[00:24] <mgz> spiv: thanks. posted some mps for you to look over at your leisure.
[06:39] <lifeless> spiv: hi
[06:40] <lifeless> spiv: did you check for a branch.conf with %7E, or was it clearly stacking?
[06:40] <spiv> lifeless: there was a %7E in one branch.conf
[06:40] <lifeless> spiv: the 11 second thing sounds suspiciously like sftp + branch locking + the url changing issue that broke checkouts
[06:42] <spiv> lifeless: but the stacked-on locations were definitely pointing to each other (modulo how the URL was escaped)
[06:42] <lifeless> spiv: is that even possible, even with branch renames?
[06:42] <spiv> So maybe the %7E is a problem, but there also appears to be a stacking loop.
[06:42] <spiv> Which is surely a problem too :)
[06:43] <spiv> lifeless: I think so
[06:43] <spiv> lifeless: with changing the dev focus around and enough rename steps
[06:46] <spiv> See also <https://bugs.launchpad.net/launchpad/+bug/692085>, which features someone doing a complicated dance with dev focuses and renames (not sure why!) and ends up with a self-stacked branch.
[06:47] <spiv> lifeless: I agree it sounds simliar to that checkout issue too, although I haven't managed to dig up the relevant bug yet
[06:47] <lifeless> kk
[06:48] <spiv> lifeless: I perhaps don't remember enough details, and I stopped looking for it when I saw the apparent mutual stacking.
[06:48] <lifeless> the 'same branch' check looks for url string equality
[06:48] <lifeless> a normalise call was either added/removed- I don't remember - on the lookup of something checkout related
[06:49] <lifeless> it made 'bzr update' blow up
[06:49] <lifeless> fullermd remembers the bug
[06:53] <fullermd> Hmmwhut?
[07:21] <fullermd> I don't recall ever being involved with any bugs on stacking...  I have it on my mental "do not touch" list.
[07:21] <lifeless> not stacking
[07:21] <lifeless> bzr update and the %7E vs ~ thing
[07:22] <fullermd> Hm.  I vaguely remember hearing of such a thing, not sure I got anywhere near it.
[07:22]  * fullermd checks what LP has to say...
[07:24] <lifeless> fullermd: last time this was discussed here, Ithought you piped up witht he bug #
[07:24] <lifeless> perhaps it was peng
[07:26] <fullermd> Hmmm...   here's an ugliness about co's of co's with local vs remote paths: https://bugs.launchpad.net/bzr/+bug/325296
[07:26] <fullermd> Nothing about recursive livelocking though.
[07:29] <fullermd> Unrelated, but on a bug gardening note: https://bugs.launchpad.net/bzr/+bug/582656
[07:30] <fullermd> Still showing a 'New' status on one of its product/effects/whatever's.  Should it?
[07:32] <fullermd> Annoying how LP shows the same bug multiple times (and widely separated) in its listout   :|
[07:35] <lifeless> thats the per context thing
[07:35] <lifeless> there is a bug on it
[07:35] <lifeless> it has a reasonable reason behind it, solvable with some good ui work
[09:39] <Manikandan> when i use this cmd
[09:40] <Manikandan> i got the following error
[09:40] <Manikandan> root@cdac[openstack]#bzr branch lp:nova
[09:40] <Manikandan> ssh: connect to host bazaar.launchpad.net port 22: Connection refused
[09:40] <Manikandan> bzr: ERROR: Connection closed: Unexpected end of message. Please check connectivity and permissions, and report a bug if problems persist.
[09:42] <spiv> Manikandan: the service is up for me
[09:43] <spiv> Manikandan: you may have a firewall between you and bazaar.launchpad.net:22?
[11:01] <techbreak_> bzr branch lp:ubuntu/maverick/computer-janitor
[11:01] <techbreak_> Permission denied (publickey).
[11:01] <techbreak_> bzr: ERROR: Connection closed: Unexpected end of message. Please check connectivity and permissions, and report a bug if problems persist.
[11:01] <techbreak_> help me with this please
[11:06] <spiv> techbreak_: the SSH authentication is failing
[11:06] <techbreak_> spiv, what am I supposed to do ?
[11:06] <spiv> techbreak_: have you added your SSH public key to your Launchpad account?
[11:06] <techbreak_> spiv, yes
[11:08] <spiv> techbreak_: try 'ssh -v -l YOUR_LP_USERNAME bazaar.launchpad.net' and take a look at which keys it tries
[11:09] <techbreak_> spiv, http://paste.ubuntu.com/573882/
[11:10] <spiv> Also, what's your launchpad username?  I can take a look at your equivalent of https://launchpad.net/~spiv/+sshkeys and see if there's any really obvious problem.
[11:12] <spiv> Well, it's trying a key (/home/techbreak/.ssh/id_rsa), but it apparently doesn't match your publickey(s) on Launchpad.
[11:13] <ChrisCauser> Hi guys
[11:14] <ChrisCauser> Just wanted to say that I've just downloaded the PPA of the bazaar 2.4 and it is fantastic. It's so much zippier than the 2.2 I was using before. Keep up the good work. You have a fantastic codebase
[11:14] <spiv> ChrisCauser: thanks! :)
[11:17] <ChrisCauser> Spiv: No problem. I use bazaar everyday to integrate with a subversion repo and it does it much better than any other solution out there. I like how it made subversion faster even before I upgraded to 2.4!
[11:31] <techbreak_> spiv, so what am I supposed to do ?
[11:32] <spiv> techbreak_: <spiv> Also, what's your launchpad username?  I can take a look at your equivalent of https://launchpad.net/~spiv/+sshkeys and see if there's any really obvious problem.
[11:33] <techbreak_> spiv, gupta-chandan1
[11:33] <spiv> techbreak_: but basically, as I said above, the SSH public key(s) you have imported into Launchpad don't match your local SSH key.
[11:34] <techbreak_> spiv, how am i suppose to import matching one then  ?
[11:34] <techbreak_> :O
[11:36] <spiv> techbreak_: um
[11:36] <spiv> techbreak_: you should remove that publickey
[11:36] <spiv> techbreak_: and read https://help.launchpad.net/YourAccount/CreatingAnSSHKeyPair
[11:36] <techbreak_> spiv, how do I remove that publickey ?
[11:37] <techbreak_> spiv, same link I followed earlier and created key
[11:37] <spiv> techbreak_: no, you didn't create that key
[11:37] <spiv> techbreak_: that's a key from paramiko's demo's directory
[11:37] <techbreak_> spiv, huh? whats that now ?
[11:37] <spiv> techbreak_: it's important for your security that generate your own key
[11:38] <techbreak_> spiv, hmmm
[11:39] <techbreak_> spiv, ok now tell me what shall I do now and how  ? bare my Noobity, I am just a beginner
[11:39] <spiv> techbreak_: https://help.launchpad.net/YourAccount/CreatingAnSSHKeyPair
[11:40] <spiv> techbreak_: it explains things more clearly than I can :)
[11:40] <techbreak_> spiv, ok I will do that :)
[11:40] <spiv> techbreak_: I'd like to know how you found that publickey to upload, though
[11:40] <spiv> techbreak_: it seems like an unusual thing to do!
[11:40] <techbreak_> spiv, wait i will tell you :)
[11:41] <techbreak_> spiv, gpg --fingerprint gupta.chandan1@gmail.com
[11:41] <techbreak_> spiv, and then gpg --send-keys <mykey>
[11:42] <spiv> techbreak_: that's GPG, that's a different set of keys entirely
[11:43] <techbreak_> spiv, huh ? am I suppose to send some other key too ?
[11:43] <spiv> At a technical level the public key cryptography GPG does has a lot in common with SSH, but it's different software with different keys.
[11:44] <spiv> At some point you presumably went to https://launchpad.net/~gupta-chandan1/+editsshkeys and pasted in the contents of 'user_rsa_key.pub' from the demos/ directory of a library called paramiko.
[11:45] <spiv> You almost certainly *meant* to paste in the contents of ~/.ssh/id_rsa.pub
[11:47] <spiv> techbreak_: I'm off to bed.  Good luck!
[11:48] <techbreak_> spiv, thaks a lot :) good night.. have sweetest dream :)
[11:48] <jelmer> g'night spiv!
[11:48] <spiv> techbreak_: (and if you figure out where you went wrong, feel free to tell me in this channel, I'll see it in the morning)
[11:48] <techbreak_> spiv, can i have your email id ? or anything to contact if i get further problem ?
[11:49] <techbreak_> spiv, okie fine if you are regular in this channel :) i will surely tell .. whether i succeed or not ;)
[11:49] <spiv> (perhaps it's something we can improve)
[12:29] <zyga> hi
[12:29] <zyga> I'm trying to push a branch to lp:~zkrynicki/ubuntu/lucid/launch-control/packaging
[12:30] <zyga> I'm doing this for a number of my packages (I push the debian packaging to a branch under ubuntu lucid)
[12:30] <zyga> yet for some reason it fails on this branch, the error message is: "bzr: ERROR: Invalid url supplied to transport: "lp:~zkrynicki/ubuntu/lucid/launch-control/packaging": No such source package launch-control."
[12:30] <zyga> am I doing something wrong? The patter is ~username/distro/series/package/branch-name
[12:31] <maxb> zyga: Why not push the Debian packaging under debian?
[12:32] <maxb> Also, apparently there is no such source package as launch-control
[12:32] <zyga> maxb, that's not my problem - let's solve one problem first
[12:32] <maxb> You know you can push to lp:~zyga/debian/sid/sourcepackage/branch right?
[12:32] <zyga> maxb, there are no source packages for any of my things (I'm writing packaging for the first time) but previously it also accepted launchpad project names
[12:32] <maxb> zyga: I'm fairly sure it never did that
[12:33] <zyga> maxb, it did this for about half dozen other packages I made today and yesterday
[12:33] <zyga> maxb, please have a look at https://code.launchpad.net/~zkrynicki
[12:33] <jelmer> zyga: are you sure those were packaging branches, not "regular" branches?
[12:33] <zyga> at the ubuntu/lucid brnahces
[12:33] <zyga> jelmer, please explain the difference
[12:33] <maxb> I infer that source packages already existed for the ones you've pushed already
[12:34] <zyga> nope
[12:34] <maxb> zyga: An "interesting" "feature" of launchpad is that source packages in PPAs count for this
[12:34] <zyga> I've started packaging the stuff I'm writing for linaro
[12:34] <zyga> ah
[12:34] <zyga> so if I push a package to a ppa first
[12:34] <zyga> then it's enough :D
[12:34] <zyga> that's something I might have done earlier
[12:34] <zyga> (build with pbuilder, push to ppa, push to lp)
[12:35] <jelmer> yeah, that's a bug :-/
[12:35] <maxb> see https://launchpad.net/ubuntu/+source/launch-control-tool for example
[12:35] <maxb> jelmer: But it's such a *useful* bug :-)
[12:35] <zyga> I like this bug :>
[12:35] <zyga> but in fact there are two bugs :/
[12:35] <zyga> there is no sane way to push packaging in a way that would link it from the project
[12:36] <zyga> and project -> packages is a messy path with possibly different names
[12:36] <zyga> (it just happens that some projects have same source package name as lp project name)
[12:36] <zyga> right?
[12:36] <zyga> maxb, now what is the advantage to pushing to debian/sid?
[12:37] <zyga> maxb, my situation is rather strange as I need to target lucid explicitly and I cannot use more recent features without backporting things like dh_python2 that I'd rather not do (packaging is not my expertise)
[12:37] <maxb> You said "debian packaging". I inferred you meant "Packaging for Debian"
[12:38] <zyga> maxb, well I meant the packaging as in the debian directory for bzr-builddeb in merge mode (should have mentioned that)
[12:38] <maxb> ah :-)
[12:38] <zyga> while I'd love to push those packages to debian too
[12:38] <zyga> I do not know the way
[12:39] <zyga> (that would also solve future problem for the next LTS as I'll need to target that and it would come from debian for free
[13:32] <zurgutt> Question: is there a size limit for bzr controlled project? i have one that is around 9 Gb, 18Gb with repository, 50 revisions, biggest file around 1Gb of sql.  bzr commit has started to segfault.  Ideas?
[13:33] <jelmer> zurgutt: there shouldn't be a size limit
[13:34] <jelmer> zurgutt: there are some issues with memory usage, but you should just get a MemoryError in that case
[13:34] <jelmer> zurgutt: Can you file a bug about this, with a gdb backtrace?
[13:38] <mgz> could be the dirstate pyx bug.
[13:38] <zurgutt> if you tell me how to do it
[13:38] <mgz> generally sizes > 4GB aren't well tested.
[13:40] <zurgutt> it has actually created a big problem for me so need to find solution, either fix problem, drop some history or convert to other system if thats only option, im used to bzr tho
[13:45] <mgz> is there a general guide about how to file bugs on actual crashes anywhere? does apport do the right thing?
[13:46]  * mgz doesn't know gdb
[13:49] <mgz> zurgutt: if nothing else, file a bug with the details on your repo and however much of .bzr.log gets output during the commit before it dies, and ask for more instructions
[14:53] <viciousprimate> joing #ubuntu-devel
[14:53] <viciousprimate> bah
[14:55] <jelmer> viciousprimate: boing :)
[14:55] <jelmer> zurgutt: something like "gdb --args /usr/bin/bzr commit" should land you in a gdb prompt
[14:55] <viciousprimate> :)
[14:55] <jelmer> zurgutt: after that, "run" to start bzr and then "bt" to print the backtrace when it's crashed
[14:58] <zurgutt> jelmer: roger, thanks, ill try
[15:03] <zurgutt> gdb --args /usr/bin/bzr commit
[15:03] <zurgutt> (gdb) run
[15:03] <zurgutt> Starting program:  commit
[15:06] <jelmer> zurgutt: sorry
[15:06] <jelmer> zurgutt: gdb --args /usr/bin/python /usr/bin/bzr commit
[20:00] <maxb> jam: Thanks for the bzr-cvsps-import review (though that MP was already Merged). In that instance I considered the commit message itself to convey sufficient information, hence not supplying an explicit cover letter.
[20:01] <quotemstr> Why does bzr complain about a " valid .bzr control directory, but not a branch or repository." when trying to upload to launchpad?
[20:01] <quotemstr> The local branch is fine.
[20:01] <jam> maxb: when reviewing via email, you don't see the commit message
[20:01] <jam> I actually reviewed it about a week ago
[20:01] <maxb> hmm. that seems like a bit of a flaw in the process
[20:01] <jam> but launchpad rejected it for various reasons, so I was resending my "Submit Request Failure" queue
[20:02] <maxb> quotemstr: Can you show the whole error message please?
[20:02] <quotemstr> bzr: ERROR: At lp:~dan-colascione/emacs/misc-23/ you have a valid .bzr control directory, but not a branch or repository. This is an unsupported configuration. Please move the target directory out of the way and try again.
[20:03] <jam> maxb: given that almost 100% of my submissions involve many commits, I think it is intentional
[20:03] <jam> the point of the review is to summarize what happened
[20:03] <jam> etc
[20:03] <jam> it *does* tell me if you set an explicit recommended commit message for the proposal
[20:03] <maxb> quotemstr: I would imagine a previous attempt to upload the branch failed, leaving only the basics of a .bzr control directory.
[20:04] <quotemstr> Worked fine after I deleted the branch and recreated it.
[20:04] <quotemstr> Thanks.
[20:13] <achiang> hello, having some issues branching based on an rspec
[20:13] <achiang> bzr tags gives me this partial output:
[20:13] <achiang> 1:0.133.10           ?
[20:13] <achiang> 1:0.133.11           ?
[20:13] <achiang> i want the .10 version
[20:14] <achiang> so, i say: bzr branch update-manager test -r "1:0.133.10"
[20:14] <achiang> where update-manager is FROM and test is TO
[20:14] <achiang> and i get: bzr: ERROR: Not a branch: "/home/achiang/Projects/update-manager/0.133.10/".
[20:21] <achiang> any clues?
[20:27] <maxb> achiang: "number:path" is a revspec format, so that's what it's parsing it as
[20:27] <maxb> For tags containing colons, you'll need -r tag:blah
[20:28] <achiang> maxb: what value do i use for "blah" ?
[20:28] <maxb> e.g. -r tag:1:0.133.11
[20:29] <achiang> oh, i see
[20:30] <achiang> am i just being dense?
[20:30] <achiang> achiang@oak:~/Projects/charlotte-x86/update-manager$ bzr branch update-manager charlotte -r tag:1:0.133.10
[20:30] <achiang> bzr: ERROR: The branch update-manager has no revision <RevisionSpec_tag tag:1:0.133.10>.
[20:30] <maxb> oh... I have a hunch
[20:31] <maxb> Could you try "cd update-manager && bzr branch . ../test -r tag:1:0.133.10" ?
[20:32] <achiang> bzr: ERROR: The branch . has no revision <RevisionSpec_tag tag:1:0.133.10>
[20:32] <maxb> hrm
[20:32] <achiang> does it have something to do with that ? in the output of bzr tags?
[20:32] <maxb> oh!
[20:32] <maxb> yes
[20:33] <maxb> So, there's a tag there in the branch pointing at a revision-id, but that revision-id is not present in the branch
[20:33] <achiang> hm
[20:34] <achiang> that is disturbing. i just did a straight bzr branch from lp...
[20:34] <maxb> of what?
[20:35] <achiang> https://code.launchpad.net/~ubuntu-core-dev/update-manager/main
[20:39] <maxb> achiang: thats .... interesting
[20:40] <achiang> maxb: so, any clues on how to resolve this? i do kinda need to fork the branch at that revspec
[20:41] <maxb> based on looking at the earlier 0.133.* tags, I would suggest that that revision exists on some lucid-updates branch
[20:43] <putrycy> Hi! I think that it's is not possible to make "partial push". I.e. let's assume that I have two files - both modified. And I would like to push changes just from one of them. I think that it is impossible (=hard and not straightforward) if main branch and the main have diverged.
[20:43] <achiang> hrm, ok. i guess i'll ping mvo tomorrow about this
[20:43] <putrycy> Please, correct me if I am wrong
[20:44] <putrycy> if my branch
[20:44] <putrycy> and the main branch
[20:45] <maxb> achiang: Have you tried lp:~ubuntu-core-dev/update-manager/lucid
[20:46] <maxb> putrycy: You seem to be mixing two separate issues. If you have local modifications to two files, you can certainly commit one but not the other.
[20:46] <maxb> Meanwhile, your branch diverging from main is a separate issue
[20:47] <maxb> achiang: Yes, 0.133.11 is in that branch
[20:47] <achiang> maxb: ok, i'll take a look. need to eat before my stomach digests itself.
[20:47] <achiang> maxb: thanks for the help!
[20:48] <putrycy> maxb: I know I can commit. But I'm talking about situation where one change is needed to push to the main branch. Is it possible?
[20:48] <maxb> I'm sorry, you're not explaining your problem well enough for me to be able to help.
[20:59] <putrycy> maxb: It seems that I know how to explain it: As I said before I have two branches - a main one that is shared with other users and a local one. I use the local one to commit to it my small changes and from time to time I push my work to the main branch
[20:59] <putrycy> And now I have a part that I could push to the main brach
[20:59] <putrycy> and a part that is not ready yet
[21:00] <putrycy> Question: how to push to the main branch only changes that are ready to show them to other users
[21:00] <putrycy> ?
[21:02] <putrycy> In version control systems like SVN I can commit for example only one directory. What would be equivlent with pushing changes from that directory in the case of usage of bazaar
[21:02] <putrycy> right?
[21:03] <maxb> If you're using bzr in a svn-ish way, then you'd just commit the single directory in exactly the same way
[21:03] <putrycy> and If I don't use it in a svn-ish way?
[21:04] <putrycy> what then?
[21:04] <maxb> If you've already committed a mix of changes, then you have created a problem for yourself and will need to recommit changes in sensible logical groupings
[21:06] <putrycy> OK
[21:06] <putrycy> thanks for reply
[21:07]  * jelmer_ waves
[21:18] <poolie> hi all
[23:25] <spiv> Hi poolie, welcome back
[23:27] <poolie> hi, thanks
[23:27] <poolie> how are you?
[23:34] <spiv> Pretty good, finally.  After the gastro last week I think my appetite has returned to normal today.
[23:59] <poolie> and you're pilot this week?