[07:08] <vila> hi all
[07:21] <vila> heya poolie
[07:22] <poolie> hi there
[07:22] <poolie> time flies
[07:22] <poolie> how are you?
[07:24] <vila> fine, was in Paris this past week-end ;)
[07:25] <poolie> lucky you
[08:43] <Merwin> Hi ! I pushed a modification on a remote repo, but I want to revert it locally and remotly
[08:43] <Merwin> Should I use bzr revert and the commit+push?
[08:52] <Merwin> vila ?
[08:53] <vila> Merwin: ?
[08:53] <Merwin> Good morning ;)
 Hi ! I pushed a modification on a remote repo, but I want to revert it locally and remotly
 Should I use bzr revert and the commit+push?
[08:57] <vila> the main question is about your remote branch, who is using it and did they already pull the modification ?
[08:57] <Merwin> Nope, I told them to not pull
[08:57] <Merwin> What if they did? :
[08:57] <vila> if not, uncommit/revert&tweak&whatever/commit/push --overwrite
[08:57] <vila> if yes, they will need to pull --overwrite instead of a simple pull
[08:58] <Merwin> ok
[08:58] <Merwin> And if they have not-pushed commits? :D
[08:58] <vila> they'll need to merge and resolve the conflicts locally
[08:59] <vila> you can also avoid the whole issue by just committing a reversed merge of your unwanted change
[09:00] <vila> i.e. your branch history will then contain both the change and the removal of the change instead of not mentioning the change at all :)
[09:00] <vila> It flows better with distributed workflows but ultimately is a mater of personal taste
[09:06] <Merwin> vila, you mike like "commit an 'inversed' commit" ?
[09:06] <Merwin> Here, I commited something I want to cancel, but they worked and modified the repo since my commit
[09:07] <Merwin> So I have to undo what has been done in the commit, and recommit
[09:07] <Merwin> Is there a way to do this automatically ?
[09:09] <vila> with X being your unwanted commit: bzr merge -rX..X-1 .
[09:09] <Merwin> Hum, I'll try
[09:10] <Merwin> bzr merge -r371..371-1 .
[09:10] <vila> 371..370
[09:10] <Merwin> bzr: ERROR: Requested revision: '371-1' does not exist in branch
[09:10] <Merwin> Ah
[09:10] <Merwin> Lol, I look stupid right now :p
[09:11] <vila> hehe, I don't think so, failing is the best way to learn
[09:12] <Merwin> Nice, it worked very good
[09:12] <Merwin> I don't understand why I have to use merge here...
[09:12] <Merwin> But it worked :p
[09:13] <vila> it's called a sherry-pick (merge a revision range instead of a branch) and is the easiest and safest way to reverse a change
[09:13] <vila> heya mgz !
[09:13] <mgz> morning!
[09:27] <Merwin> vila, http://friendpaste.com/2TAVKpXwcJV3PdPlrlEOBn
[09:27] <Merwin> This is because of the cherry-pick ?
[09:30] <vila> meh, check which branch is used for which command, 'bzr missing :push' should clarify a bit
[09:30] <vila> but the 'diverged branches' is probably because you uncommit at some point or because someone else pushed there
[09:31] <vila> Merwin: hmm, have a look at 'bzr config' too, you probably have a 'submit_branch' option defined now that is... confusing stuff, get rid of it
[09:32] <vila> Merwin: 'bzr config --remove submit_branch'
[09:32] <Merwin> I think that when I did the cherry pick from '.', it changed the default branch location
[09:32] <Merwin> Specifying the good location resolved the problem
[09:33] <Merwin> vila, thanks
[10:06] <jelmer> 'morning poolie
[10:15] <poolie> o/ jelmer, mgz
[10:16] <mgz> hey poolie
[10:39] <poolie> jelmer, the bug for junk might be a dupe, but i support it either way
[10:40] <jelmer> poolie: I'll have a look
[10:41] <jelmer> poolie: I vaguely recall that jam once described it to me as a feature of baz or tla
[10:41] <jelmer> maybe that wasn't entirely the same thing
[10:50] <jam> jelmer: baz/tla did have a non-precious ignored file as a separate concept from a precious ignored file
[10:50] <jam> I think the final decision we had come up with, was actually to have a precious ignored
[10:50] <jam> and make standard ignored 'junk'
[10:50] <jam> at least I thought that was what poolie had settled on
[10:50] <jam> because prob 95% of ignored files today are junk, and just a few cases where they are passwords, etc.
[10:51] <mgz> yup, that would be neat.
[12:45] <dholbach> hiya
[12:46] <dholbach> I was just wondering: can you see the private ubuntu bzr bugs?
[12:46] <jelmer> dholbach!
[12:46] <dholbach> hey jelmer
[12:46] <jelmer> dholbach: only those of us who are also ubuntu members
[12:46] <jelmer> s/members/developers/
[12:46] <dholbach> ok, I think I'll have a look at them now and see if there's anything private in there and unprivate
[12:46] <jelmer> dholbach: in other words, only poolie and me
[12:48] <jelmer> dholbach: thanks!
[12:49] <dholbach> ok, they're public now
[13:00] <dholbach> jelmer, am I the first reporting problems with the new bzr/bzr-builddeb combo when it comes to UDD? :)
[13:01] <dholbach> for some reason almost every udd merge I do brings me a crash (I collected 3 different ones now)
[13:01] <dholbach> ah sorry, you're patch piloting - nevermind then :)
[13:04] <jelmer> oops
[13:04] <jelmer> actually, mgz is patch piloting this week
[13:04] <jelmer> dholbach: I haven't seen any other bug reports on it yet - can you send in bugs for the issues?
[13:05] <dholbach> yep, sent them already (923706, 923691, 923688)
[13:06] <jelmer> thanks!
[13:52] <Alan502> Hello, I've been trying to commit some changes to a project but I keep getting this error: bzr: ERROR: Cannot lock LockDir(chroot-64962512:///%2Bbranch/classlink/.bzr/branch/lock): Transport operation not possible: readonly transport can somebody help me?
[13:53] <jelmer> hi Alan502
[13:53] <jelmer> Alan502: what are you trying to do exactly?
[13:53] <Alan502> jelmer, hello :)
[13:53] <Alan502> i'm trying to commit changes to a project
[13:53] <Alan502> by using bzr commit -m "comment"
[13:55] <fullermd> I'd guess a checkout over http.
[13:55] <Alan502> bzr checkout?
[13:55] <jelmer> Alan502: as fullermd says, it does indeed sound like you're trying to do a commit but the location you're committing to doesn't allow writing.
[13:56] <Alan502> I see
[13:56] <jelmer> Alan502: are you perhaps committing in a local branch that's bound to something else?
[13:57] <Alan502> hmmm
[13:58] <Alan502> jelmer, I am trying to commit to a launchpad project that me and my friend created
[13:58] <jelmer> Alan502: how did you check out the launchpad project?
[13:58] <Alan502> I need to register on that project first at launchpad.net right?
[13:58] <Alan502> I just pull to get the files
[13:58] <jelmer> Alan502: did you create a checkout or an entirely new branch of the project?
[13:58] <Alan502> it says "not a branch"
[13:58] <Alan502> nooo
[13:59] <Alan502> i haven't created a checkout
[13:59] <jelmer> Alan502: how did you create what you have locally?
[13:59] <Alan502> how do I do it?
[13:59] <fullermd> Let's just shortcut to 'bzr info'   8-}
[13:59] <Alan502> bzr pull lp:name
[14:00] <Alan502> jelmer, let me check bzr info :)
[14:00] <Alan502> http://pastebin.com/0KrrcXfg
[14:01] <fullermd> Right, checkout.
[14:01]  * fullermd wins!
[14:01] <Alan502> aha?
[14:01] <Alan502> xD
[14:01] <jelmer> Alan502: do you have write access to lp:classlink?
[14:01] <jelmer> Alan502: are you matthewowebb ?
[14:01] <Alan502> I guess that is the problem, I have to set that in launchpad right?
[14:01] <fullermd> I'm guessing you don't _want_ a checkout, considering you're pushing to a different location as well.
[14:02] <Alan502> nope
[14:02] <Alan502> Can I get write access as a different user=
[14:05] <jelmer> dholbach: thanks for the bug reports. I fixed one, will have a look at the other two a bit later.
[14:08] <Alan502> i'm trying to use the gui on the ubuntu repositories :P
[14:30] <vila> jelmer: what's the possible values for quilt-tree-policy ?
[14:32] <vila> ha, found in the tests: applied or unapplied...
[14:32] <vila> jelmer: meh, it's still unclear to me which one I should use for the importer :) There is no 'leave-me-alone' acceptable value ;)
[14:41] <jelmer> vila: the default
[14:41] <jelmer> vila: quilt-tree-policy is disabled by default
[14:41] <jelmer> vila: you probably mean quilt-smart-merge ?
[14:41] <jelmer> vila: that's just a boolean option.
[14:42] <vila> I meant the one that leads to the lxc failure
[14:42] <jelmer> vila: you want 'quilt-smart-merge=False' for that
[14:42] <vila> okydoky
[14:51] <vila> jelmer: ouch, I missed the ssl.ca_reqs=optional case
[14:52] <vila> jelmer: your second remark is still about the 'optional' case right ?
[15:02] <jelmer> vila: yes
[15:08] <vila> jelmer: checking the lazy option: what the use case for 'optional' ?
[15:09] <vila> sounds like a  feature for python's ssl to allow backwards compatibility by using it as a default value
[15:22] <jelmer> vila: it's useful if you talk to servers that don't actually have ssl certificates
[15:24] <vila> err, you mean a certificate signed by an unknown authority ? I don't think you can setup an https server without any certificate, am I wrong ?
[15:24] <vila> otherwise that'd be an http server ;)
[15:32] <dholbach> jelmer, you rock!
[16:22] <SamB> vila: Well, it *could* just make up a certificate on the spot. (Though that *would* consume a lot of CPU/entropy to no purpose ...)
[16:23] <vila> No, a certificate should be signed, you can't do that on the spot unless you also create the authority on the spot ?
[16:38] <vila> but searching for certificateless ssl returns some results, there is apparently a rather obscure way to do that
[16:40] <gypsymauro> hi
[16:41] <gypsymauro> how can I tell to bzr that my local files are correct and overwrite the modified on the repository?
[16:51] <jelmer> HI gypsymauro
[16:51] <jelmer> gypsymauro:  just commit your local changes?
[17:00] <gypsymauro> jelmer: I've already committed changes form another host.. but I forget to done an update so now I want to take as good just my local changes
[17:01] <jelmer> gypsymauro: merge?
[17:03] <gypsymauro> jelmer: of course, but 'cause I'm sure that local is perfect I just wont to overwrite remote :)
[17:05] <jelmer> gypsymauro: in that case "bzr push --overwrite"?
[17:13] <gypsymauro> jelmer: can I pm you? just for not showing my paths :)
[17:14] <gypsymauro> bzr push --overwrite bzr: ERROR: Working tree ..... has uncommitted changes (See bzr status). Use --no-strict to force the push.
[17:14] <gypsymauro> it's a good idea?
[19:40] <mgz> okay, have a handle on bzr-explorer changes needed for filesystem watcher, just need to split up the code across the different bugs
[22:32] <nDuff> Is authentication.conf expected to work for bzr-svn?
[22:34] <poolie> o/ jam
[22:36] <poolie> jam, what we did vs tla was:
[22:36] <poolie> the default is 'precious' because they might be precious (seems safer though a bit more annoying)
[22:37] <nDuff> ...funny thing is that the logs indicate that it's seeing saved credentials: Obtaining username and password for SVN connection '<https://REDACTED:443> Authorized'(username: 'duffy')
[22:37] <poolie> allow them to be orphaned as a compromise between not losing them and not being annoying
[22:37] <poolie> nDuff, i would guess libsvn will read it
[22:37] <nDuff> ...but below that: PermissionDenied: Permission denied: ".": OPTIONS of 'https://REDACTED/svn/repos/puppet': authorization failed: Could not authenticate to server: rejected Basic challenge
[23:19] <lifeless> nDuff: perhaps something in .subversion/auth ?
[23:21] <nDuff> lifeless: poked at it a bit -- the svn command-line tool is having no trouble using the cached credentials, which implies to me that that configuration is in a good state; still no love from bzr-svn.
[23:46] <jelmer> nDuff: hi
[23:46] <jelmer> nDuff: what platform are you on?
[23:57] <nDuff> jelmer: I'm using the bzr-svn-1.1.0-1 package (with bzr-2.4.1-1ubuntu1) from Linux Mint 12 (x86_64).