#instantbird log on 10 22 2019

All times are UTC.

06:04:47 <-- ivan has quit (Quit: Leaving)
06:24:31 --> ivan has joined #instantbird
10:09:09 <-- florian has quit (Client exited)
10:44:29 --> florian has joined #instantbird
10:44:29 * ChanServ sets mode +qo florian florian 
11:14:32 --> clokep has joined #instantbird
11:14:33 --> fredw has joined #instantbird
11:49:38 <-- fredw has quit (Client exited)
12:24:17 <clokep> Hey florian, curious if you'll have time soon-ish to take a look at bug 1533872 or if I should try to get someone else to review it.
12:24:20 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1533872 maj, --, ---, clokep, ASSI, Update and re-integrate the Matrix JavaScript SDK
12:32:40 <florian> clokep: hello
12:32:57 <florian> there are gigantic patches there for the update, but you set review flags on only 2 small patches
12:33:02 <florian> which kind of review do you need?
12:34:01 <clokep> florian: https://bugzilla.mozilla.org/show_bug.cgi?id=1533872#c7
12:34:03 <instantbot> Bug 1533872 maj, --, ---, clokep, ASSI, Update and re-integrate the Matrix JavaScript SDK
12:34:17 <clokep> A few comments above that describes what each patch is.
12:34:41 <florian> ah, so all the changes in the patches that have review flags were typed by you?
12:40:44 <clokep> florian: Yep!
12:41:31 <clokep> Uhh...plus the "README.md" file in the patch to update the dependencies.
12:41:44 <clokep> Which has the methodology I followed. I should probably have put that in a separate patch.
12:46:54 <florian> do we need anything in about:license ?
12:47:33 <clokep> florian: I do not believe so, we didn't have it before. :P
12:47:37 <clokep> But everything is Apache or MIT.
12:47:44 <clokep> (I'll check again though.)
12:50:57 <florian> how do you feel about the performance of including all these dependencies?
12:51:49 <florian> but I guess there's no viable alternative anyway...
12:51:56 <clokep> I don't think there's really an alternative, yeah.
12:52:11 <florian> nothing seems to be lazy loaded
12:52:13 <clokep> I tried to not include the "bluebird" one and use native promises...but they use random other pieces of bluebird that have like extra Promise methods.
12:52:31 <florian> I looked a bit at the unhomoglyph thing
12:52:32 <clokep> florian: I suspect we should lazy load the main matrix module?
12:52:48 <clokep> And that would fix "most" of that issue?
12:52:52 <florian> as soon as it's loaded, it loads a JSON data file that contains a list of 6000+ characters
12:53:12 <florian> and then creates a big regexp based on that. Eagerly :(
12:53:37 <florian> and there's const unhomoglyph = require('unhomoglyph'); at the top level of matrix/lib/matrix-sdk/utils.js
12:53:49 <florian> with such a name, I suspect this file of being loaded from various places
12:53:54 <clokep> For what it's worth, the people working on that seem to be willing to make changes...
12:54:21 <florian> yeah, I was gonna say, the best bet to improve stuff seems to be to upstream patches
12:54:42 <florian> as I don't think we have the resources to maintain a fork
12:59:30 <florian> so... I guess you are all set ;)
12:59:32 <florian> good luck!
13:00:09 <clokep> :+
13:00:11 <clokep> :-P
13:00:21 <clokep> They already fixed the import thing...so that was nice
13:00:26 <clokep> (The one that had the .js file on the end...)
13:00:36 <florian> they should use eslint
13:00:44 <clokep> I think they do...
13:00:56 <clokep> Yeah: https://github.com/matrix-org/matrix-js-sdk/blob/develop/.eslintrc.js
13:01:37 <florian> how do you feel about Magnus' patch?
13:01:54 <florian> I almost r+'ed, but given you are no longer the author, you can do it yourself :)
13:02:30 <florian> oh
13:02:36 <florian> I wonder why that ; was missing then
13:04:02 <clokep> florian: Cause that code you commented on isn't part of the matrix-js-sdk.
13:04:05 <clokep> It's part of a dependency.
13:04:12 <clokep> That's is super shitty and that I'd like to get rid of.
13:04:15 <clokep> But baby steps. :))
13:04:50 <clokep> florian: So...I see what Magnus is saying in light of the other bug...but I don't have a strong feeling one way or the other. :P
13:06:01 <clokep> I find all this code silly. Why does it always refer to contacts? :P
13:06:06 <clokep> Half the time it is conversations!
13:06:13 <florian> not having an additional random attribute is better IMO
13:07:10 <florian> "Why does it always refer to contacts?" is this a rhetorical question? Or are you curious about the history?
13:09:00 <clokep> I suspect it is because it was adapted from the Instantbird contact list, which only showed contacts. :)
13:09:09 <florian> correct
13:09:13 <clokep> But then they wanted to show contacts and conversations in a single pane and get rid of tags, blah, blah.
13:09:19 <clokep> It makes the code confusing to follow though.
13:09:49 <clokep> florian: Thanks for the reviews by the way, your suggestions all make sense. :)
13:10:15 <florian> np, sorry for the delays
13:10:35 <florian> should I put an r+ flag on Magnus' proposed patch, or do you want to handle it?
13:10:55 <clokep> florian: I want to test it first. ;)
13:11:00 <florian> ok
13:11:05 <florian> that's better :)
13:11:11 <clokep> So I'll take care of it! Thanks!
13:11:23 <clokep> If it makes you feel any better...that matrix patch also removes like 4 dependencies?
13:13:00 <florian> I saw you were using shims to get rid of pointless dependencies :)
13:26:53 <clokep> By the way, all of this code using promises makes it much nicer to deal with. :)
13:28:38 --> Logicoma has joined #instantbird
13:48:33 <-- florian has quit (Client exited)
13:58:25 * clokep does a rebuild to make sure this stuff still works...
14:34:06 --> florian has joined #instantbird
14:34:06 * ChanServ sets mode +qo florian florian 
15:03:12 <clokep> And of course it doesn't...
15:03:13 * clokep sighs.
15:55:36 <clokep> florian: By the way, it doesn't seem to load any of those "other" modules until I try to connect a Matrix account. I'm not sure where that lazi-ness would come from though.
15:55:49 <clokep> Are JS modules treated as lazy now?
15:56:16 <clokep> Ah, we have this at the top of the matrix protocol file: ChromeUtils.defineModuleGetter(this, "MatrixSDK", "resource:///modules/matrix-sdk.jsm");
15:57:58 <florian> :)
15:59:01 <clokep> Looks like past you already had the same concern. ;)
16:24:16 <florian> heh
16:24:22 <florian> you mean I already cared about performance? ;)
16:27:27 <clokep> Yep!
16:34:19 <-- florian has quit (Ping timeout: 121 seconds)
16:34:43 --> florian has joined #instantbird
16:34:43 * ChanServ sets mode +qo florian florian 
16:55:59 * clokep is now known as clokep-test 
16:56:17 * clokep-test is now known as clokep 
16:56:32 * clokep is now known as clokep-test 
16:56:37 * clokep-test is now known as clokep 
17:04:36 <clokep> florian: So checking out Magnus' patch...it seems to mostly work (the conv label updates and resorts into the proper location), but I end up with *no* conversation selected, and chat-conversation.js element throws a few errors (this.conv is undefined @ line 83, this.docShell is undefined at line 425). So I think something is missing in the "update"...or the "old" element is still getting notifications, but has no UI attached to it
17:04:36 <clokep> or something.
17:06:45 <florian> :(
17:16:20 <clokep> I'm unsure how to debug this, but it seems fishy.
17:18:48 <florian> might be a change between how XBL destructors worked and the new post de-XBL behavior
17:21:52 <clokep> Maybe.
17:27:10 <-- florian has quit (Ping timeout: 121 seconds)
17:36:56 --> florian has joined #instantbird
17:36:56 * ChanServ sets mode +qo florian florian 
20:28:20 --> chrisccoulson has joined #instantbird
20:29:27 <-- clokep has quit (Client exited)
22:11:52 <-- Logicoma has quit (Ping timeout: 121 seconds)
22:54:07 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)