[11:58] benji, some small conflicts in https://code.launchpad.net/~benji/lpsetup/bug-1023519-store-version-information/+merge/114736 [12:00] gary_poster: thanks; looking [12:01] welcome [12:03] gary_poster: fixed and pushed [12:03] hi gary_poster [12:04] i re-ran lxc-install last night and it failed on make schema. looks like it is trying to do that in the wrong directory [12:06] you know, Launchpad really should email you when an MP has a conflict [12:12] bac ok [12:12] bac benji frankban http://tinyurl.com/yellowsquad [13:02] http://codesinger.blogspot.com/2012/07/yellow-squad-weekly-topics-july-6.html [14:44] benji: done [14:45] bac: thanks [15:15] bac: I responded tou your (excellet) review https://code.launchpad.net/~benji/lpsetup/bug-1023519-store-version-information/+merge/114736/comments/246891 [15:16] if all that sounds good I will go ahead and mark the MP as approved [15:18] benji: sounds good. thanks. [16:03] gary_poster: i got my branch to work and have created a new branch and MP: https://code.launchpad.net/~bac/lpsetup/fix-inithost-install-lxc-deux/+merge/114896 [16:04] bac, reviewing [16:07] bac fully approved (vote and status) [16:14] thanks gary_poster [16:14] * bac -> BLT [16:23] gary_poster: unfortunately I can not create a fake repository under /tmp/, because then init-repo complains about repository not being inside my home directory. [16:24] gary_poster: I can think about 2 solutions: 1) create the repository inside a temporary dir in the home or 2) this is an integration test that must be run using juju [16:25] frankban, as to #2, we want it part of the integration juju tests anyway (also, I guess? I thought you were working on an integration test) [16:25] frankban, but as to 1, yeah, why not? [16:26] frankban, you can pass dir to tempfile functions, looks like [16:26] gary_poster: yes [16:26] gary_poster: what I'd like to do is: [16:26] 1. create a fake source repository using a template (as we discussed) under /tmp/ [16:27] 2. test init-repo passing --source [fake source] with and without --checkout [16:27] 3. check the command does not raise errors [16:28] 4. check for the fake file to be correctly present in the checkout (or in the branch if --no-checkout is passed) [16:28] 5. destroy the source, destroy the target, done. [16:29] frankban, sounds great, except make step 1 work with source in ~/[tempdir] (or checkout/[tempdir]?), right? [16:29] (the fake file is just one test file added and committed to the source [16:30] frankban, please as I said on call, eventually I want fake repo to have more contents [16:30] gary_poster: the source file can be in /tmp/, it can be everywhere, the repository to create must be under ~ [16:30] frankban, oh, right. [16:31] my sentence "please as I said on call," was edited badly. just meant "as I said on call" [16:31] gary_poster: yes, we can reuse the fake source for integration tests in the future [16:31] great [16:32] gary_poster: so, for this unit tests, i am going to create a temp repository in the home, right? [16:33] frankban, right. To be ore specific, you are going to use tempdir to create a tempdir in ~; then you will make repo in ~/[tempdir], right? [16:33] s/ore/more/ [16:35] gary_poster: yes, repository will be something like tempfile.mktemp(dir='~') [16:36] frankban, perfect [16:37] gary_poster: for the lpsetup package, it's near my EOD and I will try to add these changes to my branch, so... [16:38] * gary_poster listens [16:41] gary_poster: I don't think I can take care of that [16:42] frankban, sure, of course! would you like to pass it off to someone else, or save it for Monday, or pass it off to someone else, and in case they don't finish it you'll do it on Monday? :-) [16:45] gary_poster: I meant I can not take care of building a new lpsetup package, I can do that on Monday, or I pass it off if you need the new package today [16:46] frankban, oh duh, sorry [16:46] frankban, cool, thanks for letting me know [16:46] I didn't read carefully enough and missed the transition [17:09] gary_poster: cool, np [19:48] gary_poster, benji: regarding versioning, it seems we need to keep track of 1) contents of deleted files, 2) names of modified files, and 3) names of added files on a per-version basis. when the version changes we can 1) restore the deleted files, 2) unmodify the ones we changed if we put fences around beginning and end of our changes, 3) remove the added files. [19:49] we cannot restore modified files to the original since the dev or system could've made other changes in the meantime. perhaps this is obvious and you've already noted it. [19:49] bac, agreed. benji, we had a list of four items on a google doc. do you still have that hanging around? [19:50] four things that we thought we ought to save [19:50] probably can find [19:50] gary_poster: yep, one sec [19:50] bac, they are in this bug: https://bugs.launchpad.net/lpsetup/+bug/1023519 [19:51] <_mup_> Bug #1023519: Store version information < https://launchpad.net/bugs/1023519 > [19:51] i read that bug this morning but it doesn't state explicitly what i was thinking above [19:53] bac, true. what you wrote has thought through some of the points more thoroughly. [19:53] bac, perhaps it would be worthwhile to add that as a comment to the bug, or to update the description. Beyond that... [19:53] sure [19:55] bac, I was trying to think if I could see a card that ought to be made from that, but I don't see one--unless we don't add the footers yet? [19:55] we do not add footers [19:55] if we don't add the footer "fence" that would be worth a card [19:56] +1 for card then, bac. Thank you.