[00:00] I'm looking at it. [00:00] I'm vaguely disgusted by lines 70-84 of the diff [00:00] Oh, you're just moving code. [00:01] Yeah, moving body elements into the body :) [00:02] </script> [00:02] Excuse me while I review my breakfast. [00:02] Thankyou Internet Explorer :) [00:06] wgrant: r=me [00:06] StevenK: Thanks. === StevenK changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: StevenK | Firefighting: - | Critical bugtasks: 4*10 [00:24] wallyworld__: I've handled PUBLICSECURITY -> UNEMBARGOEDSECURITY in my branch. [00:24] StevenK: aewsome, thank you :-) [00:25] StevenK: there's a card for that i think [00:26] 510 line branch [00:33] wallyworld__, wgrant: Either of you want to review https://code.launchpad.net/~stevenk/launchpad/move-to-informationtype/+merge/96271 ? [00:33] i can [00:34] StevenK: Why did you use sponge? [00:34] sed -i... [00:34] StevenK: btw, a branch of mine should be out of ec2 rsn and wil need to be merged into yours [00:34] Arse [00:52] wallyworld__: When is your branch due out of ec2? [00:53] StevenK: i ec2 landed it around 7am [00:53] My time or yours? [00:53] not should be done very soon i hope [00:53] s/not/so [00:53] my time of course :-) [00:53] wallyworld__: Run 'bin/ec2 ls' ? [00:54] up for 3:54 [00:54] Right, it should land in the next 15 or so [00:54] wallyworld__: I'll hold off on the branch you approved, then. [00:54] StevenK: thanks, otherwise there will be conflicts [00:54] Wait for yours, merge, fix conflicts and re-check. [00:55] yes, plus rename new uses of AccessPolicyType [00:55] in tests etc [00:55] Right [00:55] there will only be a few, if any, can't remember [00:55] I was going to grep for it again [00:56] ok [00:56] I can share the one-liner I used, if you wish [00:56] ok [00:56] i don't use sed that much [00:56] wallyworld__: for f in $(bzr grep -l AccessPolicyType) ; do sed -e 's/AccessPolicy\(Type\)/Information\1/g' < $f | sponge $f ; done [00:56] i love a got bit of regex [00:56] bzr grep -l AccessPolicyType | xargs sed -i -e ''s/AccessPolicyType/InformationType/g' [00:56] good [00:57] Bah, xargs [00:57] i like xargs better too [00:57] what does sponge do? [00:59] wallyworld__: sponge soaks up all standard input and then writes it to a file [01:00] wallyworld__: If you do 'sed < file > file' what happens is the shell opens file for writing with truncate, so when sed gets a hold of it, it's empty [01:00] oh, interesting [01:01] wallyworld__: http://pastebin.ubuntu.com/872355/ [01:02] right [01:16] StevenK: merge done [01:16] * StevenK updates his devel with fear and trepidation [01:17] StevenK: if you get yours into ec2 within say 30 minutes or a bit longer, you'll make the next bb run [02:09] StevenK: https://code.launchpad.net/~wgrant/launchpad/bug-869631/+merge/96280 [02:10] wgrant: Excellent work, r=me [02:10] Thanks. [02:44] wallyworld__, wgrant: Either of you want to review https://code.launchpad.net/~stevenk/launchpad/drop-garbo-accesspolicy/+merge/96285 ? [02:45] StevenK: Weren't there additional product and distribution grants to remove security.cfg? [02:45] This should probably just be a straight antimerge of the rev. [02:45] It was [02:51] wgrant: bzr di -r 14900..14901 doesn't show the distribution/product addition [02:52] StevenK: Hm, I thought it was in the original MP. Maybe it isn't any more. [02:52] Or it was added with the workitem migrator [02:52] That's more likely. [02:52] I remember adding it [02:57] wgrant: Removed them manually [02:57] StevenK: You'll probably break the workitem migrator. [02:57] "Good" ? [02:57] Heh [02:58] Yes, r14893 (the workitem migrator) added those two [02:58] Which is why they're not in your diff. [02:58] Ah [02:59] * StevenK wonders if bzr uncommit and push --overwrite will cause issues [02:59] Will work fine. [03:03] wgrant: Since we've just spent 15 minutes talking about it, do you want to rubber stamp it? [03:03] If X stops crashing. [03:03] Unlikely. [03:03] At least with -radeon it doesn't hang the kernel. [03:03] I suggest NVidia. [03:03] Is approved. [03:04] wgrant: Thank you. [03:04] Nah, these crashes are unrelated to the drive. [03:04] r [03:04] I can see it conflicting horribly with my branch that is in ec2 now, but meh [03:29] wgrant: Can you explain the trigger stuff you mentioned on the call this morning? [03:34] StevenK: You may not need a trigger, but it's possibly easiest. [03:34] StevenK: We need to migrate from the two flags to the enum. [03:35] There are multiple triggers and bits of application code that use the flags. [03:35] The easiest way to migrate may be to map flag changes to enum changes in the DB. [03:38] So I add a bit to bug_mirror_legacy_access ? [03:40] Where is bigjools? Apartment hunting? [03:40] sat here! [03:40] He's teasing me in another channel, at the moment. [03:41] StevenK: You could potentially work it into that. [03:41] I was straight-batting your troll back [03:41] StevenK: It's not clear how it fits into bugtaskflat. [03:41] StevenK: (which isn't landed yet, partly for this reason) [03:41] Mmmm [03:42] So we can ignore the trigger thing, and do it in code+garbo [03:42] Which allows bugtaskflat in [03:43] Our data access layer is sufficiently chaotic that that is normally impractical. [03:43] But for this limited case you may be able to do it. [03:43] Since the two flags already have custom behaviour. [03:43] Right. [03:45] bugtaskflat may continue to have the private flag, I suspect. [03:46] Calculated from whether the type is private or not. [03:46] But it may also want to have a copy of the type. [03:46] I suspect it would do better with just a copy of the type [03:47] Then Bug.private and Bug.security_related can just *die* [03:47] And good riddance [03:48] Right, those two certainly die. [03:49] But for performance and sense it may be best for BugTaskFlat.private to still exist. [03:49] bigjools: HA. I did not se that reply until now :) [03:50] nigelb: :) [03:50] wgrant: I will ignore triggers for now and deal with updating information_type in my model branch [03:50] StevenK: Sounds reasonable. [03:50] I guess it's only the appservers that really matter here. [03:51] ENOSTUB though [03:53] lifeless: Are you happy to review this DB patch, or do I get to wait until tomorrow afternoon? [03:53] lifeless: Also, want to request a deploy of your DB patch? [03:55] StevenK: stub should be around soon, I may have a look if you toss me a url [03:55] wgrant: good idea! want to toss it up? [03:55] lifeless: stub will not be around. It's a public holidays in .th today. [03:56] lifeless: Sure [03:57] wgrant: thanks [03:57] StevenK: ah, so what is the url? [03:57] lifeless: Still sorting out the MP, give me a few. [03:59] lifeless: https://code.launchpad.net/~stevenk/launchpad/db-add-information_type/+merge/96290 [04:05] we should probably deprecate comments [04:05] do them in the patches [04:07] lifeless: I think that's probably orthogonal :-) [04:07] anyhow, you tried to use a default, which you can't. [04:08] Well [04:08] You can add a default afterwards [04:08] But you can't set a default on create [04:08] Should be able to do it in the same patch, no? [04:08] alter table add column foo integer default 1 -> rewrites the heap. [04:09] alter table add column foo integer; alter table alter column foo set default 1; -> ok. [04:09] (not-quite-sql) [04:09] Right. [04:09] That's what I Meant. [04:09] StevenK had the former [04:09] Also, comments don't take locks, so we can keep comments.sql if we want [04:09] wgrant: meh. [04:10] wgrant: one system to rule them all [04:11] Just thinking about it, the default is dangerous, so I'll remove it [04:11] Hmmm, I think xx-builder-page.txt might get into an infinite redirect loop with the broken