[00:40] <jelmer> thumper: This one should be ok: https://code.edge.launchpad.net/~jelmer/launchpad/always-import-link/+merge/33070
[00:40] <thumper> ta
[00:41] <thumper> jelmer: while this solve the immediate problem, what I really want is a single view to register a new branch whether that be an import or mirror
[00:41] <thumper> but this is fine for now
[00:43] <jelmer> thumper: That makes sense. I guess all that's needed is a single URL in most cases?
[00:43] <thumper> normally
[00:44] <thumper> we could even have a validator that splits the URL for cvs imports
[00:44] <thumper> so they go 'pserver...location module'
[00:44] <thumper> and don't allow creation of remote or hosted branches that way
[00:45] <thumper> but we'd need push instructions
[04:18] <jtv> EdwinGrubbs: you still reviewing?
[04:19] <EdwinGrubbs> jtv: sorry, no
[04:19] <jtv> :-(
[04:20] <jtv> mwhudson, can you review a branch for me?  I was hoping to get it CP'ed.  Yesterday's reviewer left me in the lurch, and now there's no OCR.
[04:20] <mwhudson> jtv: i can take a look
[04:20] <jtv> https://code.edge.launchpad.net/~jtv/launchpad/pd-bug-617431/+merge/32991
[04:20] <jtv> thanks
[04:20] <thumper> anyone? https://code.edge.launchpad.net/~thumper/launchpad/fix-canonical-url/+merge/33079
[04:21] <thumper> jtv
[04:21] <thumper> ?
[04:21] <jtv> thumper: ah what the hell… ok
[04:24] <thumper> jtv: since I have you anyway, did you see my branch that changed a translations test?
[04:24] <jtv> no?
[04:24]  * thumper digs
[04:25] <thumper> https://code.edge.launchpad.net/~thumper/launchpad/fix-TestSharingMigrationPerformance/+merge/32828
[04:25] <thumper> jtv: the gc.get_objects() was fouling up another translations test using bzrlib and lazy imports
[04:25] <thumper> jtv: so in consultation with lifeless I rewrote the guts of the test
[04:26] <thumper> jtv: do you want to check it over to make sure I didn't screw up?
[04:28] <thumper> jtv: convenitently I'm using POFile to test the canonical_url fix too
[04:28] <jtv> thumper: I'm looking at it
[04:29] <thumper> ah fooey, spelling fail
[04:29] <jtv> where?
[04:29] <thumper> fingers not obeying brain
[04:29] <thumper> convenitently
[04:29] <jtv> sounds very catholic somehow
[04:29] <thumper> too many Ts
[04:31] <thumper> jtv: which are you looking at?
[04:32] <jtv> thumper: the last one you mentioned
[04:32] <mwhudson> thumper: does removing
[04:32] <mwhudson>         if rootsite is None:
[04:32] <mwhudson>             rootsite = 'mainsite'
[04:32] <mwhudson> not affect anything?
[04:33] <thumper> no, because by the time we get there, we have already set rootsite if it is None
[04:33] <thumper> it is done further up the method
[04:34] <jtv> thumper: your changed test for memory usage in the message-sharing migration script is stricter than the original—it'll break if we should want to mention those tables in the SQL—or possible even foreign keys to them.  But if it works, I guess it's good enough for now.
[04:34] <thumper> jtv: yes I know it is more strict
[04:34] <thumper> jtv: but it passes now
[04:34] <thumper> jtv: we can change the test if you need to reference the tables later
[04:34] <thumper> jtv: sound OK?
[04:34] <jtv> yes
[04:34] <thumper> good
[04:38] <jtv> thumper: back to your canonical_url branch… I really only have one concern: in lines 80—83 of the diff, are you going to take a lot of exceptions to get to LaunchpadLayer?  If so, is there any performance risk to that?
[04:39] <mwhudson> thumper: obj_urldata.rootsite can't be None?
[04:39] <mwhudson> i guess tests will find out
[04:40] <thumper> mwhudson: I'm pretty sure not
[04:41] <thumper> jtv: I'm not sure I understand
[04:43] <jtv> thumper: you're catching ComponentLookupError there… AIUI exceptions in python can be very very slow (although mostly from a cold cache I hope).  So do you expect these very often?
[04:44] <thumper> jtv: the overhead isn't particularly big, and we'll get an error for any view for an object on the mainsite
[04:44] <thumper> jtv: no I don't expect any performance issues here
[04:44] <thumper> jtv: remember this is only used when someone specifies view_name
[04:45] <jtv> Remember?  Is this something I knew before?
[04:46] <jtv> Honestly, *what is the point* in having the little sticking-out plastic lip on a patch cable that makes the satisfying little click when you slot it in if *it doesn't help keep the plug in the socket*?
[04:46] <jtv> I know, I know.  It gives the satisfying little click.
[04:46] <thumper> it should keep it in the socket
[04:46] <thumper> if it doesn't
[04:46] <thumper> it is broken
[04:47] <jtv> TIT.  Everything's broken.
[04:50] <mwhudson> jtv: reviewed
[04:50] <jtv> mwhudson: thanks!
[04:51] <thumper> mwhudson: do you remember the process for getting a CP?
[04:51] <mwhudson> thumper: no
[04:51] <thumper> hmm...
[04:52] <jtv> You branch off production-devel
[04:52] <thumper> jtv: I branched of devel before the rollout
[04:52] <mwhudson> thumper:  https://wiki.canonical.com/Launchpad/PolicyandProcess/EmergencyChange  maybe?
[04:52] <jtv> You MP against production-devel, requesting "code" review and a CP one
[04:52] <mwhudson> might be very out of date for all i know though
[04:52]  * thumper looks
[04:53] <jtv> You get approval to land on production-devel
[04:53] <jtv> You land, then wait for the branch to percolate through to production-stable
[04:53] <jtv> You register the request on LPS.
[04:53] <jtv> thumper: the review type for the CP is "production-change."  I think you normally go to the RM for that.
[04:54] <jtv> thumper: I reviewed your canonical_url branch btw
[04:54] <thumper> ta
[04:55] <jtv> thumper: btw if it wasn't clear, what I describe above is the current CP procedure ^^
[04:55] <thumper> jtv: yeah, I'm just thinking that my change isn't really a CP candidate
[04:55] <jtv> thumper: it needs a Critical bug and an Incident Report.
[04:55] <thumper> it is just blocking landing other work which is making our SLA longer
[04:55] <thumper> :)
[04:56] <jtv> I don't think that qualifies, no.
[15:24] <leonardr> salgado, can you do a quick review for me of https://code.launchpad.net/~leonardr/lazr.batchnavigator/build-url-without-length/+merge/33045 ?
[15:30] <salgado> leonardr, sure, but I'm about to jump into a call, so it may take a while
[15:30] <leonardr> salgado, np. anyone else, feel free to take it
[16:04] <leonardr> mars, could i maybe get you to look at that branch? -^
[16:04]  * leonardr just trying random people
[16:08] <mars> leonardr, awesome doctest :)
[16:09] <salgado> leonardr, r=me
[16:10] <leonardr> salgado, thanks. mars, nevermind
[16:10] <mars> k
[17:02] <jsackett> EdwinGrubbs: when/if you have a moment, I believe I addressed your notes on my mp.
[17:37] <EdwinGrubbs> jsackett: I've replied to your changes. BTW, do you preferred to be called Jon or J.C.?
[17:56] <leonardr> mars, maybe you can review https://code.edge.launchpad.net/~leonardr/launchpad/optimized-length/+merge/33132 for me?
[17:56]  * leonardr finds it hard to believe he's the only one with branches that need reviewing today
[18:04] <jml> leonardr, don't worry, you aren't :)
[18:04] <jml> #launchpad-reviews is only one of many ways to get a review.
[18:30] <mars> leonardr, still need that review?
[18:31] <leonardr> mars, yes please
[18:37] <mars> leonardr, looks great, r=mars
[18:37] <leonardr> great
[18:40] <jcsackett> EdwinGrubbs: reviewed your notes and made the changes. and Jon or JC work, I answer to both.
[18:57] <henninge> bac: Hi! I assume I have your rs on landing the imports reformatting branch? ;-)
[18:58] <bac> henninge: yes.  i'm pondering how disruptive this will be
[18:59] <bac> henninge: but week 1 is as good a time as any, right?
[18:59] <henninge> bac: yes. Directly after roll-out would have been best, but ... oh well ;)
[19:00] <bac> henninge: rs=bac
[19:00] <henninge> But I agree that earlier in the cycle is better than later.
[19:00] <henninge> bac: thanks ;)
[19:00] <bac> henninge: thanks to you!
[19:01] <lifeless> if you were to hook up a bzr merge plugin
[19:01] <lifeless> then it would be painless anytime
[20:07] <henninge> bac: I messed up the first ec2 run, the one that succeeded. It ran on plain devel. That ought to succeed :(
[20:08] <bac> henninge: ok.
[20:08] <henninge> bac: the real one is failing badly.
[20:09] <henninge> bac: can you make anything out of this: http://paste.ubuntu.com/480557/
[20:09] <bac> henninge: otp right now
[20:09] <henninge> ok
[20:14] <bac> henninge: what have you done?
[20:15] <bac> henninge: did your process munge a bunch of stuff you didn't expect?
[20:15] <henninge> bac: what do you mean?
[20:16] <bac> henninge: that paste.  it's frightening
[20:16] <henninge> yes, that's truw
[20:16] <henninge> e
[21:26] <EdwinGrubbs> jcsackett: your branch looks good. As long as it's passing the full test suite, it can be landed.
[21:26] <jcsackett> it's passing most of what i've touched, EC2 will tell us the rest.
[21:26] <jcsackett> thanks, EdwinGrubbs.
[21:27] <jcsackett> rather, it's passing all of what i've touched.
[21:33] <jcsackett> EdwinGrubbs: sorry about not posting the incrementals; i had assumed those were easily accessible from the mp, but hadn't really looked.
[21:35] <EdwinGrubbs> jcsackett: I believe incremental diffs will be automated shortly, so you may end up learning a procedure that is immediately abandoned.
[21:36] <jcsackett> EdwinGrubbs: won't be the first time, nor the last, i'm sure. :-)