=== _mup__ is now known as _mup_ [11:48] gmb, haha, "think for mrevell" [11:49] danilos: Indeed. And now he's the Product Manager, so I probably don't have to :) [12:30] bac benji danilos gmb call in 1 [12:30] Ok [12:30] ack [12:33] uhm, where did the call window go? [12:46] benji: do you still use 8.8.8.8? [12:46] bac: yep (and 8.8.4.4) [12:46] gmb: http://code.google.com/speed/public-dns/ [12:46] sweet [12:47] benji: Ta [13:02] danilos, almost ready [13:02] gary_poster, ok [13:03] gary_poster, call when you are [13:03] thanks will do danilos [14:00] gmb, you ready? [14:00] gary_poster: Sure [14:00] cool [14:20] benji: did you get shaken? [14:21] bac: I didn't feel that one. [14:57] This looks odd to me: [14:57] $ bzr push lp:~gary/launchpad/bug838869 [14:57] Using default stacking branch /+branch-id/24637 at lp-67338320:///~gary/launchpad [14:57] Should't that have used ~launchpad-pqm/launchpad to stack on? [14:57] pushing took a bit longer than usual, but not forever [15:27] hi gary_poster, do you have time for a pre-imp sanity check? [15:29] bac, yes in 2 min ok? [15:29] ok [15:31] bac, Skype when you want to [15:32] gary_poster: actually i'd just like you to look at this diff and see if you think it is sane: http://pastebin.ubuntu.com/679837/ [15:32] k [15:32] trying to avoid directly specifying adapters when possible [15:38] bac, I think I understand (someone else pinging didn't help the comprehension level). So, this is for the case in which one object's security is directly tied to another's. I'm going to have to review the security adapter interface, one more moment. (Looks good so far though.) [15:46] gary_poster: hold that thought. i need to go afk for a bit [15:46] bac, "forwardCheckAuthenticated" in the existing AuthorizationBase implementation confuses me in this context. OK I was going to ask you about your opinion of EditDistroSeriesDifference in light of your changes. [15:47] (as an example) [15:48] gary_poster: yes i looked at fCA but it doesn't do enough. perhaps i could defer to it but it seemed cleaner to go the way i did [15:48] gary_poster: but, as long as you don't see a much better way to accomplish this and think it has value then we can tweak it as needed. [15:48] bac, very much so. I'd be tempted to try ripping out fCA, or implementing it in terms of your cached adapter [15:49] gary_poster: ah, ok [15:49] gary_poster: bbiab [15:49] ok [15:51] I'm hungry. [17:15] benji, I just saved a link to that incident report I gave you. It has good links :-) [17:15] I put 'em there so you'd think I'd know them, but no such luck [17:16] benji, first channel jumping of the day ;-) [17:16] pfft [17:16] :-) [17:26] bac, benji, I have lunch, then will go to dr. I'll check in with you after I get back to either talk then or if not, tomorrow morning. [17:26] gary_poster: k [17:26] ok [17:55] wow this is nice: http://www.tripit.com/group/calendar -- wish canonicaladmin was so usable [17:56] (must be a tripit user to see it, i suspect) [17:59] bac, not really here, but I was thinking that I don't think your cachedproperty will ever be a benefit--adapter will be discarded 99-100% of the time I think. adapter is not persistent across anything unless you make the factory somehow persist: it is created when adaptation call is made. [17:59] You could try setting up a factory that stashed things on a request. It would be an interesting experiment. I'm not sure if it would bring any practical benfit. [17:59] benefit [17:59] * gary_poster disappears, to dr. [20:21] bac, benji, I'm back. dr. visit could have been much worse. Let's plan to have our calls tomorrow morning, though, if that's alright with you guys? [20:21] that's fine by me [20:21] cool thanks [20:21] gary_poster: that's fine; by "could have been much worse" I assume the outcome, not the office visit itself [20:22] benji, heh, well, actually both. :-P [20:22] heh, two positives then [20:23] gary_poster: you asked about replacing the use of forwardCheckAuthenticated. i was able to do so in many cases but there were a few uses where it wasn't practical, so i've left it in. [20:24] cool bac. [20:25] bac, did you see my note about the cachedproperty? Probably a property is good enough, or even better. [20:26] gary_poster: i did and made the change [20:27] cool bac [20:27] gary_poster: i also experimented with using __new__ to actually just return the desired adapter instead. i couldn't come up with non-ugly implementation so i reverted it [20:27] was fun trying, though [20:27] heh, bac, you can make the adapter a function [20:28] it's a factory [20:28] of any callable sort