> rcjsuen: when I'm installing GIT what 'line endings' choice makes the most sense ?
[19:16:26] <rcjsuen> I think I told it to not do anything.
[19:16:31] <emoffatt_> kk
[19:16:39] <rcjsuen> But you can always change it again later if you want.
[19:16:53] <rcjsuen> And EGit doesn't honour it properly anyway afaik
[21:06:59] *** yarthorn has quit IRC
[21:08:21] <emoffatt_> rcjsuen: not going to use egit anymore I think...
[21:09:08] <emoffatt_> the concepts are making a bit more sense (even if the command syntax is a bit confusing ('add' -> 'stage'...)
[21:12:24] <rcjsuen> emoffatt_: it's basically saying what you're committing, you're adding it to the list of items you're going to commit
[21:13:01] <rcjsuen> even from the context menu  it's 'Add File to Git Index'
[21:13:32] <emoffatt_> it's like us and Activities / Capabilities... we talk about it one way (stage) but implement it another 'add' :)
[21:14:34] <emoffatt_> just playing with the various parameters now...like how to clone a single branch from origin
[21:28:23] *** yarthorn has joined #eclipse-e4
[21:28:38] <emoffatt_> rcjsuen: git just downloaded 74 Meg with --no-checkout ??
[21:28:49] <rcjsuen> what did you do
[21:28:49] <rcjsuen> clone?
[21:29:15] <emoffatt_> git clone <url> --no-checkout /R3_dev
[21:29:23] <rcjsuen> you still cloned the repo tho
[21:29:45] <emoffatt_> it did make a repro but no branches (which is what i wanted :) )
[21:30:09] <rcjsuen> Not sure I follow.
[21:30:12] <rcjsuen> you cloned the repo
[21:30:17] <emoffatt_> then I 'checkout' the R3_development branch
[21:30:19] <rcjsuen> and it's thrown its stuff in .git
[21:30:38] <emoffatt_> yes...I know that we have a lot of history :)
[21:31:15] <emoffatt_> I think it did what I wanted and it was pretty quick (about 1 min)
[21:32:12] <emoffatt_> bascially I'm setting up two working dirs; one for 3.x and one for 4.x in different clones so I can compare them
[21:33:18] <emoffatt_> then I'll make two workspaces and import the projects from the appropriate clone
[22:02:11] *** yarthorn has quit IRC
[22:12:24] *** yarthorn has joined #eclipse-e4
[22:23:16] <emoffatt_> rcjsuen: have you used VIM ?
[22:23:30] <rcjsuen> Only a little.
[22:23:50] <emoffatt_> how do i resolve which one i want ?
[22:24:01] <rcjsuen> Resolve which what
[22:24:41] <emoffatt_> lol, I have a merge conflict so I can choose left (br1) or right (br2)...how do I pick which in the tool
[22:25:08] <rcjsuen> I haven't faced that problem.
[22:25:40] <emoffatt_> no prob...I'll see if there are others...the color scheme is giving me a headache anyway :)
[22:26:20] *** yarthorn has quit IRC
[22:28:51] *** rcjsuen is now known as rcjsuen_
[22:49:46] *** rcjsuen has joined #eclipse-e4
[22:53:35] *** yarthorn has joined #eclipse-e4
[22:57:56] *** yarthorn_ has joined #eclipse-e4
[22:58:08] *** yarthorn has quit IRC
[22:58:15] *** yarthorn_ is now known as yarthorn
[23:10:30] *** emoffatt_ has quit IRC
[23:22:42] *** yarthorn has quit IRC
[23:43:11] *** briandealwis has quit IRC
[23:52:13] *** briandealwis has joined #eclipse-e4





top