[00:00] * Peng_ sneaks /away [00:00] visik7: I've looked into having the Bazaar smart server run over jabber [00:01] visik7: Unfortunately it's only possible to offer services to particular people using XMPP, it's not possible to announce that you're running a Bazaar smart server [00:02] lifeless: btw, related to your [RFC] history editing vs history presentation post, did you see the recent blog post from Linus about workflows in Linux ? [00:04] no, url me up please [00:05] http://torvalds-family.blogspot.com/2009/06/happiness-is-warm-scm.html [00:05] lifeless: and in particular this rant that he links: http://www.mail-archive.com/dri-devel@lists.sourceforge.net/msg39091.html [00:08] thanks [00:09] "Don't change history" may not be very novel in the Bazaar world, but as a git user I found that rant quite useful. [00:09] I'm getting more and more convinced that what people really seem to want is well-working looms for in-progress patches [00:09] the main difference for me from the rant is that merging trunk at 'random' points is ok because trunk is always clean [00:11] jelmer, hi [00:15] jelmer: I'm people, and I certainly want that :) [00:18] I'm looking for a project hosting. I'm considering cvs, svn, hg, bzr and git hostings. I know Google Code, Gitorious and Github for now. Any other suggestions? [00:18] RenatoSilva: launchpad.net [00:22] RenatoSilva: you don't really need much to host a DVCS branch, certainly not a bzr one. Just a web server and a ssh account, and you can easily host your work. [00:26] poolie: last time I've checked launckpad I'd have to do complicated stuff to set up my repo, and with some limitations (like I can't put exe files there)... [00:27] RenatoSilva: .....What? [00:27] Yeah, I don't remeber well [00:27] Maybe I was confused [00:27] But how is the set up process today? [00:27] What repos, limitations? [00:28] I'm afraid only bazaar is supported [00:28] RenatoSilva: setup your project; then bzr push lp:~USERNAME/PROJECT/BRANCHNAME [00:28] does it have issue tracker, wiki, and other secondary infra-structure? [00:29] yes no yes [00:31] jelmer: thanks for that link [00:32] lifeless: I'll take a look again, thanks [00:32] Do you guys think LaunchPad is better than BitBucket? [00:33] yes [00:33] why [00:33] RenatoSilva: You probably will get a very biased answer to that question here. [00:34] RenatoSilva: I like it. I have never used BitBucket - so I can't compare. [00:35] well, for starters lp supports bzr. More broadly it has mailing lists, translations and integration with build farms [00:35] Also, I work @ Canonical so my answer is likely biased :). [00:37] random data point - bitbucket has 4000 projects on it apparently, lp has 12000 [00:37] I just don't know why would I want to choose one of bzr, hg, and git rather than other [00:38] I _don't_ work at Canonical, and I'd also recommend launchpad. It's nicely integrated, it's development is reasonably consistent, and the bzr integration is very useful. [00:38] poolie: btw, I'll be sending a few different release related emails out today. [00:38] RenatoSilva: so choose one; it doesn't matter which. If you find it doesn't work well and the community isn't helpful for you, swiwtch. [00:39] poolie: In about 10 minutes I'll be going nose-down on check with the world turned off. [00:39] * SamB sends jelmer a bzr-svn patch [00:39] lifeless: maybe hg, because Moin uses hg, and my projects are moin plugins [00:40] but I'd like to use cvs because I'm already using it at work [00:40] poolie: so if you have stuff you want to chat about, earlier >> later :) [00:40] RenatoSilva: unless you're merging with moin it won't matter that moin is in hg and your plugins aren't. [00:40] is it possible to work with vcs's over http? [00:41] RenatoSilva: using CVS is worse than using any of bzr/hg/git: CVS has many performance issues, few correctness guarantees, is problematic for sysadmins. [00:41] my work deny access to other ports [00:42] RenatoSilva: to varying degrees, yes. Bzr supports read and write over http. However, *launchpad* hasn't deployed the write module for http. [00:43] morning all === igc1 is now known as igc [00:44] lifeless: yeah, I mean commit over http. It would be really nice don't you think [00:46] RenatoSilva: mwhudson here can tell you about that vis-a-vis launchpad ;) [00:46] AFAIK its not in the pipeline at the moment. [00:47] RenatoSilva: you can push to launchpad over ssh - that is quite a common port - If I was you I would check if that is open. [00:47] yeah, would be nice [00:47] no time though [00:48] mwhudson: btw why were you asking about doing reconcile on launchpad? [00:48] lifeless: i thought if it was an overnight job, i could have done it last night [00:48] mwhudson: yes but why are you doing it? [00:48] lifeless: it didn't seem like a bad thing to do [00:49] ah [00:49] its not, just very expensive [00:49] lifeless: since you work at canonical (which country btw?), I'd suggest you wiki for projects documentation [00:49] right [00:49] mwhudson: I'm working on check [a precursor to reconcile] at the moment. [00:49] garyvdm: I will check if ssh is allowed, tahnks [00:49] lifeless: yeah, i know [00:49] RenatoSilva: yes, the developers of launchpad would like to do something along wiki lines [00:50] RenatoSilva: time & design are still up in the air, last that I heard. [00:50] ok, nose grinding time. If anyone needs me, SMS please. [00:51] lifeless: if you're going to use moin, you may want to wait for the 2.0 release [00:52] RenatoSilva: it's pretty unlikely that lifeless is going to be doing the wiki work [00:54] mwhudson: it's just a suggestion to the wiki guys, he could tell them [00:55] where is canonical? London? [00:58] RenatoSilva: Canonical is registered in the Isle of man, It's head office is in London, but it is a virtual company with most of it employees are from all over the world. [00:58] Alot of the bzr developers are from Australia [00:59] RenatoSilva: p.s. I don't work for Canonical. [01:09] (back) [01:09] lifeless: i don't think i want to talk to you urgently [01:11] jml, hi, how did you go with the release? want to talk? [01:13] * RenatoSilva had an energy crash [01:13] poolie: would love to talk. maybe give me 15-30mins to finish up what I'm doing now? [01:15] ok [01:15] igc, how are you doing today? [01:16] hi poolie - ok thanks [01:16] poolie: slept most the weekend so that helps [01:21] oh, BitBucket is paid, agh. [01:22] garyvdm, igc, i'm using explorer [01:22] it's really nice! [01:22] poolie: Great! [01:22] i wish there was a way to get it to open a browser on the branch's launchpad page [01:22] poolie: thanks [01:22] or, generically, it's web page, like lp-open [01:22] or is there? [01:23] poolie: it got zero love over the weekend - too tired sorry [01:23] it's really easy to do :) [01:23] poolie: we could add that pretty easily [01:23] igc: I don't think qbzr would need to be extended. You can just run bzr lp-open [01:24] garyvdm: right === Kissaki is now known as Kissaki^0ff [01:24] igc oh it wasn't a complaint [01:25] spiv: hi, good morning [01:25] spiv, i was wondering on the weekend, where did aaron get up to with the subtree design docs? [01:25] do you know? [01:26] igc, poolie: What I'm planning to do qmain is to have lots of lp related functionality, but have it implemented in a separate plugin. [01:26] *to do for qmain [01:26] BitBucket's HTTP push/pull is commit / checkout over HTTP? [01:26] checkout / commit [01:26] garyvdm: sounds good [01:27] garyvdm: and don't forget that explorer let's plugins add their own panels simply by registering them [01:27] RenatoSilva: with a dvcs, you generally commit to your local repository, and then push that to your central branch [01:27] poolie: I'm not sure; there was a small amount of chatter on the list I think (on a tangentially related thread about "merge --pull") [01:28] garyvdm: ok, but over http? [01:28] garyvdm: in bitbucket I mean [01:29] I don't know [01:29] poolie: I haven't heard much, although I think we committed the version of the doc we had at the end of UDS (in the devnotes branch) [01:29] I don't know if they let you push over http [01:29] I'm sure they let you pull over http though. [01:30] RenatoSilva: I think you need the bzr-webdav plugin to push to http with bzr [01:31] poolie: IIRC we had reached decisions that were acceptable to both of us for all the questions in the "unanswered questions" section, and Aaron was intending on going through all the use cases at the end of the doc and writing the commands to demonstrate how they'd work. [01:31] (and thus verify that the use case was actually satisfied) [01:31] garyvdm: launchpad could include such feature [01:31] jml, as RM, https://code.edge.launchpad.net/~mbp/bzr/doc/+merge/7429 [01:31] ok that'd be good [01:32] poolie: what happened to the '|' line prefix? [01:32] RenatoSilva: In theory - yes. I don't think it would be high on their priority. [01:33] i removed it so you can more easily copy the text [01:33] poolie: it wasn't necessary for formatting then? [01:34] i think it changes it from the equivalent of
to 
[01:34]  RenatoSilva: as mwhudson said re http push > "yeah, would be nice, no time though"
[01:34]  poolie: ok. looks good to me.
[01:35]  garyvdm: you don't need webdav, there's bzr+http
[01:35]  garyvdm: yeah, :(
[01:36]  garyvdm: however would that be hard? It's just about installing the plugin and setting up the http server right?
[01:37]  RenatoSilva: and figuring out how to handle authentication
[01:38]  mwhudson: Ah - yes.
[01:38]  ok, anyway, it's a suggestion
[01:38]  and performance implications
[01:38]  poolie: are you on the pone at the moment?
[01:39]  garyvdm: webdav is not what lp would use
[01:39]  garyvdm: lp would use bzr+http - the smart server.
[01:39]  yes
[01:39]  RenatoSilva: launchpad has scaling to consider; its not the same problem for a hosting site that it would be for a single user wanting commit-over-http.
[01:41]  lifeless: ok
[01:41]  as I said: suggestion
[01:41]  sure. As already discussed its desired but not done yet.
[01:42]  ok
[01:42]  lifeless: BitBucket has this feature
[01:43]  RenatoSilva: like everyone, we have a much longer list of things we'd like to do than we have time to do
[01:43]  RenatoSilva: the things you've been mentioning are already on our list :)
[01:43]  ok thank you
[01:50]  poolie: want to talk?
[01:56]  bbiab
[01:57]  Since bzr is a dvcs, but at work I can only use bzr over http, which is not supported by launchpap, I could do the following: have two repos, one at work, and another at launchpad. When changing my code at work, I commit locally, then I take the repo home, and then I push it to launchpad. Is this ok?
[01:58]  RenatoSilva: you could always push home if you install bzr-webdav for http pushes from work with a post push hook to push to LP :)
[01:59]  thumper: but is my procedure ok too?
[01:59]  * thumper dumps and runs
[01:59]  am I realizing it well?
[01:59]  RenatoSilva: sure
[01:59]  ah ok
[02:00]  thumper: to carry the repo home by email or pendrive is more easy procedure than pushing twice, isn't it?
[02:00]  is it?
[02:00]  thumper: I imagine it's just about copying the files
[02:01]  RenatoSilva: What thumper was saying is that you can set up you home server so that when you push to it, it automatically pushes to launch pad
[02:01]  thumper: keep my PC online is a non-go
[02:01]  garyvdm: ^
[02:01]  garyvdm: oh...
[02:01]  garyvdm: I see
[02:02]  RenatoSilva: ok, in which case a pen-drive repo is a good idea
[02:02]  garyvdm: seems nice
[02:03]  thumper: - I think you would need to use a on tip change hook to do that.
[02:03]  Does a push push the history, or does it work like a "commit" of the local HEAD?
[02:03]  garyvdm: sure, but it'd work
[02:04]  RenatoSilva: it pushes history, yes
[02:04]  RenatoSilva: push pushes all the revisions that the remote repo doesn't know about
[02:04]  SamB: oh, nice then
[02:04]  Now I can forget BitBucket
[02:04]  but if you had any uncommitted changes, it would not push them
[02:05]  SamB: ok, we just have to commit locally before pushing...
[02:05]  which is generally a good thing, unless you lose the working directory ;-)
[02:07]  Is bzr easy to learn? If I read a quick guide (where?) in 20 minutes, would I be able to set up a project in launchpad and do the basics (including the above procedure)?
[02:07]  RenatoSilva: Oh yes
[02:07]  * garyvdm finds url to quickstart guide
[02:08]  bzr in 5 min: http://doc.bazaar-vcs.org/latest/en/mini-tutorial/index.html
[02:08]  what format string should I use for a new bbc format with the 1.16rc for `bzr init-repo --no-trees --2.0?`
[02:09]  garyvdm: thank you, I just found it at the same time. I hope it's enougth
[02:09]  thumper: --2a maybe?
[02:10]  thumper: What do you mean?
[02:11]  is http://bazaar-vcs.org/BzrEclipse usable?
[02:11]  stable?
[02:12]  Peng: I mean that I used to say --development-rich-root
[02:12]  Peng: but now we have an alpha format
[02:12]  Peng: I was wondering what the short name is for the init-repo commadn
[02:13]  RenatoSilva: I believe so. (I don't use Eclipse)
[02:13]  RenatoSilva: you may also want to look at http://bazaar-vcs.org/QBzrEclipse
[02:18]  garyvdm: ok thanks
[02:20]  ok, I'm leaving. Thank you guys for the help. Bye.
[02:59]  poolie: I guess you missed my earlier ping.
[03:00]  hi
[03:00]  was on the phone
[03:00]  talk in 15m?
[03:00]  or maybe just now?
=== garyvdm_ is now known as garyvdm
[03:47]  abentley: please bounce bb when you get a chance
[04:09]  * igc lunch
[04:35]  igc: Please can you subscribe to the qbzr mailing list with you @internode.on.net addresses so that we we have to manually approve messages sent with that address?
[04:35]  *your
[04:35]  + a don
[04:36]  + a don't in there somewhere...
[04:37]  Speaking of qbzr...
[04:37]  I had qlog croaking yesterday...
[04:38]  fullermd: On the latest of lp:qbzr?
[04:38]  jml, when you file bugs, please mark them confirmed and set some reasonable priority
[04:38]  * fullermd checks to see if it's repeatable.
[04:38]  someone else can always override you if they disagree
[04:38]  fullermd: maybe bug 386618
[04:38]  Launchpad bug 386618 in qbzr "[regression] qlog FILE is broken for files from qbzr src (trunk revno.770)" [Critical,Fix committed] https://launchpad.net/bugs/386618
[04:38]  like bug 387117
[04:39]  Launchpad bug 387117 in bzr "ListOption doesn't respect param_name" [Undecided,New] https://launchpad.net/bugs/387117
[04:40]  poolie: ahh ok.
[04:40]  Hm.  Well, it doesn't do it now...
[04:40]  poolie: bzr doesn't use the New status as a drive-to-zero style inbox then?
[04:41]  Yeah, going back to r774 gives what I remember.  So I guess it's fixed  :)
[04:41]  :-)
[04:50]  garyvdm: done
[04:51]  thanks
[05:05]  jml, in practice no
[05:05]  i have mixed feelings about whether we should
[05:06]  however, whatever we're doing with it, there's no point having another roundtrip when a developer files a bug to classify it
[05:06]  any of us including you can make an assessment of a bug we're filing
[05:12]  :D
[05:46]  hi, I want to use http://en.wikipedia.org/wiki/Bazaar_(software) to improve Russian wikipedia page. Can somebody confirm this page is enough up-to-date?
[05:49]  garyvdm: hi
[05:49]  Hi bialix
[05:49]  I'm thinking about your revision selector
[05:50]  your approach is interesting, I'm just want to ask why we don't want to use existing qlog for this?
[05:51]  Yes - that is a option.
[05:51]  e.g. when user need to select revision then we open separate qlog-like window and allow to select revision or range of revisions there?
[05:51]  igc also had some ideas regarding this.
[05:51]  we can remove some elements, e.g. search
[05:51]  * bialix looks
[05:52]  do you have bug number by hands?
[05:52]  I'm not sure If loged a bug - let me see
[05:52]  poolie: around?
[05:52]  garyvdm: it's indeed files
[05:53]  filed
[05:53]  Bug #328598
[05:53]  Launchpad bug 328598 in qbzr "Create a revisions selector widget. Use in commands like push/pull." [Medium,In progress] https://launchpad.net/bugs/328598
[05:54]  well, using igc suggestion it's really make sense to create separate window to select
[05:55]  while I'm not sure about date
[05:55]  So whether we display a dialog or a custom widget in a combo box does not mater too much
[05:55]  why somebody need date when there is full log?
[05:55]  The latter is easer.
[05:55]  um, no the difference is huge
[05:56]  qlog-like dialog allow you to browse revisions, see their details and get the diff
[05:56]  everything like in qlog
[05:56]  IIUC combobox only allow you to see summary for revisions
[05:56]  He want's the date option so we cater for people who have branches that take a very long time to load like mysql/ooo
[05:57]  bialix: yes - a dialog gives a more space to work with.
[05:58]  ah, this makes sense for Date then
[05:59]  I'm thinking in my mind: tabs at the top to select Revno (log) | Date | Tag | Other (branch:, submit: etc)
[06:00]  garyvdm: tabs are evil :-/
[06:00]  such construct is better suited for separate dialog IMO
[06:00]  pygi: not in this case
[06:00]  LineEdit at the bottom to the left of the ok button - which shows you what you have selected/can edit manualy.
[06:01]  Hi pygi
[06:01]  bialix: do you have any gdb skills
[06:01]  LineEdit is OK
[06:01]  a bit
[06:01]  hi bialix, pygi, garyvdm
[06:01]  hi igc
[06:01]  glad to see you
[06:01]  I hope you're better
[06:01]  bialix: I am thanks
[06:02]  igc, you send some mails from you internode account
[06:02]  this address is not subscribed to qbzr ml
[06:02]  bialix: it should be now
[06:02]  I've tried to subscribe you directly but can't
[06:02]  ok
[06:02]  bialix: ha ha - I asked igc to subscribe that acocunt just now :-)
[06:02]  thanks to garyvdm pointing that out earlier :-)
[06:03]  garyvdm: :-D
[06:03]  igc, can you confirm http://en.wikipedia.org/wiki/Bazaar_(software) this page is good enough to use for translation to Russian?
[06:04]  we desperately need to improve Russian wiki page
[06:04]  it's hard to write something original in wikipedia, so I've decided to just translate English one
[06:04]  bialix: I'm getting a segfault with some code I have written. I can't figure out how to get gdb to give me a stack trace.
[06:05]  segfault or traceback?
[06:05]  bialix: I can spot 1 error - I think bzr-git is more advanced than it claims.
[06:06]  hmmm
[06:06]  I'm getting a segfault - and I want to get a trackback for that segfault
[06:06]  I'm not sure I understand correctly
[06:06]  usually when you get segfault the python died, is not?
[06:06]  Yes \
[06:06]  so you can't get traceback
[06:07]  Apparently you can with gdb - but I can't figure out how.
[06:07]  oh
[06:07]  sorry
[06:07]  I've misread first time
[06:07]  gdb, not pdb
[06:08]  no, sorry
[06:08]  Ok - never mind - I should probably go and read the manuals....
[06:08]  my linux skills is rather small
[06:08]  bialix: that page looks ok to me
[06:08]  so I'll translate it, thanks igc
[06:08]  The code that is segfaulting: lp:~garyvdm/qbzr/trees - run qbrowse - expand a folder.
[06:09]  * bialix looks
[06:09]  igc, do you have experience in building deb packages for PPA?
[06:10]  bialix: no sorry
[06:10]  igc: ping reminder on command hooks
[06:10]  garyvdm: segfault on windows too
[06:11]  you did something bad with PyQt I guess
[06:11]  lifeless: sure. next on my list after looking at an eol bug report
[06:11]  usually to segfault python one have to deal with C-extansions
[06:12]  maybe incorrect usage of malloc/free
[06:12]  maybe you try to read data that was deleted from memory
[06:12]  or outside bounds of array
[06:13]  garyvdm: thanks for fixing regressions, I'm unblocked now
[06:14]  we need to think hard about unittesting
[06:14]  bialix: sorry about those.
[06:14]  I've used another branch to commit, in the end we have DVCS!
[06:15]  Yes re unit tests
[06:16]  There is just so many to be written - It is quite daunting.
[06:16]  yes, a bit
[06:16]  how to eat the elephant?
[06:17]  I spent about 3 days writing tests for loggraphprovider, and I only covered the first 2 methods :-(
[06:17]  this is disapponting, I know
[06:17]  Ok - so I very new a writing tests.
[06:17]  ask vila for help
[06:18]  vila helped me alot at uds!
[06:18]  vila likes tests
[06:18]  * bialix summons vila!
[06:18]  nope, does not work
[06:18]  ok, Gary I have to go to work, see you later
[06:18]  Ok - bye
[06:37]  lifeless: I'll get on to that Command.hooks re-review now
[06:52]  in CVS each dir is a module. In Bazaar, each dir is a branch?
[06:55]  no, branch is a directory
[06:55]  then yes, don't?
[06:56]  are sub dirs of the branch branches themselves?
[06:56]  I meant to point out that it's the other way around :)
[06:56]  sub dirs of a CVS module are modules themselves.
[06:57]  so no, not every directory is a branch
[06:57]  but every branch is a directory
[06:57]  bzr is nothing like CVS :)
[06:57]  what's a module / project in CVS repos? It's just a dir in the repo tree. Is this the same with Bazaar or other tools?
[06:57]  luks: ok I've got it
[06:57]  there isn't anything like that
[06:58]  in bzr you generally work with the filesystem
[06:58]  so if you want to group branches, you can use a directory
[06:58]  it seems that branches are like the CVS repos themselves, not their modules
[06:58]  you can't really compare cvs and bzr
[06:59]  ok
[07:11]  hi all
[07:25]  hi vila
[07:26]  hi poolie !
[07:26]  igc: I replied
[07:26]  igc: one more from you should do it
[07:54]  hi vila!
[07:54]  igc: ping
[07:55]  igc: See above
[07:55]  lifeless: looking now
[07:55]  danke
[08:10]  EODing
[08:11]  poolie: check is refactored to ~ where I want it; tuning somewhat now - reconcile's core needs fixing too, and the performance is odd - tracking down details now.
[08:17]  night lifeless
[08:22]  where is the output files for bzr init?
[08:23]  where are
=== abentley1 is now known as abentley
[08:28]  oh bzr is smart, .bzr is /ah under windows
[08:28]  is there any problem if I remove the hidden gflag?
[08:29]  flag
[08:51]  RenatoSilva: I don't think that will be a problem, but if you encounter one: 1) Restore it, 2) File a bug
[09:02]  good morning bzr experts
[09:04]  Spabby: hi
[09:06]  vila, i'm writing up a patch for the bug process stuff
[09:07]  poolie: cool
[09:13]  can anyone help me install the automirror plugin please? I have followed the instructions for installing plugins on the bzr website and also in the readme but I cannot get it working!
[09:18]  Spabby: can you elaborate ? Note that I don't know the plugin, but we'll see
[09:20]  yes
[09:20]  one second sorry
[09:21]  firstly it says to install to ~/.bazaar/plugins and the directory didn't exist
[09:21]  so I created the directory
[09:21]  and copied the automirror directory there
[09:21]  I then added the post_commit_mirror parameter to the branch.conf file
[09:22]  but when I commit nothing happens
[09:22]  Spabby: do you have bzr installed on the remote server ?
[09:22]  yep
[09:23]  if I commit to the remote server it works fine
[09:23]  and if I do a bzr update on the remote server that also works great
[09:23]  Spabby: more importantly (as far as I understand the plugin intent), do you have a remote branch there ?
[09:23]  yes
[09:23]  did you look at your .bzr.log file ?
[09:23]  ah
[09:23]  no
[09:23]  is that in the .bzr dir?
[09:23]  'bzr version' will tell you where it is
[09:25]  ah
[09:26]  I suspect the plugin is in the wrong place ;)
[09:27]  no it looks right
[09:27]  should it list plugins installed after "looking for plugins" line?
[09:32]  what's this output in $bzr launchpád-login?
[09:32]  [-                   ] https >      0KB     0KB/s |
[09:33]  jml: argh; could you please make sure the RM notes make it clear to preserve NEWS headings?
[09:33]  Spabby: 'bzr plugins -v' should answer that
[09:33]  * igc dinner
[09:33]  jml: otherwise everyone landing stuff collides for the first period after; and hilarity *doesn't* ensue.
[09:34]  RenatoSilva: it's an artifact of the progress reporting, I think there is already a related bug on launchpad.net for it, but may be not for the launchpad-login command
[09:34]  super duper the plugin is installed
[09:35]  right so it must be the config that is borked
[09:35]  vila: I only receive this output, and push is not working
[09:35]  I don't know if I'm authenticated or not
[09:35]  RenatoSilva: Are you behind a firewall ?
[09:36]  I'm at home
[09:36]  hum let me see
[09:36]  RenatoSilva: some homes come with firewalls :)
[09:37]  on the windows :)
[09:40]  this is frustrating the hell out of me, it should be so simple!
[09:40]  vila: what is the exact process to allow
[09:42]  bzr.exe? pageant.exe? putty.exe?
[09:42]  RenatoSilva: Ha :-/  Relevant bug is #186920 I think
[09:42]  ubottu: bug #186920 ?
[09:42]  can I update a remote branch using bzr?
[09:42]  Launchpad bug 186920 in python "bzr launchpad does not handle proxy when used for name resolution " [Unknown,Confirmed] https://launchpad.net/bugs/186920
[09:42]  please?
[09:43]  Spabby: As long as you're not using lp, yes
[09:43]  :-/
[09:43]  i mean I can ssh in and do an bzr update
[09:43]  but can I do it locally or do I have to do that
[09:43]  vila: what is the process to allow? pageant? bzr? putty?
[09:44]  Eerk, sorry guys, Spabby: forget my last message
[09:45]  RenatoSilva: sorry, I meant proxy, not firewall
[09:45]  ok I have written a bash script that will commit changes and update remote host, it will do!
[09:46]  No proxy. I deactivated firewall and it didn't worked
[09:46]  Spabby: you might like to try the push-and-update plugin
[09:46]  spiv, I'll be honest I don't understand the difference between "push" and "commit"
[09:46]  Spabby: or perhaps the bzr-upload plugin
[09:47]  Spabby: https://launchpad.net/bzr-push-and-update and https://launchpad.net/bzr-upload
[09:47]  RenatoSilva: what does your .bzr.log file say ?
[09:48]  Spabby: commit creates a new revision, push publishes it
[09:48]  Spabby: commit adds a new revision to a branch with the contents of your working copy on disk.  push updates one branch from another branch.
[09:48]  vila: where?
[09:48]  ok so I definitely want to be using commit
[09:48]  RenatoSilva: 'bzr version' will tell you
[09:48]  as I want to be working locally until a module is complete then committing it to the central repo
[09:49]  vila: is this log cleanned from time to time btw?
[09:49]  RenatoSilva: yes
[09:50]  RenatoSilva: renamed to .bzr.log.old to be precise, and .bzr.log.old is deleted if present at that time
[09:51]  for lp login: failed to import pycurl: No module named pycurl
[09:51]  failed to instantiate transport  for 'https://launchpad.net/': DependencyNotPresent(Unable to import library "pycurl": No module named pycurl)
[09:51]  RenatoSilva: shouldn't be a problem, what bzr version are you using ?
[09:52]  1.15-2
[09:52]  RenatoSilva: that's strange anyway, I thought the windows installer included pycurl
[09:52]  RenatoSilva: how did you install ?
[09:53]  for push: http://pastie.org/512276
[09:53]  vila: hum, 1.15-2 is a self-extracting zip
[09:54]  RenatoSilva: is this still about 'command-line: line 0: Bad configuration option: ClearAllForwardings
[09:54]  ?
[09:54]  y
[09:55]  RenatoSilva: then fix your openssh config :P
[09:55]  although bzr might be passing that option in
[09:56]  RenatoSilva: according to the log you pasted, openssh is used, not putty/pageant
[09:56]  how do I see the ssh client?
[09:56]  set
[09:57]  Yes, bzr will pass that option in if it is spawning openssh.
[09:57]  it seems that I need some python dependecies, pucurl etc
[09:57]  pyculr
[09:57]  pucurl
[09:57]  pycurl is an optional dependency.
[09:57]  pycurl!
[09:57]  RenatoSilva: no, you can ignore pycurl
[09:57]  not bzr related at all but can you remember the name of the scripting software that allows you to basically record your command line as a macro?
[09:57]  mental block
[09:57]  paramiko and pycrypto too?
[09:58]  its ok I remember, it's expect
[09:58]  RenatoSilva: those two are pretty essential for ssh
[09:59]  maybe I should try the installer 1.15-1 instead of my 1.15-2
[09:59]  IIRC paramiko is only necessary for SFTP (as the actual SSH bits can be done with openssh/putty/whatever).
[10:00]  And IIRC pycrypto is only necessary for paramiko.
[10:00]  I'm curious about which version of OpenSSH you have that doesn't recognise the ClearAllForwardings option, though.
[10:01]  how do I set the SSH client?
[10:01]  spiv: it may be an ssh.com version or some such
[10:01]  lifeless: right
[10:01]  RenatoSilva: BZR_SSH=plink
[10:01]  lifeless: (although we try to detect that too, IIRC)
[10:02]  RenatoSilva: environment variable
[10:02]  LarstiQ: shouldn't it be the full path?
[10:02]  RenatoSilva: set the BZR_SSH environment variable to one of paramiko, openssh, ssh, plink.
[10:02]  RenatoSilva: no, just the type (paramiko, openssh, ssh, plink)
[10:04]  RenatoSilva: I'm curious, what does "ssh -V" report?
[10:04]  LarstiQ: bzr: ERROR: [Error 2] O sistema nÒo pode encontrar o arquivo especificado (file not found)
[10:04]  RenatoSilva: I guess plink.exe isn't on your %PATH% then.
[10:05]  spiv: I only have ssh because of MSYS: OpenSSH_2.9p2, SSH protocols 1.5/2.0, OpenSSL 0x0090602f
[10:05]  2.9!?
[10:05]  spiv: that's why I said: I need the full path!
[10:05]  Ah, that is a pretty damn old OpenSSH
[10:06]  RenatoSilva: why is setting %PATH% harder than setting %BZR_SSH%?
[10:07]  spiv: why is bzr_ssh not in `bzr help configuration`?
[10:07]  (I mean, it would be nice if it were possible to specify a full path, but that feature doesn't exist yet.  Patches welcome!)
[10:07]  LarstiQ: a bug, I guess.
[10:08]  spiv: yes :P
[10:08]  spiv: I set the apth
[10:08]  LarstiQ: also, for no good reason (probably just historical) it doesn't get looked up by the regular config machinery, i.e. you can't set it in bazaar.conf
[10:08]  * LarstiQ nods
[10:08]  The server's host key is not cached in the registry. You have no guarantee that the server is the computer you think it is. The server's rsa2 key fingerprint is: ssh-rsa 1024 ******.
[10:08]  Connection abandoned. bzr: ERROR: Connection closed: please check connectivity and permissions
[10:08]  LarstiQ: it's just environment or autodetect or default. (in that order)
[10:09]  RenatoSilva: do you have access to server logs? Are you using the correct user? Is the key you supplied allowed to login? etc
[10:09]  RenatoSilva: so perhaps all that's needed is for you to manually do "ssh bazaar.launchpad.net" and accept the host key so it gets saved in the registry.
[10:12]  I opened a session with PuTTY
[10:13]  bzr: ERROR: Target directory lp:~renatosilva/+junk/solenoid-moin-theme already exists, but does not have a valid .bzr directory. Supply --use-existing-dir to push there anyway.
[10:13]  it's better error now :)
[10:14]  ooohhhh \o/ it's working
[10:14]  RenatoSilva: cool
[10:14]  spiv: I had to store the server keys
[10:17]  my first branch \o/
[10:17]  https://code.launchpad.net/~renatosilva/+junk/solenoid-moin-theme
[10:17]  thank you guys, you are _so_ helpful
[10:20]  btw: http://moinmo.in/ThemeMarket/Solenoid
[10:32]  http://bazaar.launchpad.net/~renatosilva/%2Bjunk/solenoid-moin-theme/annotate/head%3A/htdocs/css/msie/style_ie8.css
[10:32]  why does it show the 1st line only?
[10:34]  thats very odd
[10:34]  could you please file a bug at https://bugs.launchpad.net/loggerhead ?
[10:36]  lifeless: it seems that for some reason the file has Mac-style  line breaks
[10:37]  RenatoSilva: mention that in the bug report :)
[10:38]  spiv: this is a dump of the file "/* Width was too short */\rdiv.wrapper
[10:40]  RenatoSilva: I can see the file content already... a bug report is what's needed at this point.
[10:44]  spiv: https://bugs.launchpad.net/loggerhead/+bug/387225
[10:44]  Launchpad bug 387225 in loggerhead "File content not entirely displayed" [Undecided,New]
[10:44]  I will commit the windows-line-break version
[10:45]  how do I link to the bug?: the complete URL?
[10:46]  spiv: http://bazaar.launchpad.net/%7Erenatosilva/%2Bjunk/solenoid-moin-theme/revision/1 does the file list look like it has something odd at the top to you?
[10:46]  RenatoSilva: link to it where
[10:47]  in the commit comment
[10:48]  This change is because of this bug: xxx
[10:48]  however you like :). you can also say '--fixes lp:' in the command line when you commit
[10:49]  in eclipse you can put just 'bug 123', when shown in bugzilla a link is created to that bug
[10:49]  Launchpad bug 123 in rosetta "There's no direct way to see the project info when translating it" [Medium,Fix released] https://launchpad.net/bugs/123
[10:49]  lifeless: the commit doesn't fix, but is caused by the bug :)
[10:49]   
[10:49]  RenatoSilva: oh right. In Ubuntu the convention is
[10:50]  LP: #
[10:50]  who told ubottu to write that?
[10:51]  lifeless: yes
[10:51]  bug 12345
[10:51]  Launchpad bug 12345 in isdnutils "isdn does not work, fritz avm (pnp?)" [Medium,Fix released] https://launchpad.net/bugs/12345
[10:51]  oh
[10:51]  RenatoSilva: if you say "bug " in a bug comment then it is linked correctly, perhaps loggerhead could do that too
[10:51]  in fact, I think it's done for the comments that LP shows on the branch page, it's just because loggerhead is a bit separate that it doesn't currently
[10:52]  is a bug id unique among the projects?
[10:52]  in launchpad, yes
[10:52]  lifeless: there's an empty 
  •  that's causing it, I think.
[10:52]  the root perhaps
[10:52]  anyhow bug sent
[10:52]  Yeah.
[10:53]  LarstiQ: so, know where jelmer is?
[10:53]  lifeless: in fact almost certainly the root, judging from how it renders other directories.
[10:55]  james_w: so I can just type exactly "Thsi commit is because of bug 387225."
[10:55]  Launchpad bug 387225 in loggerhead "File content not entirely displayed" [Undecided,New] https://launchpad.net/bugs/387225
[10:55]  lifeless: "/" shows along with other files as something you can click & expand to see changes.
[10:55]  RenatoSilva: yeah, it won't do anything yet, but it probably should :-)
[10:55]  Anyway, that's enough internet for the evening...
[10:56]  RenatoSilva: you may also want to use the "--fixes lp:" thing, as that tells LP to link the branch and the bug
[10:56]  actually, "Launchapd bug"
[10:56]  james_w: the branch is not a fix
[10:56]  ok, then don't :-)
[11:03]  spiv: bug 387227 if you want to comment
[11:03]  Launchpad bug 387227 in loggerhead "odd rendering of files list" [Undecided,New] https://launchpad.net/bugs/387227
[11:07]  hwo to store a default merge location?
=== bialix_ is now known as bialix
[11:09]  RenatoSilva:  --remeber.
[11:09]  Is this stored under my prefs or under the branch?
[11:09]  tiased: the branch
[11:14]  * RenatoSilva read the 5-min tutorial in a few hours o.O
[11:28]  thank you guys, I'm leaving, bye.
[11:28]  ciao
[11:30]  lifeless: no, last time I saw him was previous saturday in Amsterdam. I have noticed him being on irc less lately.
[11:48]  hello.
[11:49]  Is it me or the PPA has not been updated for bzr 1.15.1?
[11:49]  I see 1.14.1 and 1.15 only
[11:54]  you are right. The beta ppa has 1.16-rc1
[11:55]  How odd.  the MD5 for the 1.16rc1 tarball is a dead link...
[11:57]  And bzrtools too.
[12:37]  igc: ping
[12:37]  igc: is this branch still needed? https://code.launchpad.net/~ian-clatworthy/qbzr/explorer-prototype
[12:37]  bialix: no
[12:38]  mark it as abandoned please
[12:38]  bialix: I can delete it if you like
[12:38]  bialix: ok
[12:39]  thx
[12:39]  btw, qversion in the trunk
[12:41]  luks: I see you have uploaded packages for qbzr 0.11, thank you. We need packages for 0.12 too, because it's version compatible with bzr 1.16
[13:10]  Hi I am having some problems with some scripting with bazaar
[13:11]  I am making a PHP script which merges in changes from one branch to another
[13:11]  but exec, shell_exec and friends don't return any script output - even though when I execute these commands in my shell they actually return output
[13:11]  How can I get the output out of these commands?
=== TheJosh is now known as TheJosh1337
[13:13]  it may be that the output you're looking for is on stderr
[13:13]  As is customary, the PHP manual tells you everything but what you care about, but AFAIK all those commands give you stdout, but not stderr.
[13:13]  most chatter is on stderr
[13:13]  And in some cases, bzr might give you different output if it's not run on a terminal, I'm not sure.
[13:14]  so why does bzr output to stderr? Isn't stderr for, like errors?
[13:14]  and can you trick bzr into thinking it is a terminal? set TERM env or something?
[13:14]  stderr is for output that isn't the main output
[13:15]  Try using proc_open().
[13:15]  When I run the command with 2>&1
[13:15]  I get the output, but I get other junk too
[13:15]  No handlers could be found for logger "bzr"
[13:15]  [27093] 2009-06-15 21:38:58.811 INFO: Nothing to do.
[13:15]  Nothing to do.
[13:15]  (of course, like most PHP executing functions, that still calls a shell for the execution.  Sigh.  A continuing irritatnt...)
[13:16]  thats very odd output
[13:16]  what is this 'logger "bzr" stuff'. I don't want that
[13:17]  I guess I can regex it out but that's a bit crappy
[13:17]  we use python's logging facilities to do output
[13:17]  something is causing bzr's setup routines not to run
[13:18]  or it may be related to trminal detection (but even then we should setup a logger)
[13:18]  you can fool bzr yes, it honours TERM etc
[13:19]  so If I set TERM=bash, that might fix it?
[13:20]  * fullermd doesn't see that in his termcap...
=== mrevell is now known as mrevell-lunch
[13:21]  hmm
[13:22]  TERM=xterm
[13:22]  that will fool it
[13:22]  Something like vt100 is usually the LCD, but yah, that'll work fine.
[13:22]  Probably doesn't matter too much unless you get exotic.
[13:24]  setenv('TERM=vt100') doesn't fix it. I'll have to try with proc_open which allows setting env vars
[13:25]  I just upgraded a bzr.dev in format "pack-0.92" at revno 3638 to "1.14" and 76.7 MB later pulled to revno 4441. And it completed. How about that.
[13:25]  I'd guess the std{err,out} issues are more likely to be your issue than terminal detection, though.  I'd try with just that first.
=== ddaa1 is now known as ddaa
[13:36]  I changed my code to use proc_open and now it doesn't return anything at all
[13:36]  can I set the logger somehow, so it doesn't throw the stupid messages?
[13:38]  TheJosh1337: file a bug about the messages please. Also what version of bzr are you using?
[13:38]  2.6.21
[13:38]  my bad
[13:38]  1.6.1
[13:39]  that's quite an old version isn't it.
[13:39]  Fairly, yah.
[13:39]  I am running Ubuntu Intrepid
[13:40]  I guess its time to upgrade to Jaunty
[13:42]  well im off then - Have an upgrade to do
[13:42]  cya
=== mrevell-lunch is now known as mrevell
=== KX is now known as MemoBot
=== ddaa1 is now known as ddaa
=== MemoBot is now known as EmoBot
=== EmoBot is now known as MemoBot
[13:59]  How do I change the push branch listed in `bzr info`?
[13:59]  bzr push --remember 
[14:01]  thanks
[14:05]  lifeless: is there another way to set the push branch location?
[14:05]  gioele: $ vi .bzr/branch/branch.conf
[14:08]  gioele: bzr help configuration
[14:11]  lifeless: ah, great. I'm adding that to the Location wiki page
[14:13]  difference between the public and the submit branch? Both seem to be used by the 'bzr send' command
[14:13]  Y'know, it may be time to stop having a separate section in there for branch6...
[14:35]  jelmer: ping, unless we chat, I'm afraid we will stay out of phase :-) I'm sure we're close to agreeing though :)
[14:35]  vila: hi!
[14:36]  vila: Did you see my last reply?
[14:39]  Does bazaar always retrieve full history of a branch, when I branch from it?
[14:39]  cyberixae: if you don't specify --stacked, yes
[14:41]  jelmer: well, if he creates a new branch in the same repository, there is no need for it to do any fetching, is there ?
[14:41]  "The new branch will depend on the availability of the source branch for all operations."
[14:41]  "all operations"?
[14:41]  jelmer: got my patch, btw?
[14:41]  What is the point of having one then?
[14:41]  SamB: yep, it's already in the 0.6 branch
[14:41]  SamB: thanks
[14:41]  no problem
[14:42]  it was an easy fix
[14:42]  Look, our project has some big files in the main branch and I was considering to move them out of the way into another branch and deleting them from the main branch.
[14:43]  But this is useless, if the history is downloaded everytime anyway
[14:43]  Is this correct?
[14:43]  well, not quite useless
[14:43]  at least you wouldn't have to have them in the working directory then!
[14:44]  but yeah, they'd still get downloaded
[14:44]  the speed of download and amount of disk space were the concerns here.
[14:45]  well, not having it in the working directory would obviously save some disk space
[14:45]  but not really any download
[14:45]  why?
[14:45]  the version history is compressed?
[14:45]  well, probably
[14:46]  But the files are compressed already anyway
[14:46]  but what I mean is that instead of having the copy in the history *and* the one in the directory, you'd just have the one in the history
[14:46]  we have some jpegs, mp3s and oggs in there
[14:46]  ah.
[14:46]  if this were #git, I might be suggesting "git filter-branch"
[14:47]  ... does anyone know of something like that for bzr?
[14:52]  oh great masters of [bzrlib] unittesting! can you say me how to test NotBranchError?
[14:53]  SamB: you mean fast-import-filter
[14:55]  bialix: test what about it?
[14:55]  bialix: btw, any news on your bzr-git windows porting effort?
[14:55]  jelmer: I'm writing helper method for qbzr to open working tree and show errors NotBranchError/NoWorkingTree errors in GUI
[14:56]  IIRC there is fake tree created at temp dir root
[14:56]  so how can I force NotBranchError in my unit test?
[14:56]  can a branch contain anotehr branches?
[14:57]  bialix: there shouldn't be a fake tree created at the temp dir root
[14:57]  jelmer: re bzr-git: it means your bzr<->git roundtripping means for me that I should wait until each my patch lands before I can start to work on next one
[14:57]  bialix: the bzr testsuite relies on that not being the case
[14:58]  bialix: I think you mean *lack* of roundtripping, with roundtripping there wouldn't be a problem :-)
[14:58]  I mean rebasing
[14:58]  bialix: You can send me multiple patches in one go though, as long as you don't have any unpublished work based on the stuff you've sent me
[14:58]  bialix: bug 387241 is fixed now
[14:58]  Launchpad bug 387241 in bzr-explorer "bzr explore fails to open branch with out-of-date working tree" [High,Fix released] https://launchpad.net/bugs/387241
[14:59]  can I tag a previous revision?
[14:59]  bialix: but qgetupdates doesn't work
[15:00]  jelmer: you've rejected my test script, it was a little blocker for me. and then I digressed on other things
[15:00]  bialix: there's a radio button to refresh the working tree, it runs, it completes but the OutOfDateTree exception doesn't go away
[15:00]  igc: hm?
[15:00]  gile a bug?
[15:00]  igc: I'm about to finish simpler qupdate dialog
[15:00]  igc: qile a bug ? ;-)
[15:01]  :-P
[15:01]  s/gile/file/
[15:01]  bialix: shall do. why qupdate when qgetupdates already exists?
[15:01]  qgetupdates is too much for this IMO
[15:01]  it's a swiss knife
[15:02]  what does not work actually?
[15:02]  qgetupdates seems to update the *branch* but not the working tree
[15:02]  jelmer: how you mock GTK widgets for testing?
[15:02]  igc: there is radio button with 2 variants
[15:03]  second variant should update wt
[15:03]  bialix: I think we basically create them but don't call show
[15:03]  bialix: maybe qgetupdate just needs more options put into an extension panel
[15:03]  bialix: but jam should be the expert on that, he did the testing stuff
[15:03]  igc: qgetupdates and qgetnew designed by Mark Hammond
[15:03]  bialix: anyhow, I'll file a bug
[15:03]  * SamB wants "bzr clog"
[15:03]  clog for what?
[15:04]  jelmer: did you see my email about saved commit data?
[15:04]  bialix: color log
[15:04]  ah
[15:08]  * bialix reads about unittesting in Qt but found QtTest approach a bit limited
[15:11]  igc: q about English
[15:11]  igc: you're using Adds, Modifies in bzr explorer view
[15:11]  why not Added, Modified?
[15:11]  just curious
[15:12]  I want to understand difference
[15:13]  bialix: it should be Added, Modified - I'll tweak
[15:13]  No need to get tense about it   8-}
[15:16]  I'm new to bzr. I tagged a specific revision, but when I commit or push, it stands that there's nothing to change. Why?
[15:16]  Well, you don't commit tags, so that's half the answer.
[15:16]  The other half is that push doesn't say when it pushes tags.  I think it just does anyway, but it may not if there aren't revisions to push too; can't remember.
[15:17]  fullermd: that's why I also tried to push
[15:17]  fullermd: it's kind of bug isn't it? for example...
[15:18]  fullermd: I pushed 2 revisions, but 'forget' to tag rev 1 as 'RC_2009.6.14'
[15:19]  fullermd: then I pulled the branch, tagged revision 1, and then I want the local and remote branch (in launchpad) to contain the new tag
[15:19]  fullermd: maybe forcing the commit? how about it?
[15:20]  igc: btw, I'm working with bzr explorer all day today. It's very nice shaping
[15:20]  There's no commit.  Tags are on a different timeline.
[15:20]  push probably _should_ push the tag anyway, even if there are no new revisions.  It may; I don't remember.
[15:20]  Did you check?
[15:20]  bialix: nice to hear
[15:20]  I need qswitch very soon otherwise I can't work with my plugins
[15:20]  I'm using lightweight checkouts for plugins
[15:20]  * igc does his fix commit and push using explorer --gtk
[15:22]  fullermd: how do I browse tags?
[15:22]  s/fix/first/
[15:22]  RenatoSilva: `bzr tags [$LOCATION]`
[15:22]  fullermd: oh, bzr log
[15:22]  must be time for some sleep :-)
[15:22]  igc: Well, hopefully it was a fix too   ;)
[15:23]  fullermd: I guess it was :-)
[15:24]  igc: look at new qupdate from bzr+ssh://bazaar.launchpad.net/~qbzr-dev/qbzr/qupdate/
[15:25]  fullermd: I pulled it onto a new dir, and the tags is there...thanks
[15:26]  I just can't see it trhough launchpad
[15:26]  bzr tags -d lp:your-lp-url-here
[15:27]  LP may have the tags cached, I dunno.  But if they show up in a new pull, they're there, so it'll work out.
[15:28]  bialix: as I've pulled and it had the tag, such command should work
[15:28]  bialix: and works
[15:28]  bialix: I can't see it through the web interface tough
[15:29]  That may be a LP bug.  Might try getting somebody there to check.
[15:30]  a loggerhead bug maybe
[15:33]  fullermd: https://bugs.launchpad.net/loggerhead/+bug/246739
[15:33]  Launchpad bug 246739 in loggerhead "tags are not available" [Wishlist,Triaged]
[15:35]  bialix: I'll took a look tomorrow sorry
[15:35]  night all
[15:36]  ok, night
[15:36]  I have a patch in the middle of my tree that I'd like to revert.  Is there any way to just throw away that patch?  Or even better, just mark the patch as thrown away?
[15:38]  * RenatoSilva found weird that he could push the tag without launchpad-login
[15:39]  balor: shelve?
[15:41]  bialix: That seems to shelve all changes since the revision in the middle of the tree.  I just want to remove one revision from the revision history.
[15:41]  oh, you can't
[15:41]  bialix: thanks
[15:41]  you need to reqrite entire history
[15:41]  reqrite
[15:42]  rewrite
[15:42]  rats
[15:42]  :)
[15:42]  you can look at rebase plugin
[15:42]  So branch at -r n-1 and then replay -r n+1
[15:42]  but old and new history will be incompatible after replay
[15:42]  yes, something like that
[15:46]  can I change commit comments?
[15:46]  after commit?
[15:47]  yes
[15:49]  you can't
[15:49]  history is immutab;e
[15:49]  immutable
[15:49]  you only can uncommit and then commit again with new message
[15:50]  yes, thant's ok
[15:50]  bialix: uncommit -r 10 for a range of 1..20 will uncommit all the 11..20 commits right?
[15:51]  look at `bzr uncommit -h`
[15:51]  it doesn't mention that
[15:52]    If --revision is specified, uncommit revisions to leave the branch at the
[15:52]    specified revision.  For example, "bzr uncommit -r 15" will leave the
[15:52]    branch at revision 15.
[15:52]  my mistake sorry
[15:52]  then it's like I tould you
[15:53]  vila: ping
[15:54]  bialix: pong
[15:55]  is there any way I can trigger NotBranchError in the tests?
[15:56]  raise NotBranchErro ? I think I don't understand the question, that's too obvious >-/
[15:56]  ok, lets pastebin first
[15:57]  http://pastebin.com/m1f001ece
[15:57]  this test fails because error is not raised
[15:58]  right
[15:58]  here is open_tree http://pastebin.com/m695017ea
[15:58]  why?
[16:00]  because there is a workingtree ?
[16:00]  in temp root?
[16:00]  what should I do then?
[16:00]  bialix: try checking what WorkingTree.open_containing(directory)[0] is returning
[16:01]  1sec
[16:01]  bialix: in temp root ? that's strange, but put a breakpoint before open_containing and look there
[16:01]  or better put the breakpoint in your test and do 'pp self.test_dir' and look there
[16:02]  it returns 
[16:02]  as I remember bzrlib creates fake working tree at the temp dir root
[16:03]  to prevent leaking tests outside testing jail
[16:03]  nope, fake repo, not fake wt AFAIR
[16:03]  but in fact it's WT here
[16:03]  cyberixae: for a stacked branch, it does not download the full history when you branch it.
[16:03]  should I intialize empty bzrdir then?
[16:03]  cyberixae: if you perform an operation that needs it, like, say, log, it will at that point contact the stacked-upon branch to retrieve it.
[16:04]  cyberixae: if that branch is no longer reachable, you can't log the revisions you didn't have in the first place
[16:04]  but open_containing search upwards by design, why don't you use open() instead ?
[16:04]  bialix: ^
[16:04]  why I should?
[16:05]  I'm using open_tree as substitution for WorkingTree.open_containig
[16:05]  bialix: well, open_tree docstring says: "Open working tree at specified directory." not 'specified directory or above'
[16:05]  bialix: ha ok
[16:05]  ok, i'll change docstring
[16:07]  bialix: wow, you're right, I was wrong, bzr creates a working tree not a repo
[16:07]  new docstring:http://pastebin.com/m289b92c0
[16:10]  bialix: cool
[16:10]  vila: self.make_bzrdir does not help
[16:10]  I'm getting NoWorkingTree with it
[16:11]  so... it's imporssible?
[16:11]  what happens if you use /I/do/not/exist as a path ? still NoWorkingTree ?
[16:11]  um
[16:11]  let me check
[16:12]  yep, it works
[16:12]  thank you!
[16:12]  bialix: you're welcome
[16:13]  I'm inventing simple mocks to testing qbzr
[16:13]  I'm curious how you test bzr-gtk
[16:13]  bialix: way to go !
[16:14]  bialix: from memory: classes than can raise dialogs define a self.dialog_class that tests override with nockups
[16:14]  mockups even
[16:14]  aha
[16:14]  so I'm using something similar then
[16:14]  bialix: yup ;)
[16:15]  oki
[16:25]  woot!
=== Kissaki^0ff is now known as Kissaki
[16:58]  morning vila
[16:59]  jam: hi ! Looking at your better heads right now
[16:59]  I mean reviewing
[16:59]  sounds good
[17:07]  Better head?  Is there something wrong with the one he has now?
=== jelmer is now known as Guest44014
[17:08]  fullermd: tsk tsk headS, we need more like him, so we multi-head him
[17:08]  Is that really advisable?  I'm of two minds about that...
[17:09]  two is bad, use three to avoid ties
[17:10]  pff, good enough for zaphod...
[17:17]  jam: lp is driving me mad, there is a '_counters = [0,0,0,0,0,0,0]' in your code that doesn't appear in the mail for your merge proposal, grr
[17:17]  jam: I review it from your branch >-/
[17:18]  vila: well, I updated my branch a lot since the submission
[17:19]  Ha ! That explains a lot :) Well, so I review the good bits, right ?
[17:20]  so the original idea of whether we want it or not
[17:20]  is certainly present
[17:20]  the latest version has some bug fixes
[17:20]  and more optimizations
[17:23]  jam: for both implementations ?
[17:23]  vila: the test suite passes for both implementations... so yeah
[17:23]  I didn't do as many optimizations  for the python code
[17:23]  but it at least has the correctness fixes
[17:23]  There wasn't a lot of algorithm diff between python & pyrex
[17:23]  mostly just Dict/List tuning
[17:25]  oh, and heaps are quite a bit faster if you can use peek + heapreplace() rather than pop + push
[17:25]  since it doesn't resize the list
[17:26]  (possibly causing reallocs, etc.){
[17:27]  jam: review sent, I've got revno 4411 here
[17:27]  vila: yep, 4411 is my latest
[17:27]  oh damn, missed jelmer again
[17:32]  jam: EOD for me then
[17:32]  vila: have a good night
[18:00]  Hi! After a bzr update some of my branches don't work anymore. I get errors when I try branch, log, checkout and some other command. I get "KeyErrors" and an Exception-Backtrace
=== The_User|afk is now known as The_User
[18:04]  sorry, it works after a downgrade to 1.8
[18:28]  hmm
[18:29]  there is a guy in my office who is using git
[18:29]  our code is all in svn
[18:29]  i am wondering how i should pitch bzr to him
[18:29]  he doesn't seem like a git zealot, just a guy who heard there was a better thing than svn
[18:47]  I use bzr at work against a svn repo, largely so I have "bzr shelve"
[18:58]  dash: "SVN is a great system. Just like Eminem is a popular musical artist. You know, if we're in 2002."
[18:58]  mneptok: the issue is git, not svn :)
[18:58]  dash: sorry, i read "never heard"
[18:59]  there's nothing *worng* with git. you just have to accept is was built with a certain audience in mind. if you're willing to adjust your thinking to be that audience, it will work for you.
[19:00]  personally, i want my tools to match my workflow and mindset, not vice versa. bzr does that.
[19:00]  well, bzr has a high command/interface correlation with svn
[19:01]  e.g. svn revert blah => bzr revert blah || git checkout -- blah
[19:01]  Actually, I think one of the options to git reset is more like revert...
[19:02]  yeah, that was the main thing I came up with
[19:02]  (and another is like uncommit, and another is like pull...)
[19:02]  1) 'svn foo' can be replaced with 'bzr foo' in nearly all cases 2) bzr gets by without history editing a lot better
[19:03]  crossplatform usage?
[19:07]  reduced inter-language fragility?
[19:09]  do the dev tools you're using support either git or bzr?
[19:12]  I don't think any of my dev tools support svn, cvs, rcs or sccs either... yet still I manage :P
[19:12]  I bet they do ;-)
[19:13]  but either way, if they support bzr, that would be additional ammunition
[19:20]  How do I break a lock if the break-lock command I'm told to paste throws an "Unsupported protocol" error?
[19:20]  dash: bzr also has professional paid support. git does not. you may never need it, but it's nice to know the net is there.
[19:23]  I'm sure you could find somebody good with git who'd be happy to take your money...
[19:24]  Laney: against launchpad>
[19:25]  james_w: right
[19:25]  Laney: you ignore what it tells you to do and use a different URL instead
[19:25]  bzr break-lock lp:~user/project/branch
[19:25]  apparently I locked it 864 hours ago
[19:25]  (u-d-t)
[19:26]  ok that worked, thanks
=== thekorn_ is now known as thekorn
[19:32]  hi Tak
[19:33]  jelmer: hello
[19:38]  Pfft, can't diff while pulling. Darn dirstate lock.
[19:48]  mneptok: hm! an interseting point
[19:48]  anyway
[19:48]  the dev environment is eclipse on OSX
[19:48]  how's eclipse support these days
[20:28]  * fullermd boggles.
[20:29]  * Peng_ scrabbles.
[20:29]  Or something!
[20:30]  So I just updated my bzr.dev, which was 4 revs behind (those made today)
[20:30]  Combined, the commit mails for them, which include the diffs, come to about 150k.
[20:30]  The pack file pull created was 7 megs.
[20:31]  fullermd: Yeah, something has been wrong for a while now.
[20:32]  fullermd: Nowadays I pull from lp:bzr since it behaves normally and then http://bazaar-vcs.org/bzr/bzr.dev/ in case there's anything newer.
[20:33]  I wonder if it has to do with changes to make sure stacking or CHKs have all the necessary data which somehow screwed things up for other formats too.
[20:33]  * Peng_ shrugs.
[20:33]  Someone should probably care about this, but that someone ain't me.
[20:34]  For me the annoying part is that it always transfers 5 MB (or 8 MB this time). I don't notice the disk space.
[20:38]  Hm, yeah.  branch'ing it off and pulling from lp:bzr gives a 400k pack.  That makes a lot more sense.
[20:39]  It's nice to know I'm not just crazy. :)
[20:40]  * fullermd scribbles up an email.
[20:40]  Currently when I try view a file in loggerhead on launchpad, I get "internal server error"
[20:40]  * Peng_ starts running bzr pull with various -D options.
[20:40]  e.g. http://bazaar.launchpad.net/~garyvdm/qbzr/trees/annotate/head%3A/lib/browse.py
[20:40]  garyvdm: There's some stacking-related bug.
[20:40]  garyvdm: At least that's the popular theory, IIRC.
[20:41]  Peng_: is there a work around?
[20:42]  garyvdm: Don't think so.
[20:42]  Is there maybe a cat page, I don't need annotate.
[20:43]  Oh - there is a small little icon on the right hand side.
[20:44]  If you have some bandwidth, you can run "bzr cat" and annotate on remote branches.
[20:46]  Hehe, running bzr pull from a GigE connection is fun. :D
[20:47]  Peng_: I'm posting a question to stack overflow, and I want to put a link to the code.
[20:48]  garyvdm: Ah.
[20:52]  fullermd: FWIW, I here's a log of pulls (just revision 4443) with various -D options (though I didn't combine them): http://cheezum.mattnordhoff.com/tmp/.bzr.log
[20:52]  Err, s/I //
[20:53]  May be a useful followup to my mail.
[20:54]  Where'd you send your mail? The list?
[20:54]  Yah.
[20:54]  * Peng_ pokes at his mail client.
[20:59]  any idea why bzr would eat up all my RAM and then go get itself OOM killed?
[21:00]  solarion: because it was doing too much allocation?
[21:00]  er, no idea!
[21:00]  "Build phase:Adding file con"
[21:00]  what were you doing ?
[21:00]  bzr co --lightweight 
[21:00]  solarion: Would the file "con" happen to be larger than your RAM?
[21:02]  it is effing annoying
[21:07]  there it goes again
[21:10]  jelmer: hi, aware of any rules when extending what object types dulwich gets? (i'd like to add some own stuff for my personal hacks)
[21:11]  ronny: basically just register in dulwich.objects
[21:12]  jelmer: aware if git has any rules for those extensions
[21:12]  ronny: in dulwich.objects.type_map and dulwich.objects.num_type_map
[21:12]  last time i asked that in #git they acted like i was xenu
[21:12]  ronny: I don't know what the C git implementation will do with objects of types it doesn't know about
[21:13]  ronny: dulwich will raise a KeyError I think if it encounters objects of types it doesn't know about
[21:18]  thumper, mwhudson: How hard would it be to do imports to a non-default repository format on launchpad?
[21:18]  thumper, mwhudson: (on a case-by-case basis)
[21:28]  jelmer: we don't really have anywhere to store the information that there's a different format wanted for this import
[21:29]  jelmer: it could be done, but would be work
[21:29]  mwhudson: ok
[21:29]  mwhudson: (some bzr-git imports will work into the 2a format, since it doesn't squash xml-invalid characters)
[21:29]  jelmer: why do you ask?
[21:29]  ah
[21:30]  well, hopefully there will be a new default format fairly soon :)
[21:30]  we can make all git imports 2a more easily
[21:31]  do new SVN imports use bzr-svn yet?
=== afk is now known as mthaddon
[21:49]  jam: <3 for lp:~jameinel/bzr/1.15-pack-source
[21:51]  Peng_: did you test if it works for you?
[21:55]  jam: No. If you think it's safe enough, I'd be happy to, though.
[21:55]  Peng_: should be safe, the tests still pass :)
[21:56]  jam: OK, but if it is somehow broken, will it lead to tracebacks or horrible corruption?
[21:56]  The former is ok. The latter, not so much.
[21:57]  Peng_: if it was horribly broken, I suppose you could get some missing objects
[21:57]  like it wouldn't fetch a referenced text key
[21:57]  etc
[21:57]  you could always fill those in later if it was really bad
[21:57]  or you could just do  a test in a copy of your repo
[21:57]  and tell me if the download was correct :)
[21:57]  and not worry about using it 'live' as much
[21:58]  oh, and you can 'cp -rl' if space is a concern
[21:58]  (we don't append to pack repos, so they are fairly safe to hardlink)
[21:59]  Pulled 1933 KB. That's definitely an improvement over 8+ MB!
[21:59]  jam: I've already run "cp -ai" on the repo like 8 times today. 1 more isn't gonna hurt!
[21:59]  jam: Oh. I didn't check that it didn't horribly corrupt the repo, though.
[22:00]  jam: Want me to?
[22:00]  Peng_: as long as you can "bzr log -vp" the newly pulled revisions
[22:00]  I'm pretty confident
[22:00]  Argh, I already deleted it. :P
[22:00]  since that would have to extract all the data that we should have referenced
[22:00]  Peng_: well as you said, only 1MB :)
[22:00]  jam: Yeah, but now I have to copy the repo again
[22:01]  jam: "bzr log -vp -r -1" worked.
[22:02]  Will "bzr pack" completely eliminate the duplicate data?
[22:02]  Or whatever?
[22:03]  Peng_: yes
[22:03]  'bzr pack' will remove the duplicates
[22:03]  lifeless: in bzr.dev 4441, aliases now override main command names.  Concretely, "bzr help shelve" now uses the bzrtools version.
[22:03]  jam: it seems that it also backs up previous packs
[22:03]  I've seen cases where bzr pack doubles repo size...
[22:04]  pygi: rm .bzr/repository/obsolete_packs/*
[22:04]  jam: I know :p
[22:04]  and yes, it is 2x the size until that gets cleaned out
[22:04]  others don't tho :)
[22:04]  but I'm pretty sure Peng_ knows that, too :)
[22:05]  jam: can't we get bzr pack clean? :P
[22:05]  at least in 2.x, where we supposedly focus more on the UI :)
[22:06]  jam: Good to know. I'm probably too lazy to run around "bzr pack"ing everything, though... :P
[23:21]  abentley: I will file a bug immediately and look at it shortly.
[23:22]  abentley: to be clear; bzrtools has old-shelve which has an alias shelve?
[23:24]  jam: FYI, I ran "bzr check" too, and it was ok. (Bunch of inconsistent parents, but that's always the case.)
[23:25]  lifeless: "shelve1", not "old-shelve", but yes.
[23:27]  jam: morning
[23:27]  jam: still around?
[23:30]  I could swear I had reconciled that repo... where did the inconsistent parents come from? new revisions?
[23:32]  morning lifeless
[23:32]  morning poolie
[23:34]  Peng_: did it get upgraded to rich roots rececntly?
[23:34]  lifeless: Nope, still 1.9.
[23:34]  It's totally possible that I didn't ever reconcile it. I have a lot of repos lying around.
[23:43]  hello all
[23:44]  poolie: hi.
[23:48]  Hi poolie, jml
=== visik71 is now known as visik7
[23:50]  garyvdm: hi