[12:55] <BrianB04> I'm already liking bazaar...it's so simple.
[12:56] <lifeless> cool
[12:56] <lifeless> I like to hear such things
[01:00] <igc> morning all
[01:03] <lifeless> hi
[01:03] <lifeless> jelmer: ping
[01:05] <lifeless> jelmer: your code.launchpad.net/~jelmer/bzr-svn/trunk is borked
[01:37] <jelmer> lifeless: thanks, I'll have a look
[01:39] <abentley> lifeless: the new annotation stuff is in.
[01:55] <lifeless> jelmer: I think you renamed something and didn't tell launchpad
[01:55] <lifeless> jelmer: also you should tell launchpad what branch is used for each release series
[01:57] <poolie> helol
[02:01] <lifeless> hi
[02:02] <igc> hi poolie
[02:31] <igc> poolie: w.r.t. your partial review, did you want the -D options in a separate 'debug-options' topic, in a separate "table" or just all together in the existing table separated by a newline say?
[02:32] <igc> abentley: if you're happy with Nam's tweak to the hook doc in his pre-commit patch, let me know and I'll merge the change today
[02:42] <poolie> igc,just in the same thing separated by a newline
[02:42] <poolie> would be fen
[02:42] <poolie> fine
[02:42] <igc> thanks
[02:42] <lifeless> igc: abentley is not here
[02:42] <igc> ok
[02:52] <poolie> spiv, ping?
[02:58] <spiv> pong
[03:01] <poolie> lifeless, should we be using ppas at all for bazaar packaging?
[03:01] <poolie> spiv, can i call?
[03:02] <spiv> Sure.
[03:04] <lifeless> poolie: no advantage at this point
[04:01] <igc> poolie: just confirming, the indentation of NEWS items was explicitly reduced by 1 char in 2773. Any reason for that?
[04:01] <poolie> igc, because i noticed that the rest rendering had them at different bulletpoint levels
[04:01] <poolie> in different sections of the file
[04:02] <poolie> i think i got it right
[04:02] <igc> np - just impacting merging of course.
[04:02] <poolie> yes http://doc.bazaar-vcs.org/bzr.dev/en/release-notes/NEWS.html#in-development
[04:02] <poolie> it's better but there are still some that are a bit worng
[04:04] <igc> ok - I'll clean it up now as a trivial change. I need to fix some other ReST formatting anyhow in developers/update.txt so I'll do both while I'm at it
[04:06] <poolie> jml, hi?
[04:06] <jml> poolie: hi
[04:09] <lifeless> igc: just sent a commit patch up that you might like to review
[04:10] <igc> lifeless: got the email re memory but no patch thru yet
[04:11] <igc> lifeless: I have 4G on my main benchmarking machine of which 1G is given to a VM btw
[04:12] <igc> lifeless: ok, patch thru now - will review today
[04:14] <lifeless> are you able to benchmark outside the vm ?
[04:15] <igc> I'm doing that - in the other 3G
[04:15] <lifeless> cool
[04:15] <igc> the VM is Windows
[04:15] <lifeless> we still spend 13% in libz
[04:16] <igc> but I'm not surprised that compression is an overall win - thanks to reduced IO
[04:16] <lifeless> oh yes
[04:17] <lifeless> raw pack is 990MB when annotations are used
[04:22] <jdong> tailor config files have just been more painful than swallowing needles for me...
[04:24] <lifeless> why not use bzr-svn ?
[04:24] <lifeless> 'pull' is about as simple as it gets :)
[04:24] <jdong> lifeless: because I only need a -20-revision horizon....
[04:25] <jdong> also, for my KTorrent needs.... bzr-svn would have to mirror the *entire* KDE xD
[04:26] <jdong> but for smaller trees, bzr-svn is bliss
[04:32] <lifeless> jdong: why would it mirror everything ?
[04:33] <jdong> lifeless: because KTorrent is a subdirectory of the entire KDE subversion repository....
[04:33] <jdong> lifeless: at least from what I understand, I would have to branch the entire KDE svn, not just a subdirectory
[04:37] <lifeless> if its structured well bzr-svn can treat them as nested trees
[04:40] <ubotu> New bug: #136890 in bzr "merge --uncommitted ignores specific file" [Undecided,New]  https://launchpad.net/bugs/136890
[04:41] <jdong> lifeless: hmm I'll have to look more into that then
[05:14] <lifeless> igc: another one up
[05:39] <igc> lifeless: got it - will review now
[06:14] <igc> lifeless: what's our policy w.r.t. changing return values on public methods? Should we use another name like add_lines_get_details() instead of just changing what add_lines returns?
[06:15] <lifeless> igc: we aim for compatability where the tradeoffs are worth it
[06:15] <lifeless> this is very much an internal aspect of implementation, I don't think compatability is worth it
[06:16] <igc> I'm ok with that.
[06:17] <igc> still, might be worth putting it in the API breaks part of NEWS for completeness, yes?
[06:20] <lifeless> I did
[06:20] <lifeless> :)
[06:23] <igc> /wishes he looked at the diff for NEWS in meld and not just bzr cdiff :-)
[06:27] <poolie> spiv, hi?
[06:27] <poolie> igc: "Can we get this change through as an incremental improvement and have
[06:27] <poolie> this discussion separately?" yes, in principle, i'll read the code
[06:27] <igc> cool
[06:34] <lifeless> spiv: pack patch you may be interested in up
[06:34] <poolie> spiv, how's the format doc?
[06:36] <spiv> poolie: Will have it mailed in a few minutes.
[06:39] <igc> lifeless: can you quickly explain the changes to test_weave.py? StoreText and StoreTwo are redundant tests I gather so deleting them is fine, yes?
[06:40] <lifeless> poolie: I'm done for the day; started at 6am. 3 patches up, user time is now 3m5seconds
[06:40] <lifeless> igc: thats right. The other change was to use the new return value api
[06:41] <igc> ok - all approved. I'll mark it in BB as such.
[06:41] <lifeless> thanks
[06:45] <thumper> poolie: ping
[06:55] <igc> lifeless: if you're still there, that change might be breaking sefltest test_bundle. I'll dig a bit deeper
[07:00] <lifeless> igc: it does. I'll fix and submit tomorrow morning
[07:00] <lifeless> ciao
[07:00] <igc> bye
[07:15] <bialix> lifeless: hi
[07:28] <igc> bialix: lifeless has wrapped up for the day after an early start
[07:29] <poolie> spiv, thanks for your thorough review of C patience diff
[07:29] <bialix> poolie: hi
[07:30] <poolie> hi bialix
[07:31] <igc> thanks for the review poolie
[08:33] <davidmccabe> Sorry, this seems like a dumb question, but I couldn't find with google:
[08:33] <davidmccabe> I have two branches here; how do I see the diffs between them?
[08:35] <beuno> davidmccabe, bzr missing would be one way
[08:35] <davidmccabe> never mind, that was too easy :)
[08:36] <davidmccabe> it *looks like* I can just do, bzr diff wc-1 wc-2
[08:36] <beuno> yeap, that would output a diff
[08:36] <beuno> missing tells you the difference on the commit level
[08:37] <davidmccabe> ah, I was looking for a diff.
[08:37] <davidmccabe> to recollect all my changes at the end of a feature branch.
[08:37] <davidmccabe> actually, I am comparing VCSs on speed here.
[08:37] <davidmccabe> My actual VCS for work is SVN. It took fifteen minutes.
[08:37] <davidmccabe> git does it in under a second.
[08:38] <davidmccabe> bzr does just fine at 1.29 seconds.
[08:38] <beuno> ah, right, bzr isn't focused too much on performance at the moment, although a lot of work is being done on it
[08:38] <davidmccabe> It performs great!
[08:38] <davidmccabe> git is written in C with a bit of assembly, and it's only a few times faster for this operation.
[08:38] <davidmccabe> (must be IO-bound?)
[08:39] <beuno> well, it'll be fun to watch when they do focus on performance  :D
[08:43] <poolie> davidmccabe, beuno, we are very focussed on performance in our current work
[08:43] <poolie> we were not so much in the past
[08:43] <poolie> but look at all the performance patches on the list in the last week
[08:52] <beuno> poolie, right, I've seen all the working going on, I just still have in my head the so much repeated phrase "we are not focusing on performance"  :D
[08:53] <beuno> poolie, would that mainly be the "pack" feature?
[08:54] <poolie> pack, commit performance, hpss, inventory changes
[08:54] <poolie> iter merge
[08:55] <beuno> that just got in, so it's not in 0.90, right?   I have a few <1gb branches at the office I would like a performance boost on  :D
[08:56] <poolie> pack will be in 0.92, not the last release or this one
[08:56] <poolie> but if you want to try it from robert's branch that would be enormously useful
[08:56] <vila> hi all
[08:56] <poolie> hi vila
[08:57] <beuno> poolie, sure, I could run some tests if that would help, they mainly contain a lot of binaries (PSDs, PNGs and all that)
[08:57] <vila> poolie: sorry for the delay in coming back, huge backlog :-/
[08:58] <poolie> vila, np
[09:00] <vila> poolie: not sure I get all the details right, but it looks like I will be most helpful to bzr by: 1) looking at the FTP password bug, 2) multiple connections for update, 3) auth-ring spec, 4) recursive hash on directories
[09:00] <vila> in that order of priority
[09:00] <vila> any comment ?
[09:01] <poolie> that sounds great
[09:01] <poolie> i'm working on 45
[09:01] <poolie> 4
[09:02] <vila> poolie: I re-read your directory-fingerprints.txt yesterday, nice work, I'd try to comment on it asap
[09:07] <poolie> is Lukas here?
[09:08] <vila> poolie: do you except your work to land in 0.91 or later ?
[09:08] <poolie> later, it's based on packs and they will likely be in the next release
[09:09] <vila> ok
[09:31] <ubotu> New bug: #136942 in bzr "What's in the trash" [Wishlist,Triaged]  https://launchpad.net/bugs/136942
[09:41] <poolie> vila, interesting suggestion
[09:41] <fullermd> I claim royalty rights on the name when somebody writes "bzrchaeology".
[09:42] <vila> poolie: Just had the need for it :)
[09:43] <vila> fullermd: I will argue that I use the name archeology for years ! Whether for VCS purposes or just trying to understand code :)
[09:43] <fullermd> Yes, but you didn't pun it with 'bee zee arr'.
[09:44] <vila> fullermd: and don't start me about explaining young developers that, no, people are not stupid, even if their code looks so. Once, with the help of archeology, you understand why they wrote it this way *at the time* their code was not stupid, nor them :)
[09:44] <fullermd> Oh, you're whistling in the dark, there.  I stand by my conviction that pretty much everybody in the world _is_ stupid   :)
[09:45] <vila> fullermd: this may be true but is unrelated :)
[09:45] <vila> or is it in a self-referring way ???
[09:45] <ne1uno> commits/developers=iq
[09:46] <fullermd> Oh, no.  I'm a genius.  Everybody else is a moron.  The fact that you don't immediately recognize my brilliance simply proves the point   ;)
[09:46] <vila> fullermd: ooops, just realized my remark was ambiguous :) I meant: are you calling people stupid by looking at their code ?
[09:47] <fullermd> Oh, no.  I don't have to look at code.  I assume _everybody_ is stupid.
[09:47] <vila> fullermd: hmmm, you remind me of someone...just can't remember who...
[09:52] <fullermd> What, somebody else has attained my exhalted level??  Never!  We'll joust for the title!
[10:11] <poolie> spiv, can i read about the protocol yet?
[10:11] <spiv> poolie: Yeah, I sent a merge/rfc about it.
[10:11] <poolie> oh great, thanks
[10:12] <poolie> spiv, can you look at http://bundlebuggy.aaronbentley.com/request/%3C46D86283.4080602@utoronto.ca%3E
[10:13] <poolie> actually, i see a bug, but if you can say if the basic intent is right that'd be good
[10:13] <spiv> Yeah, that's high on my mental todo list.  Like Robert I'll need to page in a little bit too, but not as much as Robert...
[10:17] <poolie> ok, don't worry til your stuff is up
[10:17] <poolie> i can read it
[11:00] <ubotu> New bug: #136952 in bzr "Bzr doesn't merge the most recent revision" [Undecided,New]  https://launchpad.net/bugs/136952
[11:31] <poolie> spiv, is http://bundlebuggy.aaronbentley.com/request/%3C20070826091222.GC11127@steerpike.home.puzzling.org%3E
[11:31] <poolie> the most current fix for that bug?
[11:31] <poolie> for trunk
[11:31] <spiv> poolie: yes
[11:31] <poolie> thanks
[11:31] <poolie> igc, thanks for all the reviews
[11:32] <igc> np
[12:27] <BrianB04> Morning all.
[12:33] <gabe_> morning BrianB04
[12:33] <grimboy> Morning, Bri. Doing a checkout or branch of a branch I've just pushed I get the error bzr: ERROR: No such file: 'http://code.discussium.org/discussium/.bzr/repository/knits/8f/models.py.old-20070902211246-zblkha1c6hz8yag7-30.kndx'. The file is available where I pushed from put not where I pushed to. I've tried using both sftp and bzr+ssh but both do the same thing. Any idea what I'm doing wrong?
[12:34] <grimboy> s/put/but/
[12:46] <Kamping_Kaiser> is bzr <current releases> still compatible with the 0.8.2 in Ubuntu 6.06?
[12:47] <gabe_> seems fine for me
[12:47] <gabe_> i'm on dapper
[12:47] <Kamping_Kaiser> i was just making sure there hasnt been any storage format breaks
[12:47] <dato> Kamping_Kaiser: how do you mean, "compatible". whether branches created with current versions can be read by 0.8?
[12:48] <Kamping_Kaiser> dato, and vica versa
[12:48] <Kamping_Kaiser> as will be the case
[12:48] <dato> vice versa, yes, always.
[12:48] <dato> branches created by the default format up to 0.90 can be read by 0.8 as well.
[12:49] <dato> however 0.15 and onwards have a newer format that is incompatible with 0.8, but it has to be enabled by hand.
[12:49] <dato> finally, 0.91 will make that new format the default.
[12:50] <dato> (but you can still create branches in the old format with 0.91 and upwards; you just need to pass the appropriate flag to bzr init)
[12:50] <Kamping_Kaiser> would the version of bzr in debian 4.0 or ubuntu 7.04 be compatible? i'm asuming so
[12:51] <dato> the very same comments apply.
[12:51] <Kamping_Kaiser> sweet
[12:51] <Kamping_Kaiser> *goes to find out versions in said distro releases*
[12:51] <dato> debian 4.0 has 0.11
[12:51] <BrianB04> I'm back, sorry about that, quick shower. Ya know, I offically love Bazaar...it's so simple once you get an idea of how things work.
[12:52] <dato> Kamping_Kaiser: in any case, there are backports for both debian (in backports.org) and for ubuntu (in bazaar-vcs.org/releases/debs)
[12:52] <dato> Kamping_Kaiser: in case you're interested
[12:52] <Kamping_Kaiser> 0.15 in feisty
[12:52] <Kamping_Kaiser> dato, i'll bear it in mind, thanks
[12:53] <BrianB04> Otherwise Kamping_Kaiser: Why not just grab the source from the bazaar site?
[12:53] <Kamping_Kaiser> i just init'd a project before i came and asked here
[12:53] <Kamping_Kaiser> BrianB04, what, build it myself?
[12:54] <BrianB04> Kamping_Kaiser: Yea, not that difficult. And, you can do a quick google search for something like 'build bzr in x distro' and usually will find (Especially for debian/ubuntu) howtos of exactly what you need to apt-get to make it build properly.
[12:54] <dato> BrianB04: well, why build yourself if both distros provide updated backports??
[12:54] <Kamping_Kaiser> BrianB04, i prefer not to build (esp. if thers backporst)
[12:56] <BrianB04> dato: Unfortunetly, they are not always updated, that's the big problem. You know what I always found a little weird, when I was first looking into bazaar, that Ubuntu does not keep the most up to date in repos...and they are 'owned' by Canonical who does Bazaar.
[12:56] <dato> BrianB04: what "repos"?
[12:57] <BrianB04> In one of the main not like Backports, ya know?
[12:57] <dato> BrianB04: latest versions have always been in bazaar-vcs.org/releases/debs, ttbomk
[12:58] <dato> you'd have to ask some ubuntu person why they are in bazaar-vcs.org instead of feisty-backports, eg.
[12:58] <BrianB04> Ah, I haven't been in Ubuntu in awhile...stupid computer now doesn't like Ubuntu, in the least.
[12:58] <Kamping_Kaiser> the -backports repos are pulled from a later release, by someone in the backports team with motivation
[12:58] <Kamping_Kaiser> i asume the bzr-cvs ones are done as a mater of course
[01:00] <BrianB04> But, I'm happy that finally launchpad mirrored my bazaar repo after the race condition yesterday. Hey, I have a trivia question regarding bazaar/drcs if someone knows, when did the whole DRCS thing start? Back when Bitkeeper came in to being to 'help' out the kernel team, or did DRCS start before then?
[02:22] <grimboy_uk> When doing a checkout or branch of a branch I've just pushed I get the error bzr: ERROR: No such file: '*blah*/.bzr/repository/knits/8f/models.py.old-20070902211246-zblkha1c6hz8yag7-30.kndx'. The file is available where I pushed from but not where I pushed to. I've tried using both sftp and bzr+ssh but both do the same thing. Any ideas what I'm doing wrong?
[02:32] <markvandenborre> I'm setting up a central bzr repository for two people
[02:32] <markvandenborre> how do I avoid having to fiddle with that repo as root?
[02:33] <markvandenborre> we are two people, each with our own ssh account on this machine
[02:33] <gabe_> markvandenborre: ensure both people are the members of a same group
[02:34] <markvandenborre> gabe_, yes, of course,...
[02:34] <gabe_> and that group has write permissions on the repository
[02:34] <markvandenborre> but then you still need to set the default group for newly created files
[02:34] <markvandenborre> with umask-ish stuff
[02:34] <markvandenborre> I don't see if that is possible on a directory basis
[02:35] <markvandenborre> or rather: how that is possible
[02:35] <dato> ensure that the directory is mode g+s and g+w
[02:35] <dato> (numeric mode 2775)
[02:35] <dato> bzr *should* take care of the rest, or so I hear
[02:36] <gabe_> works for me with the users being in the same group
[02:47] <AfC> 02755 and umask 0002 worked for us
[02:50] <lifeless> grimboy_uk: are you using http to pull from ?
[02:59] <BrianB04> Hey, are bzr branches cheap?
[03:03] <AfC> Short answer: yes
[03:30] <matkor> BrianB04: Even cheaper if you have them in one repo
[03:31] <AfC> Of course, more to the point, they're real branches :)
[03:37] <BrianB04> God, I hate going outside of stories when developing in rails, but...blech I have to for this.
[03:42] <markvandenborre> dato, bzr commit -m "Test met gebruiker Onno"
[03:42] <markvandenborre> Password:
[03:42] <markvandenborre> bzr: ERROR: Permission denied: u'/srv/bzr/.bzr/repository/lock/pending.62hc7mnjpa6jcyrq91v1.tmp': [Errno 13]  Permission denied
[03:44] <markvandenborre> I'm still locked to this one user "mark"
[03:45] <markvandenborre> I know I could chgrp -R admin ./ in /srv/bzr
[03:45] <markvandenborre> (both users are in the admin group-
[03:45] <markvandenborre> )
[03:45] <markvandenborre> but that would not be the cleanest way to do it, would it
[03:46] <markvandenborre> what is the cleanest way when it comes to permissions
[03:47] <dato> markvandenborre: find /srv/bzr -type d -print0 | xargs -r0 chmod 2755
[03:47] <dato> markvandenborre: find /srv/bzr -not -type d -print0 | xargs -r0 chmod 664
[03:47] <dato> do that and try
[03:49] <markvandenborre> you make me change all directory rights g+ws, and all file rights ug=rw,o=r
[03:50] <dato> aha
[03:50] <markvandenborre> would that be the cleanest way?
[03:50] <markvandenborre> I wonder if this is such an uncommon setup
[03:50] <markvandenborre> that I have to fiddle with it like this...
[03:51] <markvandenborre> (btw, thx for your help, dato!)
[03:51] <dato> well, maybe the permissions were bad, to this should be a one time thing
[03:58] <markvandenborre> dato, the problem is that mark.mark is still the user/group rights on many files, so onno can't touch them
[03:59] <markvandenborre> (after setting up the repo with g+ws as you first suggested)
[03:59] <dato> ah, then if you don't have root there, you will need mark to fix
[03:59] <markvandenborre> I have root here
[03:59] <markvandenborre> but I was wondering how to fix this in a sustainable way
[03:59] <markvandenborre> so that mark and onno can work together without touching anything at the shell level once it is set up well
[04:00] <markvandenborre> presuming they stay within an existing repository, that is
[04:01] <dato> as I understand it, the starting point is (1) all files and directories g+w; (2) all directories g+s; (3) all files and directories group "admin"
[04:01] <dato> AfC mentioned something about the umask, but I've hear jam say bzr preserves the g+w bit, so it should not be needed
[04:02] <markvandenborre> ok
[04:02] <markvandenborre> how documented is this?
[04:03] <dato> I don't know.
[04:04] <AfC> (group whatever-you-want-it-to-be)
[04:08] <grimboy_uk> lifeless, Yes.
[04:11] <grimboy_uk> But I've looked at it on the server through the file system and the file that's on my computer that it's complaining about not having (*blah*/.bzr/repository/knits/8f/models.py.old-20070902211246-zblkha1c6hz8yag7-30.kndx) isn't there.
[04:28] <markvandenborre> I have this bzr with central workflow running now, as described at http://doc.bazaar-vcs.org/latest/en/user-guide/centralized_workflow.html
[04:28] <markvandenborre> thx, step 1 is complete
[04:52] <BleSS> niemeyer: hi! I contacted with you ago 2 weeks about issues on python-mcrypt and python-mhash, and I sent you an email but withou answer
[04:54] <BleSS> and you said me that you were to looking/working about them on past week, I would to know if you aren't maintain those projects netiher upload to PyPi
[04:54] <BleSS> because I started to built a python wrapper over your programs for nothing
[05:06] <niemeyer> BleSS: First, this isn't the right channel to talk about it..
[05:06] <niemeyer> BleSS: Second, I said I wouldn't be able to look at these projects before the sprint was over.. it didn't mean I would look at them immediately after the sprint was over
[05:07] <BleSS> niemeyer: yes, it's true, sorry, but as I was not answered by email after of 3 emails ...
[05:07] <niemeyer> BleSS: I have dozens of things to look at, literally, and these projects aren't at the top of my list.
[05:07] <BleSS> niemeyer: excuse me but you said me that you were to looking them on the last week
[05:07] <niemeyer> BleSS: I appreciate your effort on them, and don't want you to feel frustrated because I'm not giving you attention
[05:08] <BleSS> niemeyer: then, for the next time you must say the truth and answer the emails
[05:11] <markvandenborre> a strange beginner question:
[05:11] <markvandenborre> I want to use bzr for web development
[05:12] <radix> generally one uses it to manage revisions ;-)
[05:12] <niemeyer> ...
[05:12] <niemeyer> Aug 21 11:02:02 <niemeyer>      So I apologize for the lack of response
[05:12] <niemeyer> Aug 21 11:02:08 <BleSS> no problem
[05:12] <niemeyer> Aug 21 11:02:19 <niemeyer>      Ok.. I'll try to look at these issues next week
[05:12] <niemeyer> Aug 21 11:02:23 <BleSS> if you're busy i don't disturb
[05:12] <niemeyer> Harsh.. harsh..
[05:13] <radix> !?
[05:13] <markvandenborre> radix, for managing revisions of web development files of course
[05:13] <niemeyer> radix: Discussion above..
[05:14] <radix> wow
[05:14] <radix> markvandenborre: oh, ok :-)
[05:14] <markvandenborre> but how do I get the clean source tree
[05:14] <markvandenborre> without any .bzr stuff
[05:15] <markvandenborre> so that I can use a checkout directly in my web tree?
[05:15] <markvandenborre> I'm probably not expressing myself too correctly here
[05:15] <mwhudson> well, 'bzr export' will do that
[05:16] <mwhudson> but i'm not sure that's what you actually want
[05:16] <markvandenborre> ah, thx
[05:16] <markvandenborre> mwhudson, now I can at least look things up
[05:16] <abentley> export is a one-time operation, so you won't be able to use "pull" or "update" in that directory.
[05:17] <abentley> But "pull" and "update" require a .bzr directory, even if it doesn't have a lot in it.
[05:18] <abentley> I'm not sure why you want to avoid ".bzr".
[05:24] <gabe_> ls
[05:45] <BrianB04> I don't think that a webserver could even see a .bzr directory.
[05:46] <markvandenborre> abentley, I wonder if this has any security implications
[05:47] <abentley> BrianB04: They typically don't show it on directory listings, but yes, you can access a .bzr over http.  If not, most public branches wouldn't work.
[05:48] <BrianB04> abentley: Yes, that's true but how many people are just going to try that?
[05:48] <abentley> markvandenborre: Well, you can always make .bzr/* a forbidden URL in your server config.  But you certainly don't need to store all the version history there, anyhow.  You can make a lightweight checkout instead.
[05:59] <markvandenborre> abentley, ahh... another good thing you bring up
[05:59] <markvandenborre> a lightweight checkout
[06:03] <markvandenborre> what is the fastest operation: lightweight checkout, or export?
[06:05] <markvandenborre> think of a web site in development that needs to be updated quite frequently from a bzr repository
[06:05] <markvandenborre> maybe there is a better/faster way?
[06:09] <markvandenborre> ... any good doc pointers pertaining to this also welcome
[06:09] <markvandenborre> !
[06:15] <ubotu> New bug: #137044 in bzr "Saved FTP passwords not used, password not prompted" [Undecided,New]  https://launchpad.net/bugs/137044
[06:15] <ubotu> New bug: #137045 in bzr "bzr push hangs some minutes if server data quote is exceeded" [Undecided,New]  https://launchpad.net/bugs/137045
[06:22] <markvandenborre> a bzr update from a remote sftp server (pubkey auth) where nothing has changed consistently takes about nine seconds
[06:22] <markvandenborre> is that normal?
[06:22] <markvandenborre> this is on bzr 0.15-0ubuntu2
[06:25] <AfC> markvandenborre: Step 1: upgrade to the current release, 0.90
[06:26] <AfC> markvandenborre: you are using code that is more than 4 releases old.
[06:26] <markvandenborre> AfC, or from another perspective, I'm using the latest ubuntu on a production server...
[06:26] <AfC> markvandenborre: well. then that's an Ubuntu bug, then.
[06:27] <markvandenborre> maybe there is something more recent in a backports repo somewhere
[06:27] <radix> markvandenborre: yes, the latest bzr packages for ubuntu are made available
[06:27] <markvandenborre> oh, nice
[06:27] <radix> you'll find info in the download section of bazaar-vcs.org
[06:28] <markvandenborre> thx!
[06:28] <markvandenborre> radix, brilliant!
[06:29] <sits> I seem to be getting KeyError errors trying to do
[06:29] <sits> bzr branch 'https://code.launchpad.net/~asac/intellinuxwireless/ipw3945.asac'
[06:29] <sits> on Gutsy. Any ideas?
[06:36] <markvandenborre> AfC, regarding performance
[06:36] <markvandenborre> I have just upgraded to the latest .90 release (thx again for the suggestion!)
[06:37] <markvandenborre> averages for a bzr update from a central repo with one very short text file and three revisions is just below 6 seconds on average now
[06:37] <markvandenborre> is there anything else I should check regarding performance?
[06:38] <markvandenborre> oh, I still need to bzr upgrade, apparently
[06:39] <markvandenborre> any other suggestions
[06:39] <markvandenborre> ?
[06:43] <abentley> markvandenborre: I expect that updating a checkout is faster than updating an export, because updating a checkout will only affect modified files.  But updating an export requires outputting all files each time.
[06:44] <markvandenborre> abentley, this is a checkout already, but thanks for the suggestion
[06:45] <abentley> You asked for a comparison earlier.  I was responding to that.
[06:45] <markvandenborre> ah, thx!
[06:46] <abentley> sits: Please pastebin your traceback.
[06:46] <abentley> ubtotu: paste
[06:46] <abentley> ubotu: paste
[06:46] <ubotu> pastebin is a service to post large texts so you don't flood the channel. The Ubuntu pastebin is at http://paste.ubuntu-nl.org (make sure you give us the URL for your paste - see also the #ubuntu channel topic)
[06:50] <abentley> markvandenborre: The repository is remote?
[06:57] <sits> http://paste.ubuntu-nl.org/36208/
[07:15] <markvandenborre> abentley, yes, the repo is remote
[07:16] <markvandenborre> oh... too late
[07:52] <yml> hello
[07:52] <yml> I am still trying to get my complete workflow of development flowing with launchpad.
[07:53] <yml> I beleive I reach one of the final step of the loop. I am trying to push my branch from a remote host to launchpad.net.I am getting this error message:yml@ssh1:~/workspace$ bzr push "sftp://yml-nospam@bazaar.launchpad.net/~yml-nospam/survey/main-yui"Could not create directory '/home/yml/.ssh'.Permission denied (publickey).bzr: ERROR: Unable to connect to SSH host bazaar.launchpad.net;
[07:53] <yml> Does this make sense to someone?
[08:01] <luks> "sftp yml-nospam@bazaar.launchpad.net" works for you?
[08:01] <luks> I think it's something wrong without your local ssh key setup
[08:01] <markvandenborre> yml, I'm not a bzr expert, but you should probably get your pubkey auth fixed first
[08:03] <markvandenborre> make sure .to add the contents of your local ~/.ssh/id_dsa.pub to the remote ssh/authorized_keys2
[08:03] <yml> lucks you want me to type this: sftp yml-nospam@bazaar.launchpad.net
[08:03] <markvandenborre> make sure to have pubkey auth switched on in the remote sshd configuration
[08:03] <luks> yml: yes
[08:03] <luks> I think you don't have your private ssh key installed correctly
[08:05] <yml> luks : all the problems come from that my key are not in /home/yml/.ssh but in /home/yml/www/.ssh
[08:06] <yml> luks : same kind of error:
[08:06] <yml> yml@ssh1:~/workspace$ sftp yml-nospam@bazaar.launchpad.netConnecting to bazaar.launchpad.net...Could not create directory '/home/yml/.ssh'.Permission denied (publickey).Couldn't read packet: Connection reset by peer
[08:07] <markvandenborre> what would be the best way to backup a central bazaar repository?
[08:07] <yml> when I do a ssh-add -l I can see my key
[08:08] <james_w> markvandenborre: with multiple branches?
[08:08] <luks> yml: it probably needs to write to known_hosts in ~/.ssh
[08:08] <yml> luks: :)
[08:08] <james_w> yml: ssh -v can be helpful sometimes.
[08:08] <luks> hm?
[08:09] <markvandenborre> james_w, I'm really a beginner when it comes to bzr, but...
[08:09] <markvandenborre> we're two users working on one central repository from different computers
[08:09] <yml> This took me a day to set up ssh to write known_hosts in the correct directory
[08:09] <markvandenborre> with several branches in this same repository
[08:10] <markvandenborre> one for each web project
[08:10] <yml> http://forum.alwaysdata.com/viewtopic.php?id=93
[08:10] <markvandenborre> at least, that's the layout that was suggested to me as the most logical one
[08:10] <james_w> markvandenborre: ok, so for backup, either use bzr and pull each branch in turn, to a repo on another machine. Or just rsync the whole repo over or similar.
[08:12] <yml> luks: here it is the output of sftp -v => http://dpaste.com/18500/
[08:12] <markvandenborre> james_w, which one would be most bandwidth efficient?
[08:12] <luks> well, it obviously doesn't try to use ~/www/.ssh
[08:12] <markvandenborre> not that it's such a big problem right now, but I want things to be scalable
[08:13] <luks> but I really don't know how to fix that
[08:14] <james_w> markvandenborre: rsync, unless you have masses of unreferenced revisions in the repository (you often uncommit revsions that are stored there or similar).
[08:14] <yml> luks  : thank you, I am like you but me I searching the solution for the last 2 days
[08:14] <yml> :[
[08:14] <markvandenborre> james_w, thx!
[08:14] <luks> why not just use ~/.ssh?
[08:14] <james_w> if the latter then you are probably better off just cleaning up and using rsync anyway.
[08:15] <yml> luks : because I do not have access to it
[08:16] <yml> for some reason my internet provider give me access in read only to that folder
[08:16] <Peng> ~/www? As in people can download the contents of .ssh on the web? That isn't exactly good.
[08:17] <yml> and read write in /www
[08:17] <markvandenborre> Peng, if it's like that, it's even _awfully_ bad, insecure, irresponsible
[08:17] <markvandenborre> yml, but surely it's not like that, is it?
[08:18] <BrianB04> How do you merge a branch in with full log file included?
[08:18] <Peng> BrianB04: bzr merge $branch?
[08:18] <yml> Peng: It not that bad because I can configure folders in ~/www that are accessible
[08:18] <BrianB04> Peng: Didn't pull in the full log
[08:18] <luks> yml: you can probably change the location using IdentityFile in your ssh configuration
[08:19] <Peng> BrianB04: Yes it did.
[08:19] <BrianB04> Peng: Didn't show in bzr log...or...do I have to commit...nevermind.
[08:19] <Peng> Oh.
[08:19] <yml> luks : what is this IdentityFile?
[08:20] <luks> man ssh_config :)
[08:20] <luks> but I have no idea where to put the file, because I have it in ~/.ssh/config
[08:20] <yml> thank you nicely said  lol
[08:21] <luks> you will probably need to use command line options
[08:21] <luks> and convince bzr to use the ssh with your options
[08:21] <yml> luks: it look like an endless loop => chiken, egg
[08:22] <luks> ssh -i /home/yml/.ssh/id_dsa
[08:22] <luks> it seems to use BZR_SSH environment variable
[08:23] <yml> luks you mean ssh -i /home/yml/www/.ssh/id_dsa
[08:23] <luks> right
[08:23] <luks> so `BZR_SSH="ssh -i /home/yml/www/.ssh/id_dsa" bzr push ...` might work
[08:24] <yml> BZR_SSH is an environment variable that I have to set before your bzr
[08:26] <luks> actually, not BZR_SSH seems to do something different
[08:31] <yml> luks  : yes I believe it is use to specify ssh engine
[08:32] <luks> yep, but I'm obviously bored, so here you have a simple plugin - http://rafb.net/p/5qLlgx54.html :)
[08:32] <yml> beuno save me after 13 hours of unsuccesfull search
[08:32] <luks> put this to ~/.bazaar/plugins/ymlssh.py
[08:32] <luks> and then `BZR_SSH=yml-openssh bzr push ...`
[08:33] <luks> oh, wait, do you have access to ~/.bazaar?
[08:34] <luks> if not, you need to point BZR_PLUGIN_PATH to the plugin part
[08:34] <luks> *path
[08:34] <yml> luks   : what is that?
[08:34] <luks> what is what?
[08:34] <luks> the plugin?
[08:34] <yml> I start to understand
[08:34] <luks> it will call ssh with the arguments needed to load private key from where you have it
[08:35] <yml> You have created a plug in
[08:46] <yml> luks: I have copied your plugin in a file called ymlssh.py. This file is now in /home/yml/www/.bazaar/plugins.
[08:46] <luks> do you see it in `bzr plugins`?
[08:47] <yml> I have set BZR_SSH=yml-openssh
[08:48] <yml> luks: yes I can see it
[08:48] <yml> yml@ssh1:~/workspace$ bzr plugins/usr/lib/python2.4/site-packages/bzrlib/plugins/launchpad        Launchpad.net integration plugin for Bazaar/home/yml/www/.bazaar/plugins/ymlssh.pyc/usr/lib/python2.4/site-packages/bzrlib/plugins/bzrtools        Various useful plugins for working with bzr.
[08:49] <yml> luks: it does not work
[08:49] <luks> what's the error?
[08:49] <yml> I have exactly the same result yml@ssh1:~/workspace$ bzr push "sftp://yml-nospam@bazaar.launchpad.net/~yml-nospam/survey/main-yui"Could not create directory '/home/yml/.ssh'.Permission denied (publickey).bzr: ERROR: Unable to connect to SSH host bazaar.launchpad.net;
[08:50] <luks> and do you really have the private key in ~/www/.ssh/id_dsa?
[08:51] <yml> I have a key called id_rsa
[08:52] <luks> then you need to change it in the plugin
[08:52] <luks> and then, I hope, it should work
[08:53] <yml> stupid me I have just change your plugin
[08:53] <yml> and
[08:54] <yml> it does not work  :(
[08:55] <yml> same error message:
[08:55] <yml> yml@ssh1:~/workspace$ bzr push "sftp://yml-nospam@bazaar.launchpad.net/~yml-nospam/survey/main-yui"Could not create directory '/home/yml/.ssh'.Permission denied (publickey).bzr: ERROR: Unable to connect to SSH host bazaar.launchpad.net;
[08:57] <luks> what about `sftp -oIdentityFile=~/www/.ssh/id_rsa yml-nospam@bazaar.launchpad.net` ?
[08:58] <yml> same kind of pb:
[08:58] <yml> yml@ssh1:~/workspace$ sftp -oIdentityFile=~/www/.ssh/id_rsa yml-nospam@bazaar.launchpad.netConnecting to bazaar.launchpad.net...Could not create directory '/home/yml/.ssh'.Permission denied (publickey).Couldn't read packet: Connection reset by peer
[08:59] <luks> well, it looks like the private key is not the right one
[08:59] <yml> even with the complete path : yml@ssh1:~/workspace$ sftp -oIdentityFile=/home/yml/www/.ssh/id_rsa yml-nospam@bazaar.launchpad.netConnecting to bazaar.launchpad.net...Could not create directory '/home/yml/.ssh'.Permission denied (publickey).Couldn't read packet: Connection reset by peer
[08:59] <yml> ssh try to write in /home/yml/.ssh
[09:00] <luks> that shouldn't be a problem
[09:00] <yml> in addition:
[09:00] <yml> yml@ssh1:~/workspace$ echo ~/home/yml/www
[09:00] <luks> hm?
[09:01] <yml> yes it is a pb because I do not have write access to /home/yml
[09:01] <luks> not even to /home/yml/www?
[09:01] <luks> try this: sftp -oUserKnownHostsFile=~/www/.ssh/known_hosts -oIdentityFile=~/www/.ssh/id_rsa yml-nospam@bazaar.launchpad.net
[09:01] <yml> yes I have write access to /home/yml/www
[09:02] <yml> luks in your path I don't need the www
[09:02] <yml> because ~ = /home/yml/www
[09:03] <yml> Am I understanding it correctly?
[09:03] <luks> well, try it with absolute paths
[09:03] <luks> but I don't think your home is /home/yml/www if ssh tries to write to /home/yml/.ssh
[09:05] <yml> could you believe this:
[09:05] <yml> yml@ssh1:~/workspace$ sftp -oUserKnownHostsFile=/home/yml/www/.ssh/known_hosts -oIdentityFile=/home/yml/www/.ssh/id_rsa yml-nospam@bazaar.launchpad.netConnecting to bazaar.launchpad.net...Could not create directory '/home/yml/.ssh'.Permission denied (publickey).Couldn't read packet: Connection reset by peer
[09:05] <yml> ssh you are DUMB STUPID!!!!!
[09:05] <yml> lol
[09:06] <luks> try with -v and pastebin the log
[09:06] <yml> ok I will do
[09:07] <yml> http://dpaste.com/18504/
[09:07] <yml> luks here it is
[09:08] <luks> well, I don't think the private key in www/.ssh/id_rsa and the one on launchpad match
[09:10] <yml> I can regenerate a key with ssh-keygen and upload the public key to launchpad
[09:10] <yml> would that help?
[09:10] <luks> no idea :/
[09:11] <luks> probably yes, but you need to find somebody who actually knows how ssh works
[09:11] <luks> I'm just guessing
[09:11] <yml> do you thing that I will have more chance on #launchpad
[09:12] <luks> most likely yes
[09:12] <yml> Because it does not seem to be a bzr bug
[09:12] <yml> luks thank you very much for your great help
[09:13] <yml> and also for my first plugins
[09:13] <yml> :-)
[09:13] <luks> no problem :)
[09:13] <yml> This is something that I would realy like to do write a bzr plugin to do some cherypick before my commit
[09:14] <yml> bzr seems to be very open
[09:14] <luks> cherrypick as in `bzr merge -r X..Y`?
[09:15] <yml> nope cherrypick as in a commit select to take the modification you have done on file A,B and not the one of the file D,F
[09:15] <yml> like darcs is doing
[09:15] <luks> ah
[09:16] <luks> I think there already is something like that
[09:16] <Peng> "bzr commit a b"
[09:16] <luks> there is shelve, which does the reverse
[09:16] <yml> what do you mean the reverse?
[09:16] <luks> it lets you to take selected changes aside
[09:16] <yml> in fact I am missing some interactivity
[09:17] <luks> http://projects.collabora.co.uk/~asabil/bzr/bzr_record/
[09:17] <yml> like a bzr commit -i
[09:17] <luks> I think it's based on the shelve plugin
[09:18] <yml> will ask me file by file if I want to commit the modification done on this file in the current changeset.
[09:18] <luks> change by change
[09:19] <asabil> yml, luks: it doesn't commit right now
[09:19] <luks> oh
[09:19] <asabil> it will create a patch for you in the patches/ subfolder
[09:19] <luks> I just noticed it on the plugins page some time ago
[09:19] <luks> so I assumed it works like darcs record
[09:20] <yml> luks the link is dead, empty folder
[09:20] <luks> you better get used to empty folders with bzr :)
[09:20] <yml> lol
[09:20] <asabil> yml: bzr get http://projects.collabora.co.uk/~asabil/bzr/bzr_record/
[09:20] <asabil> luks: I maybe should rename it
[09:21] <asabil> but it works exactly like bzr record
[09:21] <asabil> it just doesn't commit
[09:22] <yml> asabil the folder is empty
[09:22] <luks> it isn't, there is a .bzr folder inside it
[09:23] <yml> ???
[09:23] <luks> http://projects.collabora.co.uk/~asabil/bzr/bzr_record/.bzr/
[09:23] <luks> simply run: bzr branch http://projects.collabora.co.uk/~asabil/bzr/bzr_record/
[09:24] <yml> bzr branch http://projects.collabora.co.uk/~asabil/bzr/bzr_record/
[09:24] <yml> does that work?
[09:24] <yml> ok
[09:25] <yml> This is why I have considered most of the plugins as dead
[09:25] <yml> stupid me again
[09:46] <vaidhy> I am trying to set up bzr with webdav and I could not push my changes using webdav.. Does webdav support pushing the changes or am I struck with ssh?
[09:48] <lifeless> grimeboy: that strange
[09:48] <lifeless> grimeboy: I would run bzr check locally
[09:48] <lifeless> on the server
[09:49] <vaidhy> vila you around?
[09:58] <grimeboy> Oh wait, it's a server problem.
[09:59] <grimeboy> Now to find the character it doesn't like in "models.py.old-20070902211246-zblkha1c6hz8yag7-30.kn"
[10:00] <grimeboy> Multiple dots?
[10:00] <grimeboy> No wait, most have at least two.
[10:10] <vaidhy> Anyone successfully pushed data using webdav here?
[10:14] <lifeless> grimeboy: '.py' probably
[10:15] <Peng> Ohh, right.
[10:16] <Peng> Apache likes executing anything with ".py" anywhere in the name. :)
[10:18] <grimeboy> Ah, but other files with .py seem to be alright. Oh well, time to fiddle about with .htaccess
[10:20] <fullermd> It seems semi-random.  I've seen it with .php's.
[10:20] <fullermd> I just slammed .kndx/.knit to MIME-type a/o-s in my configs to force it to leave 'em alone.
[10:21] <grimeboy> Ah har
[10:22] <Peng> Oh, that's a good idea.
[10:23] <Peng> I guess that's a benefit of Mercurial always using a smart server: that Apache doesn't get a chance to screw around like that.
[10:23] <lifeless> packs don't have this problem
[10:40] <Peng> Oops, bzr doesn't like it when I edit .bzr/branch/location and leave a trailing newline.
[10:59] <ryanakca> Hmm. On some branches, (ssh+bzr/sftp methinks), when I commit, it get's uploaded to a remote branch, without 'push'... is there a way to tell if my local checkout is that type or not?
[10:59] <lifeless> info
[10:59] <lifeless> and its whether its a checkout or not
[10:59] <ryanakca> ok... so, 'Checkout (format: dirstate)
[11:00] <ryanakca> ' means it's local? (along with '       checkout root: file:///home/ryan/deb/debian/
[11:00] <ryanakca> ') ?
[11:00] <lifeless> local has nothing to do with it ;)
[11:00] <lifeless> its a checkout
[11:00] <lifeless> so when you commit it will go to the master branch
[11:01] <ryanakca> shucks... is there a way to make a 'local' commit, just so I can check, before 'pushing'?
[11:01] <lifeless> yes you can turn it into a regular branch or you can commit offline
[11:01] <lifeless> to do the former run unbind
[11:01] <lifeless> to commit offline do 'commit --offline'
[11:02] <lifeless> uhhh
[11:02] <lifeless> don't see how merge is related
[11:02] <ryanakca> completely seperate
[11:03] <lifeless> if you haven't committed, then merging will conflate your changes and the changes you are merging
[11:03] <lifeless> so you should commit your work then merge then commit the merge.
[11:04] <ryanakca> Ok, thanks
[11:05] <ryanakca> lifeless: bzr: ERROR: unknown command "--offline"
[11:06] <lifeless> ryanakca: you forgot the word commit
[11:06] <ryanakca>  bzr commit --offline -m "LP: #106718, #136560, dropping libgtk and 9902_nspluginviewer patch"         ?
[11:06] <lifeless> why do you need to commit to test things?
[11:06] <Lo-lan-do> Hi all
[11:06] <Lo-lan-do> Still having problems with bzr-svn :-(
[11:07] <lifeless> ryanakca: you might find --fixes lp:106718 --fixes lp:136560 are useful to add to that
[11:07] <Lo-lan-do> http://paste.debian.net/36162
[11:07] <ryanakca> lifeless: ah, thanks :)
[11:07] <Lo-lan-do> jelmer: ^^^ if you have a few moments
[11:07] <ryanakca> I don't need to commit to test
[11:07] <lifeless> ryanakca: you should not need to commit to test your work though so something isn't adding up for me
[11:08] <ryanakca> ^^
[11:08] <jelmer> Lo-lan-do: the quick fix is to make _is_http_transport() in transport.py return False
[11:09] <jelmer> I'm working on a real fix
[11:09] <lifeless> ryanakca: so why do you need to do offline commits? They are a useful feature but isn't it simpler for you to just commit normally ?
[11:10] <ryanakca> lifeless: would commiting not cause some merge things in the files?
[11:10] <ryanakca> oh well, I'll just commit normally and then merge normally. I guess I don't have anything to worry about :)
[11:11] <lifeless> 'merge things' ? I don't know what you mean
[11:11] <lifeless> commit does not alter your files
[11:11] <Lo-lan-do> jelmer: Thanks, I'll do that.
[11:11] <lifeless> jelmer: did you see my version_info bug for svn ?
[11:12] <ryanakca> lifeless: herm. *tries to find the old bug* just a sec
[11:12] <Lo-lan-do> Yay, it works :-)
[11:15] <jelmer> Lo-lan-do: cool, 0.4.2 should have a proper fix
[11:16] <jelmer> lifeless: yes, hope to have a look at it somewhere this week
[11:16] <lifeless> jelmer: should be trivial - just add 'dev', 0
[11:16] <lifeless> to the end of your version tuple
[11:17] <jelmer> lifeless: is there a bundle included in the bug report ? (-:
[11:17] <lifeless> nah
[11:17] <lifeless> fire n forget
[11:17] <jelmer> heh, ok
[11:20] <lifeless> yay
[11:20] <lifeless> one patch in
[11:24] <james_w> isn't it bzr commit --local?
[11:32] <lifeless> back shortly, breakkie
[11:52] <jelmer> lifeless: hmm, where did you file that bug? I can't find it?
[11:55] <ubotu> New bug: #137157 in bzr "Cannot lock: transport is read only" [Undecided,New]  https://launchpad.net/bugs/137157