[11:11] hi gmb -- any progress on the /var/tmp problem? [11:20] bac: none yet. Just grabbing some lunch; maybe we should pair on it after standup. [11:21] gmb ok. i'll fire up an ec2 instance [11:21] K [12:09] oops [12:09] bac benji frankban gmb call RSN [12:09] getting a hangout address... [12:10] bac benji frankban gmb https://talkgadget.google.com/hangouts/_/9ee69a171ceb3e2b794d492280bead509d4fddb8 [12:11] gmb is fighting firefox? [12:11] gary_poster: Had lost connection. Joining now. [12:12] cool [12:13] Arrg [12:13] Now I'm fighting firefox [13:01] gary_poster: if you're still chatting with benji would you mention he's not on irc? [13:02] bac, I'm ready to hang out / skype / whatever whenever you are. [13:03] gmb: ok, you want to create a hangout? [13:03] Sure. [13:04] bac, Done. Invitation sent. [13:04] https://plus.google.com/hangouts/_/8ef4543f15cb72fc89bf5cb8a2e897d3e614e3e2, for reference [13:10] bac, benji won't be able to be here today after all [13:10] gmb, did you want me to join? got an invitation [13:10] gary_poster, So did everyone, apparently. [13:10] No, it's okay. [13:10] ok cool [13:10] We had a very confused Aaron. [13:11] heh [13:11] gary_poster: in that case, can you and i talk during benji's 2:00 slot? [13:11] did you include circles? [13:11] bac, sure [13:29] gary_poster: gmb did point out the script branch-rewrite.py is added as a RewriteRule to the local-launchpad apache config. so bouncing apache will cause /var/tmp/bazaar.launchpad.dev to be created. can't find a link yet but it is good to know there is another (hidden) path that causes it to be made. [13:30] bac, ah, interesting. good catch. [13:31] bac, gary_poster: The only reference I can find to /var/tmp/bazaar.launchpad.dev in the output of `make -d clean build schema` is the rm -rf of it that make build does. [13:31] ... which is pretty much what we expected, I think. [13:31] gmb: yep, on an unbroken machine [13:31] Ah. [13:31] OICWYM. [13:32] it'll only be really interesting when run via lp-setup-lxc-build [13:32] gmb, that matches what I saw, yeah. I thought that, eventually, we might want to use "make inplace" rather than "make build" or "make install" because it seems to create the directories as we want [13:32] ah, frankban, whoops. getting a hangout room... [14:27] gmb, need to step away and will ping when I return. no later than quarter till. S'ok? [14:27] gary_poster, That's fine. [14:27] thanks [14:36] gmb: i've got an instance that appears to be up. want to hangout? [14:36] bac, Sure. I'll need to hop over to my one-to-one with Gary at some point soon, but we might as well make a start. [14:37] gmb: soon being top of the hour? [14:37] Hmm. Actually, sooner. [14:37] bac, So, why don't you start poking around and I'll join you when I can. [14:37] Rather than interrupting things just as they get going. [14:37] gmb: ok [14:38] gmb just create a hangout and invite me when you're read [14:38] y [14:38] ok [15:02] gmb, meh, sorry, and I made you not hangout [15:02] but I'm ready now [15:02] getting hangout [15:02] Okay. [15:03] https://talkgadget.google.com/hangouts/_/110aa68ecccc7bb4323ad132b1079bff120478e1 [15:03] gmb ^^ [15:03] gmb, gary_poster: aha! [15:03] * gary_poster eagerly awaits aha moment [15:04] the lxc container is getting brought up and down all of the time. each time it comes up, apache starts and the branch-rewrite script gets run. (this is the difference from a local machine) [15:04] so it is apache coming up that makes that directory [15:04] why this only happens now i don't know [15:04] but bac, that happens also when we ssh in [15:04] but i can remove the directory, shut down the container, start it back up, and there it is [15:04] no [15:04] to an lxc container that its already running [15:05] bac https://talkgadget.google.com/hangouts/_/110aa68ecccc7bb4323ad132b1079bff120478e1 ? [15:12] * bac dances a jig [15:12] http://paste.ubuntu.com/1004886/plain/ [15:12] gary_poster, gmb: ^^ [15:12] this time i *did* mean to interrupt [15:13] bac :-) yay! awesome, good catch [15:13] thx to gmb's insight [15:13] now what do we do about it :-) [15:14] Maybe we don't run make install [15:14] initially [15:14] and do make inplace instead [15:17] perhaps. we may have a timing issue though. [15:18] lp-setup-lxc-build pretends to wait for the container to be "ready" but clearly apache is still coming up at that point [15:19] and if apache beats us, we're screwed unless we can delete that directory as root [15:19] which i don't think we can [15:47] gary_poster: can i join the hangout with you and gmb when you finish your weekly? [15:50] bac, We finished; let me start a new one and invite you both [15:50] (And hopefully not $world+$dog [15:50] ) [15:50] * bac hates airlines. just bought tickets for my parents and had to pay an extra $70 so they can sit next to each other. [15:57] gmb could you give me url? [15:58] gary_poster, https://plus.google.com/hangouts/_/4b1e08d9fc3d9e4c60b747ee42d066a6cc4b6056?authuser=0&hl=en-GB# [17:28] gary_poster: here is the change i'm proposing. will test on ec2 now. http://paste.ubuntu.com/1005098/ [17:28] looking [17:28] so, it was a wee bit over two characters [17:29] but i wanted to ensure rewrite.log didn't get created by apache [17:29] bac, I think that diff is reversed. Could you confirm pls [17:30] yep, the first is reversed the second is correct. [17:30] got it [17:30] i always get those wrong when doing bzr diff [17:32] looks nice bac [17:35] so, bac, after doing that, do you still feel like the "tests should pass after make inplace" approach should be pursued, or do you think this is far enough to take it? [17:35] gary_poster: i think it would be nice to not have to worry about apache if we don't have to [17:36] gary_poster: lpsetup will need to manage test vs dev box setup if that is the case. (it may already do that) [17:38] yeah, to some degree it does [17:38] robert has already asked for/suggested more [17:40] bac, I'm going to retroactively make a bug for this and assign you to it [17:40] then put it on the card [17:41] gary_poster: ok [17:41] gary_poster: i've started a PPA build for my lpsetup changes. it'll be an hour or so before it is ready [17:42] cool [17:42] and then another hour for the instance to be ready [17:42] at which point we'll be pretty close to EoD [17:42] i've created a yellow/experimental ppa for adding stuff we just want to play with that we can easily remove [17:42] cool, good idea [17:43] bac, do you agree with this summary? [17:43] Apache configuration is necessary for tests but makes it impossible to use "make clean" safely in lxc [17:44] yes [17:44] cool [18:02] bac, almost finished with bug then will call you [18:05] bac, https://bugs.launchpad.net/launchpad/+bug/1004088 [18:05] <_mup_> Bug #1004088: Apache configuration is necessary for tests but makes it impossible to use "make clean" safely in lxc < https://launchpad.net/bugs/1004088 > [18:06] bac, ping me when you are ready and I will start hangout [18:06] ping [18:06] * bac runs to get water [18:08] gary_poster: ok [18:08] bac https://talkgadget.google.com/hangouts/_/808e46bd4d2b62c66982476debd3a15c7440298f [18:56] urgh, my PPA is dragging along. an hour later it has a 52 minute eta.