[13:02] <benji> I can't find any general practitioners open yet.  I guess doctors don't like to get an early start.
[13:07] <gary_poster> heh, yeah, I find 9 AM is the magical starting time typically
[13:33] <gary_poster> frankban, just approved "tests mutate/break uri" branch
[13:33] <gary_poster> good to land
[13:33] <gary_poster> thank you!
[13:37] <frankban> gary_poster: cool, thanks
[13:37] <gary_poster> welcome.
[13:42] <frankban> gary_poster: hum... the merge is full of conflicts, do we now use "describe" inside "YUI(GlobalConfig).use..." in tests?
[13:50] <gary_poster> frankban, ah :-( yes, we do that in some and we are due for a discussion about the pattern this Friday
[13:50] <gary_poster> I think it is the right change to make
[13:50] <gary_poster> but we should talk about it
[13:50] <gary_poster> meanwhile, reconciling with the conflicts is the right practical thing to do I think
[13:51] <frankban> gary_poster: yes, I am going to do that. 
[13:52] <gary_poster> cool
[14:10] <gary_poster> bcsaller, yo
[14:22] <bcsaller> gary_poster: hey, want to talk through things at 9:30?
[14:22] <gary_poster> bcsaller, sounds good thanks
[14:33] <frankban> gary_poster: this may or may not be related, but test filtering (?grep=) does not seem to work with new style tests (those wrapped by YUI)
[14:33] <gary_poster> frankban, :-(
[14:33] <gary_poster> frankban, thank you for discovering and reporting that.  GO ahead with the merge now
[14:34] <gary_poster> We might have to revert the overarching change after discussion
[14:34] <gary_poster> BUt that can be separate
[14:43] <frankban> gary_poster: landed
[14:43] <gary_poster> frankban, yay thanks
[14:55] <bac> hi gary_poster, you have a minute?
[14:55] <gary_poster> on call bac.  will take awhile.  need something soon?
[14:56] <bac> gary_poster: just wanted to chat about approaches to charm panel testing
[15:58] <gary_poster> bac bcsaller benji frankban goodspud (?) hazmat jovan2 Makyo mattuk1972 (if you wanna?) teknico call in 2
[16:30] <gary_poster> bcsaller, gonna have lunch then will check back in with you
[16:30] <bcsaller> gary_poster: sounds good
[16:43] <benji> gary_poster: is there a way I can signal that I have done the UX review?  (artifact-wise, other than just telling the people involved)
[17:07] <Makyo> Window guys are coming by today, I guess?  Out for just a few to get things ready for them...
[17:45] <gary_poster> benji no.  I added my name as a tag after I did the review, but you already had your name there
[17:45] <benji> gary_poster: ok; for what it's worth, I have done my review
[17:46] <gary_poster> ok thanks benji
[17:46] <gary_poster> bcsaller, you available at 1:00PM to continue?  1:30?
[17:46] <bcsaller> gary_poster: lets say 1:30
[17:46] <gary_poster> cool, talk to you then
[17:50] <benji> gary_poster: I have come to believe that bug 1078978 does not exist.  Have a moment to discuss?
[17:50] <_mup_> Bug #1078978: CSS is not minified in static deployment <juju-gui:Triaged> < https://launchpad.net/bugs/1078978 >
[17:50] <gary_poster> heh, sure benji
[17:50] <gary_poster> juju-ui benji
[19:21]  * benji drags himself off to the doctor's office.
[19:30] <Makyo> gary_poster, good news: the serving-css-from-wrong-path seems to have been a problem with merging with trunk and a previous change causing strange behavior.  Should I just mark the bug as invalid?
[19:30] <gary_poster> Makyo, yes, cool and sorry you encountered that annoyance
[19:31] <Makyo> gary_poster, Good that it was something so simple. 
[19:31] <gary_poster> agreed :-)
[20:40] <benji> Is "You have big puss pockets." a compliment?
[20:41] <bac> benji: depends on where they are
[20:41] <benji> lol
[20:42] <bac> benji: are you likely to pull through?
[20:42] <benji> bac: it is touch and go, but the codine should improve things
[20:43] <bac> gary_poster, benji: either of you have time for a quick call?
[20:44] <gary_poster> bac, I can
[20:44] <bac> cool
[20:44] <benji> bac: sure
[20:44] <benji> oh, Gary wins
[20:44] <gary_poster> you have puss pockets.  I'm pretty sure those are like cooties
[20:45] <bac> gary_poster: normal hangouty place?
[20:45] <gary_poster> bac, yeah juju-ui
[20:46] <benji> heh
[21:11] <gary_poster> benji, I don't think that the bzr ignore approach will work for what Matt wants
[21:11] <gary_poster> Could be wrong
[21:11] <gary_poster> Sorry s/Matt/Makyo
[21:11] <benji> hmm, I wonder if I misunderstood the intent of the branch
[21:11]  * benji looks again
[21:11] <gary_poster> This changes the watch function
[21:11] <gary_poster> which the debug server uses
[21:12] <gary_poster> to always keep generated files up to date
[21:12] <gary_poster> I think?
[21:12] <gary_poster> yeah
[21:12] <benji> oh, you're right
[21:13] <gary_poster> cool
[21:13] <benji> in that case my comment is a subset of the comment I want to make: i.e., the only things that should be watched are those that bzr knows about
[21:14] <benji> that's the approach we take in the Makefile, and it is reasonable
[21:14] <gary_poster> that sounds nice.  OTOH, changing that involves more work
[21:14] <gary_poster> I think this is a nice incremental improvement
[21:14] <gary_poster> Though maybe if I knew how to call out to shell from JS I'd feel differently ;-)
[21:16] <benji> require('child_process').exec('rm -rf /')
[21:17] <benji> oh, and there is a second argument that is a function that is called for each line of output generated by the subprocess
[21:17] <benji> (or something like that)
[21:18] <gary_poster> cool
[21:18] <gary_poster> Makyo ^^^
[21:18] <gary_poster> thanks benji
[21:18] <benji> my pleasure
[21:22] <Makyo> benji, gary_poster, with vim, there are 10 files total that the watcher sees: 3x 4913, 3x stylesheet.less~, 4x stylesheet.less.  The files are created/deleted/modified within a few milliseconds and node does something for each of them.  Given bzr's speed, I don't know how well that'd work, because by the time it got done checking whether the first 4913 file existed, the other two would've been created and removed, and ditto the backup~ fi
[21:22] <Makyo> le.
[21:22] <Makyo> I don't know if the watcher queues the rest.
[21:23] <benji> Makyo: interesting
[21:23] <gary_poster> speed is an interesting question, fair enough
[21:23] <Makyo> I'm not quite sure how/when emacs' #stylesheet.less# works
[21:24] <benji> it sounds like a file change should place the file name in a list of pending files and schedule a regeneration for a short time in the future (e.g., 1 second)
[21:24] <gary_poster> I'm not sure that is necessary
[21:25] <benji> any more file changes will push the regeneration further out; when the alarm finally goes off we can consult bzr to see if any of the changed files are under its management, and if so, do the work
[21:25] <Makyo> I was originally thinking a whitelist would work, but then we'd have to update that list every time we add a new file...
[21:25] <gary_poster> bzr ls -V is not *that* slow
[21:25] <gary_poster> .064s for me
[21:25] <benji> yeah, it may be an unneccesray optimization
[21:26] <Makyo> gary_poster, true, and I suppose it will return false both if the file exists and is not under vc, and if the file doesn't exist.
[21:26] <gary_poster> it is not a matter of checking, I don't think Makyo . You'd use that to get your list of files
[21:26] <Makyo> gary_poster, I see.
[21:27] <gary_poster> may not work well with fs.watch though...
[21:27] <gary_poster> looking...
[21:27] <benji> if speed is a problem, the first optimization would be to cache the results of "bzr ls -V" for, say, a second; that would eliminate several runs in the flurry-of-temp-files scenario
[21:28] <gary_poster> I think if we are really running into this, we should land what Makyo has as a nice improvement of the status quo and move on
[21:29] <gary_poster> Makyo, I was wrong that you would get your list of files that way, I think.  But I still suspect that shelling out to bzr ls would be fine
[21:29] <benji> I'm a big fan of not letting the perfect be the enemy of the good.  I'm also somewhat fuzzy on where the line is between good and ok-but-will-hurt-you-in-the-future.
[21:30] <gary_poster> heh, fair enough
[21:31] <Makyo> I think that, in this case, the only future difficulty we'd have to worry about would be a different editor creating different files (and even then, it's on the editor's user whether or not they care enough - this was bugging me while debugging LESS)
[21:31] <benji> Makyo: what is the failure mode?
[21:31] <gary_poster> The bigger worry is that we are not including something in the compilation that should be included
[21:32] <gary_poster> that would be surprising
[21:32] <gary_poster> bigger worry IMO
[21:33] <benji> I also suspect the magic filename "4913" is likely to vary over time, neccesitating frequent upkeep.
[21:33] <gary_poster> Makyo, maybe a good way to do what we want here: https://lists.ubuntu.com/archives/bazaar/2008q3/047134.html
[21:34] <Makyo> benji, It's been in there since 2009, but yeah, I get that.  I can try gary_poster's suggestion, though, see how that works
[21:35] <gary_poster> same speed as ls, but more direct answer
[21:35] <gary_poster> bzr ls I mean
[21:35] <gary_poster> speed is all in Python's VM I suspect
[21:35] <gary_poster> startup cost
[21:36] <benji> heh, if the file 4913 exists vim will then try to create a file named 5036, and keep adding 123 until it finds an unused file name
[21:37] <gary_poster> heh
[21:37] <Makyo> Ooh, good catch, just found the source for that.
[21:38] <Makyo> Oh, vim...must you be so coy? 9.9
[21:42] <benji> ok, I'm getting dizzy now, probably from the codine; I'll (likely) see you guys tomorrow.
[21:42] <gary_poster> good luck benji.  feel better
[22:04] <Makyo> gary_poster, I have it working with using bzr to check.  Like me to reproprose?
[22:04] <Makyo> -r
[22:04] <gary_poster> sure Makyo 
[22:25] <Makyo> Reproposed (will remove the commented out bit if we go down this path)
[22:45] <gary_poster> Makyo, Iapproved with some important changes
[22:45] <gary_poster> have a good evening
[22:45] <Makyo> gary_poster, thanks, you too