[02:21] <lifeless> so, whos up for kicking bug mangler Tobias B ?
[02:27] <wgrant> I've fixed those and emailed him
[02:27] <wgrant> If he continues, a suspension will swiftly follow.
[02:58] <StevenK> wgrant: I'm confused. The test that fails is actually firing an event which ends up tossing an exception that is caught and then returns false. That code path for the event framework is not called for 3.10.2, so there is no exception, therefore the function returns true.
[03:01] <wgrant> StevenK: Have you traced through it in both versions?
[03:03] <StevenK> There is a maze of event functions, but I set a breakpoint on the line that causes the exception in 3.9.1, and it doesn't get hit in 3.10.2
[03:17] <StevenK> wgrant: The fire() function in 3.10.2 looks very very different -- in 3.10.2, the call to getSiblings() (which is what dies horribly in 3.9.1) is guarded by this._hasSiblings
[03:22] <wgrant> StevenK: You'll probably want to step through it and work out what's going wrong.
[03:22] <StevenK> http://pastebin.ubuntu.com/5737687/ fixes it in 3.10.2
[03:24] <wgrant> What does it mean?
[03:24] <wgrant> Shouldn't the response always have an event_key?
[03:25] <StevenK> Ah, but this test is for malformed data
[03:25] <StevenK> Which only has { something: 6 } and no event_key
[03:26] <wgrant> Y.fire behaves differently in the face of undefined?
[03:28] <StevenK> Yes, it fails to find silblings in 3.9.1, hence tosses an error
[03:28] <StevenK> It doesn't look for silblings in 3.10.2 and so behaves
[03:31] <wgrant> Can you see the rationale for the change?
[03:47] <StevenK> wgrant: Based on a quick Googling, it is make fire() more performant
[03:49] <wgrant> StevenK: What's the intent of the test?
[03:51] <StevenK> wgrant: That manager.sucessPoll() with that responseText returns false
[03:53] <wgrant> StevenK: Oh, so it catches the exception, right. That diff sounds OK, then
[03:54] <StevenK> Well, in fact it fires a namespace.longpoll_fail_event with the exception, so maybe mine should do the same too
[05:49] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/add-yui-3.10.2/+merge/167689
[05:51] <wgrant> StevenK: Can you remove the try/except?
[05:51] <wgrant> Surely that check will work for 3.9.2 as well
[05:51] <wgrant> Or are there other potential errors?
[05:51] <wgrant> 3.9.1
[05:51] <StevenK> wgrant: Just in case there are other errors, like failure to decode
[05:52] <wgrant> StevenK: Would it be better to just throw an exception rather than duplicating the failure firing?
[05:53] <StevenK> Oh
[05:53] <StevenK> Could do that
[05:57] <StevenK> wgrant: http://pastebin.ubuntu.com/5737890/
[06:09] <wgrant> StevenK: Do we prefer to throw strings or objects (eg. Error)?
[06:11] <StevenK> wgrant: I see evidence of both, even in the same JS files
[06:11] <wgrant> Yay
[06:11] <StevenK> lib/lp/registry/javascript/timeline.js:                throw new Error("resize_frame must not be empty.");
[06:11] <StevenK> lib/lp/registry/javascript/timeline.js:            throw "Invalid y_align argument: " + y_align;
[06:12] <wgrant> Yayer
[06:13] <StevenK> 38 for Error and kin, 11 for bare strings
[06:23] <StevenK> wgrant: Diff updated
[06:26] <wgrant> StevenK: k
[06:27] <StevenK> wgrant: So, launchpad.Edit for DAS.
[06:28] <StevenK> wgrant: Shall I use launchpad.Moderate or so then?
[06:28] <wgrant> Moderate might work indeed
[06:28] <wgrant> In other news, person has too many FKs referencing it.
[06:29] <StevenK> News at 11
[06:29] <wgrant> Somewhere around 200
[06:37] <StevenK> wgrant: http://pastebin.ubuntu.com/5737952/
[06:40] <wgrant> StevenK: Sounds reasonable
[06:47] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/das-setchroot-api/+merge/167464
[06:50] <wgrant> StevenK: That docstring is not good. "Set the chroot tarball used for builds in this architecture. The SHA-1 checksum must match the chroot file.", maybe.
[06:51] <wgrant> That even lets you format the docstring legally.
[06:51] <wgrant> Otherwise that looks OK, except for the file upload bits, but not much we can do about that because lazr.restful and Zope.
[06:52] <StevenK> Oh, do you want the XXX there like the other callsites to that horrible method that I will hopefully pay for in blood one day?
[06:54] <wgrant> That would be nice.
[06:55] <StevenK> wgrant: ... That I add the XXX, or pay in blood for that horrible method? Both?
[06:55] <wgrant> Both would be ideal.
[06:57] <StevenK> http://pastebin.ubuntu.com/5737987/
[06:57] <wgrant> Docstring should be formatted correctly.
[06:59] <StevenK> wgrant: http://pastebin.ubuntu.com/5737994/
[06:59] <wgrant> Right.
[07:06] <StevenK> wgrant: Diff updated
[07:08] <wgrant> StevenK: r=me
[07:10] <StevenK> wgrant: Thanks