=== Spads_ is now known as Spads
[12:35] <LeoNerd> GAHHH
[12:36] <LeoNerd> ONCE AGAIN  bzr up  did not abort on diverged history but instead made a merge
[12:36] <LeoNerd> How do I a) unfuck this, and b) stop it EVER doing it again?
[12:36] <LeoNerd> I do not *EVER* want  bzr up  to merge. if there's divergant history I want it to complain, so that I can rebase
[12:58] <Spads> LeoNerd: so bzr up kind of *is* merge
[12:59] <LeoNerd> Yes. I don't want one
[12:59] <Spads> so why did you run bzr up, exactly?
[12:59] <LeoNerd> I want it to bring me in new non-divergent history, or abort if there is
[12:59] <Spads> are you doing a bzr checkout?
[12:59] <LeoNerd> It's a checkout yes
[13:00] <Spads> is there a reason you're doing that instead of bzr branch?
[13:00] <LeoNerd> Because most of the time I'm online and therefore I want to use the central server
[13:00] <Spads> the semantics of bzr checkout are there to support people who are uncomfortable with anything but SVN, near as I can tell
[13:00] <Spads> everyone else just uses bzr branch
[13:00] <Spads> and does pull/merge/push
[13:00] <Spads> a pull will abort on divergent history
[13:01] <Spads> up is basically "merge from where I checked out to" and that is basically all it is
[13:02] <Spads> s/ to//
[13:02] <LeoNerd> Ah; maybe I should just use pull then..?
[13:03] <Spads> I think so
[13:03] <Spads> you'll need to branch first
[13:03] <LeoNerd> Nah
[13:03] <Spads> your checkout has no repo basically
[13:03] <LeoNerd> pull :bound == up, give or take
[13:03] <Spads> yeah
[13:03] <Spads> see this conflict detection is a repo opration
[13:03] <Spads> mostly
[13:03] <Spads> so it's about two repos communicating
[13:03] <Spads> (I think)
[13:03] <Spads> so a branch will have a local repo
[13:04] <Spads> and can thus negotiate the conflicts correctly
[13:04] <Spads> but if pull :bound works, then do use it!
[13:04] <Spads> but divergent history vs :bound doesn't make sense to my mind
[13:04] <Spads> history only exists in branches
[13:04] <Spads> not checkouts
[13:04] <Spads> checkouts defer history to the bound branch
[13:04] <LeoNerd> I only really get divergent history if I was offline on my laptop, like on a train
[13:04] <LeoNerd> As happened here
[13:04] <Spads> if you're doing work offline on your laptop, you want to branch locally
[13:05] <Spads> unless this is something pathological like a repo with 44G of history or something
[13:05] <LeoNerd> Hrm? A checkout is a branch
[13:05] <Spads> not really
[13:05] <LeoNerd> bzr co == bzr branch + bzr bind
[13:06] <Spads> hm
[13:06] <LeoNerd> It's not a _lightweight_ checkout.. it still has a local branch
[13:06] <Spads> ahhhh
[13:06] <Spads> okay sorry
[13:06] <Spads> yes, I was thinking lightweight
[13:06] <LeoNerd> :)
[13:06] <Spads> so yeah, see if pull :bound has the right semantics for you
[13:06] <LeoNerd> Oh, Ialready know it does. I just keep forgetting to do that
[13:06] <Spads> heh
[13:06] <LeoNerd> ((You only need :bound the first time; it'll remember afterwards)
[13:06] <Spads> bzr pull --remember is handy when you get those wrong
[13:08] <Spads> maybe bzr alias up="pull :bound" would help :)
[13:08]  * Spads does that on his one heavyweight checkout
[13:11]  * Spads has been using bzr since 2006 or so, and could swear that bzr co used to do lightweight checkouts
[13:12] <LeoNerd> Mm... I don't think it ever has done
[13:12] <LeoNerd> Though that might be a configurable option
[13:14] <Spads> iinteresting
[13:14] <Spads> TIL
[17:51] <fullermd> Your problem isn't using 'up', it's using 'ci --local'.  I still maintain that its existence is a bug   :p