[11:43] <sil2100> Hi!
[11:44] <mgz> hey
[11:44] <sil2100> Could anyone help me with a quick question bzr related? Since I remember a configuration option in bzr that disables ssh key authentication - I need it now but can't find it anywhere
[11:44] <sil2100> I remember using it in the past
[11:45] <mgz> I'm guessing you mean you want read-only http access to launchpad rather than bzr+ssh?
[11:46] <sil2100> Yes
[11:47] <mgz> remove launchpad_username from ~/.bazaar/bazaar.conf
[11:47] <sil2100> I have no launchpad_username there ;) Checked that already
[11:47] <sil2100> Ok, I did it differently
[11:48] <sil2100> I simply switched to using http://* directly
[11:48] <sil2100> But thanks!
[11:48] <mgz> really? what does `bzr config launchpad_username` say?
[11:49] <mgz> you can also remove the [Launchpad] section of authentication.conf to be sure, but that should be ignored if the name isn't in the main config
[14:23] <zyga> hey, I'm trying to use bzr join, I have two 2a branches, let's say A and B, I branch B to A/B (so it's inside the A branch), running bzr join B from root of A fails with:
[14:23] <zyga> bzr: ERROR: Cannot join B.  Trees have the same root
[14:23] <zyga> is this expected?
[14:24] <mgz> I take it A is a fork of B? and wow, you really are trying to do things in the most convoluted ways possible.
[14:25] <zyga> no
[14:25] <zyga> A and B are independent
[14:25] <zyga> mgz: what is convulted about join?
[14:26] <mgz> is that actually true? experience says at least one of them is probably going to be something like a fastimport from git.
[14:26] <zyga> it is true (the are independent), and the other part you said is also true (B is an export from git, but it was never derived from A in any way)
[14:26] <zyga> a is lp:checkbox
[14:26] <zyga> b is lp:~zkrynicki/+junk/plainbox
[14:27] <zyga> no common ancestry in any way
[14:27] <zyga> I gave up trying to export the merged tree from git
[14:27] <zyga> as I was unable to fix the bugs in bzr-fastimport and bzr-git
[14:27] <zyga> so I thought that good old bzr join should work
[14:29] <zyga> mgz: is this a bug in bzr core or is the bzr join --help description just confusing and I'm doing something wrong on the basic level
[14:29] <mgz> zyga: as reported, join works fine. what's borked is your branches.
[14:29] <zyga> mgz: how are they borked?
[14:30] <mgz> (probably, I can't get the +junk one)
[14:30] <zyga> oh
[14:30] <zyga> sorry
[14:30] <zyga> typo
[14:31] <zyga> mgz: please try again
[14:33] <jelmer> zyga: if these are fairly old bzr branches they might both have the old constant tree root file id
[14:34] <zyga> jelmer: I don't undestand what that is
[14:34] <zyga> jelmer: checkbox is failry old if you look at the history
[14:34] <mgz> jelmer: one is a fastimport thing, the other is an oldish branch
[14:34] <jelmer> mgz: ah
[14:35] <jelmer> mgz: I guess the fastimport thing should be using rich roots then?
[14:35] <zyga> mgz: both branches use 2a format
[14:36] <zyga> mgz: what is the "constant tree root file id"?
[14:36] <mgz> ...I pretty much think helping you do this is a waste of time, because you're just making your branches more and more complicated...
[14:36] <zyga> mgz: could you explain what I'm doing wrong
[14:37] <mgz> what's the benefit of embedding a the history of the old project in a subdirectory of your new fork?
[14:37] <zyga> mgz: it's a requirement of the team leader, we wish to contiune working on the fork in the old repository
[14:38] <mgz> sure, but why that way around?
[14:38] <mgz> it would be better to just merge your changes into the existing code and not screw around with the structure surely...
[14:38] <zyga> mgz: how would I merge my changes into lp:checkbox?
[14:39] <zyga> mgz: and how is any of that explaining why bzr join is not working, is any of the repositories damaged?
[14:39] <jelmer> mgz: is this a duplicate root file id issue? I think I missed an earlier part of the conversation
[14:39] <mgz> for the record:
[14:39] <mgz> (Pdb) e
[14:39] <mgz> BadSubsumeSource(Can't subsume <WorkingTree6 of /home/ubuntu/tmp/checkbox> into <WorkingTree6 of /home/ubuntu/tmp>. Trees have the same root)
[14:39] <mgz> (Pdb) e.tree.get_root_id()
[14:39] <mgz> 'TREE_ROOT'
[14:39] <mgz> (Pdb) e.other_tree.get_root_id()
[14:39] <mgz> 'TREE_ROOT'
[14:40] <mgz> jelmer: zyga posted to an internal canonical list for reasons passing understanding, so you're missing some context
[14:40] <jelmer> mgz: ah, I'll leave it to you then
[14:40] <zyga> mgz: what I posted to the caonical list was not related much, maybe apart from just trying various things to make this work, the join issue is separate IMHO
[14:40] <jelmer> mgz: (you do know about the merge-into plugin?)
[14:41] <mgz> basically he created a new project on github as a fork of an existing bzr project, added the existing project as a submodule, and is now struggling to reintegrate the codebases
[14:41] <zyga> mgz: you are incorrect
[14:41] <zyga> mgz: plainbox is not a for of checkbox
[14:42] <mgz> zyga: was going to post what I understood from looking briefly and get you to correct anything I didn't understand, maybe I should do that
[14:42] <zyga> mgz: ok
[14:42] <zyga> mgz: we can move the discussion to a proper place if you'd like
[14:42] <zyga> mgz: basically I want to "join" two projects, one of which was in git another in bzr, into bzr
[14:42] <mgz> so, we could make the join work, but I'm unconvinced that's what you really want
[14:43] <zyga> mgz: anything that works is fine, fixing bugs in bzr is good if I can do that as a side effect
[14:44] <zyga> mgz: fixing bugs in bzr-fastimport would be good as well but I've yet unable to trace that issue to something I can report
[14:44] <mgz> can you explain what plainbox is to me briefly? it uses the checkbox code, and is described in the readme as a "plain replacement", but you're saying it's not a fork?
[14:44] <zyga> mgz: ditto to bzr-git
[14:44] <zyga> mgz: it's a rewrite that our team is doing
[14:44] <mgz> bzr-git is simple (I think, you missed pasting the error), you used submodules in your git branch, which aren't supported
[14:44] <zyga> mgz: since checkbox became unmaintainable and the team lead made a decision to rewrite it
[14:45] <mgz> so it is a fork?
[14:45] <zyga> mgz: I see, I didn't know they were unsupported
[14:45] <mgz> or... a redo from scratch?
[14:45] <zyga> mgz: no, it's a rewrite, not a fork
[14:45] <zyga> yes
[14:45] <mgz> then why integrate the histories?
[14:45] <zyga> it uses some of checkbox scripts in practice (hence the submodule)
[14:45] <zyga> but the core is independent
[14:48] <zyga> mgz: so is lp:checkbox broken in any way?
[14:48] <zyga> mgz: or is the plainbox bzr export broken?
[14:49] <mgz> depends what you're trying to do.
[14:50] <zyga> mgz: well, use bzr the way it's advertised so to speak, getting bzr join to work would be a good start
[14:55] <zyga> mgz: ?
[15:00] <mgz> writing an email.
[15:00] <zyga> thanks!
[15:20] <zyga> everyone, for reference, unless something major happens, the most likely bzr branch for plainbox is going to be copied from lp:~zkrynicki/+junk/plainbox
[15:20] <zyga> we'll just need do decide if we want it in lp:plainbox or lp:~something/checkbox/plainbox
[15:23] <zyga> meh
[15:23] <zyga> wrong window :)
[16:19] <mgz> zyga: +metoo bug 1031773 and I'll get round to doing something about that
[16:20] <mgz> your variation on the underlying problem was at least less obnoixious than the one hidetaka got...
[16:20] <zyga> looking
[16:21] <mgz> didn't want to just back the code out, but that's probably the sanest option...
[16:22] <zyga> mgz: osutils.rename calls systems 'mv'?
[16:22] <zyga> mgz: or is some part of python borked to return non-unicode, localized message?
[16:22] <mgz> Python 2 returns non-unicode, localised messages, when setlocale is called
[16:22] <zyga> ahh
[16:24] <mgz> it calls rename(2) underneath
[23:51] <Nav_> Hello. I need some help at installing centralized workflow thing on a CentOS server
[23:51] <Nav_> can someone help?