[00:05] <wgrant> I recall seeing it a couple of years ago, so it has been around for a while.
[00:43] <keithy> any progress on nested trees
[05:59] <keithy> why is this site here http://whybzrisbetterthanx.github.com/
[06:09] <NfNitLoop> because the people who run .github.com think git is better?
[09:06] <pynonoir> Hi. Is there a way to do a fast-export | filter | fast-import of a bzr branch to a bzr branch without modifying internal revisions ids (to keep to possibility to do merges) ?
[09:46] <thumper> pynonoir: can I ask why you want to?
[09:51] <lifeless> time to fly
[09:51]  * lifeless flap flap flaps
[09:56] <pynonoir> thumper: I'm trying to import a svn repo since last week. Thanks to mzz, I found a solution involving "bzr svn-import"
[09:56] <thumper> ok
[09:56] <thumper> cool
[09:56] <pynonoir> thumper: but commiter names are still svn login names
[09:56] <thumper> ah, yes
[09:56] <pynonoir> and I would like to replace them by real e-mail adresses
[09:57] <thumper> right, that I don't know
[09:57] <thumper> you probably want to poke jelmer
[09:57] <thumper> there may be some magic you could feed into bzr-svn to provide that info
[09:58] <pynonoir> yes I wrote an "answer" question on launchpad
[09:58] <pynonoir> the other possibility would have been to export the branches, rewrite the commiter name, and then export it back to bzr
[09:59] <pynonoir> but then I think it will make the different branches somewhat incompatible
[09:59] <thumper> pynonoir: yeah, you don't want to revisions with the same id pointing to different data
[10:00] <thumper> pynonoir: best to get the info in when the revisions are being created
[10:00] <pynonoir> thumper, sure, I will wait either jelmer or for an answer on launchpad
[10:01] <thumper> ok, cool
[10:01] <pynonoir> thanks
[10:06] <mzz> heh, he's giving me way more credit than I deserve, all I did was wave vaguely in the direction of bzr-svn
[10:11] <pynonoir> mzz, I tried for more than a week with svn2bzr. Then I began to read and modify the source code of fast-export-from-svn. So ni, you saved me a lot of time
[10:42] <gerard_> hey
[13:11] <napster> How to list all files tracked by bzr?
[13:12] <Peng> napster: bzr inventory
[13:12] <Peng> napster: Also probably some permutation of "bzr ls"
[13:12] <napster> Peng tnx :)
[15:26] <MattCampbell> Is there an RPM package of bzr 2.0 for RHEL/CentOS 5?
[15:44] <mkanat> MattCampbell: Not that I know of, but you can rebuild the Fedora one.
[15:44] <mkanat> MattCampbell: That's what I do.
[15:44] <mkanat> MattCampbell: Oh actually, wait, nevermind, I think it's in rpmforge now.
[15:52] <MattCampbell> I think I'd rather build from source than mix RPM repositories (am already using EPEL).
[15:52] <MattCampbell> though really, I'd rather run Ubuntu on these servers.
[15:52] <mkanat> MattCampbell: You could use yum-priorities.
[15:53] <MattCampbell> already got the SRPM, running mock now
[15:56] <mkanat> Okay.
[15:57] <mkanat> You might have to do "rpm -i --nomd5" on the srpm if it doesn't work.
[16:30] <pynonoir> jelmer: Hi. Can svn-import replace subversion login names by e-mails during svn-import ?
[16:35] <Peng> bzr-fastimport certainly can...
[16:36] <pynonoir> Peng: from what I tested fast-export-from-svn cannot export branches
[16:37] <pynonoir> maybe I'm wrong
[16:59] <jelmer> pynonoir, hi
[17:00] <jelmer> pynonoir, no, replacing authors is not supported (and cannot be supported)
[17:00] <jelmer> it would break referential integrity
[17:00] <jelmer> the only way to support that would be to support changing committers in bzr revisions
[17:01] <pynonoir> then why bzr-fastimport can ?
[17:01] <jelmer> pynonoir, it doesn't do generate deterministic revisions
[17:03] <jelmer> s/do//
[17:04] <pynonoir> jelmer, I have modified bzr-svn to do this and imported my old svn repo. The result looks fine. What are the risks ?
[17:06] <pynonoir> I thought it was an option that I was not able to find. So I looked at "committer:" in the source files and replaced the commiter by a new string
[17:06] <jelmer> pynonoir: The risk is repository corruption
[17:07] <jelmer> Bazaar assumes that revisions that are referred to by the same revision id have the same contents
[17:07] <jelmer> please do not distribute any revisions created by your modified bzr-svn
[17:07] <pynonoir> I wasn't going to distribute that ugly hack
[17:08] <Kilroo> What's the hack do?
[17:08] <Kilroo> Oh, something to do with committer strings. Ok.
[17:08] <pynonoir> Kilroo, it replaces svn commiters logins by e-mails for bzr
[17:09] <Kilroo> Ah.
[17:09] <Kilroo> I need to figure out the best way to watch for progress on the out of memory issue...
[17:17] <zubin71> hi i am currently attempting a bzr pull and i am behind a proxy server; how do i do it? it shows "connection timed out" when i try
[17:17] <zubin71> please help
[17:18] <jelmer> pynonoir: ideally bazaar should support a way to annotate a commit with a new committer or commit message
[17:20] <pynonoir> jelmer: ok so until this is possible the only thing to do now is to import my branches, keep the old svn logins and continue from that ?
[17:22] <jelmer> pynonoir: yes, unfortunately. When support for updating the committer and commit message is added to Bazaar you should be able to update the existing revisions.
[17:30] <pynonoir> jelmer: ok, thanks for the advices. Importing branches was my top priority and that part is solved. Changing commiter strings was just a cosmetic issue.
[17:37] <mgedmin> so, can I see the history of two related branches at once with bzr viz?
[17:59] <unic0rn> hi
[18:00] <unic0rn> i'm using bzr 2.0.4 and i have a small issue with my branch sitting on vfat partition
[18:00] <unic0rn> i've just moved back to linux, and if i'll copy whole branch directory to my home dir, it works fine
[18:01] <unic0rn> but if i'll create symlink to that directory in my homedir, bzr status displays all files as modified, and i cannot commit
[18:01] <unic0rn> any ideas?
[18:02] <unic0rn> bzr: ERROR: [Errno 1] Operation not permitted: '/fll/sda4/devel/as3/projects/tiramisu/.bzr/repository/upload/a0y830g7imodklgnfkpm.pack
[18:02] <unic0rn> that's what i'm getting
[18:03] <pynonoir> does vfat support symlink ?
[18:04] <unic0rn> no, but that's a symlink on ext2 partition mounted from loop
[18:04] <unic0rn> pointing to a directory on vfat partition
[18:04] <unic0rn> which i have write access to
[18:04] <unic0rn> file from the error message gets created, but it has zero length
[18:05] <pynonoir> then you need an answer from one of the bzr expert here :)
[18:06] <unic0rn> i think it somehow gets confused by the fact that it's inside /home/uni/tiramisu but some system call probably returns full path to vfat partition as /fll/sda4/...
[18:06] <unic0rn> because bzr info returns parent branch as being /home/uni/eris when i'll copy whole directory
[18:07] <unic0rn> but when i'm inside symlinked /home/uni/tiramisu, parent branch is being returned as /fll/sda4/...
[18:07] <unic0rn> hmm
[18:07] <unic0rn> lemme check something
[18:08] <unic0rn> same error if i'll cd /fll/sda4/... and try to commit there.
[18:08] <pynonoir> i was going to ask this...
[18:08] <unic0rn> the only difference i can see are +x flags on all files on vfat
[18:08] <unic0rn> perhaps that's an issue
[18:08] <unic0rn> no idea
[18:09] <unic0rn> it shouldn't matter i think
[18:10] <pynonoir> so it is a linux bzr branch (don't know if branches structures are system dependant, I guess they are not)
[18:10] <pynonoir> you moved it on a vfat external hard drive
[18:10] <pynonoir> and then you cannot work directly on it
[18:10] <unic0rn> nope. it's a branch created under windows
[18:11] <unic0rn> and now i've switched OS to linux but i'm still keeping my data on vfat partition, as a matter of fact, linux boots from image files on vfat as well
[18:12] <unic0rn> that's why i would prefer to use files there directly rather than to copy them to my homedir on ext2, which resides inside image file
[18:12] <unic0rn> and when i copy that branch created under windows to my homedir, it works
[18:12] <unic0rn> but from within vfat partition mounted under linux, it doesn't
[18:12] <unic0rn> it's not actually symlink issue, because after entering that directory directly, it's same error
[18:14] <pynonoir> maybe a problem of locking file access
[18:15] <unic0rn> hmm
[18:16] <unic0rn> but why it displays all files as modified?
[18:16] <unic0rn> and when i copy that dir to my home, it's all fine
[18:16] <unic0rn> it returns crap only from within vfat partition
[18:20] <pynonoir> well, i'm not a bzr expert. This may be a bug, or not...
[18:20] <pynonoir> anyway you know a workaround (not using vfat)
[18:24] <unic0rn> http://wklej.org/id/271375/txt/
[18:26] <unic0rn> yeah, but that workaround isn't the best option. i may format 2gb partition to ext3 and use it to store sources there, but i would prefer them to be accessible from windows, just in case.
[18:27] <unic0rn> and yes, i know there are ext2/3 drivers for windows, but that's just another workaround to workaround that shouldn't be neccessary in the first place
[18:27] <pynonoir> maybe (just maybe), the bzr status tries to perform a lock on the directory just in case you decide to modify files during the check
[18:33] <unic0rn> i think i've found it. need to check
[19:12] <unic0rn> it's some access issue. it works under root, unfortunately i was unable to find a solution for it. changing uid and gid won't work with remount, so i would have to rebuild initrd to change those options
[19:12] <unic0rn> so i'll probably copy sources to ext2
[20:00] <timClicks> I've received a conflict on a massive locale file, is there a bzr command that says 'cross fingers & use the most recent commit and overwrite the old'
[20:01] <timClicks> i don't speak french, so i have no idea which version is correct
[20:18] <jelmer> timClicks, hi
[20:19] <timClicks> jelmer: hello :)
[20:19] <jelmer> timClicks: filename.OTHER will contain the version that the other side had
[20:19] <timClicks> oh, so a simple mv & bzr resolve will work
[20:19] <jelmer> timClicks: just overwrite the original file with that
[20:19] <jelmer> timClicks: yep, exactly