=== _mup__ is now known as _mup_ [00:36] wgrant: http://pastebin.ubuntu.com/1654509/ [00:43] StevenK: Mmm [00:43] StevenK: That's a bit of a regression [00:44] I wonder if LimitedView might be better reserved for the "you can only see this archive because a build was copied out of it into a public archive" case [00:44] Which might mean that an intermediate permission is required [00:45] wgrant: Well, we can override getBuildRecords on IArchive to return [] if the requestor is a subscriber and it's a private archive [00:45] StevenK: Ew [00:46] A proper solution would not involve hacks like that, and would remove the +packages hack that is in place today [00:46] Yes, which I did, and you say it's a regression [00:47] StevenK: The regression is that +index no longer works [00:48] I didn't know the baselayot template will only show the not shared with you string if you hold LimitedView [00:49] Sure [00:49] It's still a regression [00:49] Not realising that something was going to break does not make it not broken :) [00:49] Sure. But what can I do instead? [00:49] Use a non-LimitedView permission, I guess [00:49] launchpad.NotQuiteView doesn't help either [00:50] Do we have a list of them? [00:54] StevenK: lib/lp/permissions.zcml [00:55] Hmmmm [00:55] None of them are a good fit === wallyworld___ is now known as wallyworld === wallyworld is now known as Guest86404 [03:01] wgrant: Hmmmm, where is view/macro:is-page-contentless defined? My quick grep didn't turn up much [03:02] Bah [03:03] It's in app/browser/tales [03:03] And it's as I feared, it checks for !View [03:03] Luckily it is code [03:04] You could alternatively reassign View, I guess [03:04] wgrant: I've added launchpad.SubscriberView, does that make me terrible? [03:04] i think it would have upset stevea about 5 years ago [03:04] Haha [03:05] The comments in lib/lp/permissions.zcml make some interesting reading [03:05] mwhudson: Never mind that, most of the launchpad team managed to annoy stevea in Budapest [03:08] wgrant: Reassign View? [03:08] StevenK: You could make +packages etc. require more than View, potentially [03:08] Indeed [03:09] +packages requires View in the ZCML, but the view wants Append [03:09] +delete requires Edit [03:09] So it's a mess [03:09] Hmm? [03:10] Append is for uploading/copying packages in [03:10] +packages does not need Append [03:10] - # To see the +packages page, you must be an uploader, or a [03:10] - # commercial admin. [03:10] - if not check_permission('launchpad.Append', self.context): [03:10] Oh, you mean for a private archive [03:10] Yes [03:11] Anyway, Append is clearly wrong here [03:11] Because you can't protect the view with it [03:11] It is already is if it's private [03:11] No [03:11] There's a hack that you are to entirely disregard [03:11] The view is not protected [03:13] Right, +edit and +delete require .Edit, and +builds and +packages require View [03:14] And +subscriptions requires Append [03:15] Right [03:16] The key is that subscribers need to be able to see +index but not +packages [03:16] Yes [03:16] So you need multiple view permissions [03:16] And LimitedView probably doesn't fit [03:16] => you probably need another viewish permission [03:16] Not Append or Edit [03:16] So I've created SubscriberView [03:17] Right, but you need to consider how that will interact with things that like View [03:17] But that trips straight over the is-page-contentless, because that fires if you don't hold View [03:17] I think lazr.restful checks for LimitedView now instead, so it's probably fine [03:17] But stuff like is-page-contentless will not be [03:50] OOPS-cb18303a07bb367312fd6366522207cc [03:50] Hmm? [03:51] What exactly did you expect? :) [03:51] not a timeout ;) [03:51] It's an unbatched page that probably isn't linked from anywhere [03:51] Oh, it is batched [03:51] But not very well [03:52] The BTF UNION BTF UNION BTF query is horrible [03:52] Still, AFAIK it's unlinked and unreleased and well known to be horrible [03:52] You got what you ordered :) [03:56] so delete it [04:01] Person:+subscriptions 19 13.25 [04:09] I'm not sure it's necessarily worth deleting at this point [04:13] whats the current top freq timeout ? [04:14] 455 / 0 POFile:+translate [04:14] Most of the time [04:14] thats not so bad [04:14] Sometimes someone runs a dreadful bot that calls getMergeProposals on project groups, which can exceed that [04:14] well done === cyclicflux is now known as Guest70333 === Guest70333 is now known as CyclicFlux === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === yofel_ is now known as yofel === almaisan-away is now known as al-maisan === wedgwood_away is now known as wedgwood === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan === dpm_ is now known as dpm === wedgwood is now known as wedgwood_away === wedgwood_away is now known as wedgwood === al-maisan is now known as almaisan-away