[00:00] <mgrandi> bzr is just hanging, git actually gave output from what ssh said (permission denied: publickey)
[00:00] <mgrandi> although i think its something else cause now that the key is working bzr is still hanging haha
[01:23] <thumper> ping bzr folks
[01:24] <thumper> why did SmartServerStreamMedium suddenly need a timeout?
[01:24] <thumper> especially when the default params for the derived classes are to pass through None?
[01:24] <thumper> now it complains horribly if you don't specify one
[01:24] <thumper> like my old code did
[01:24] <fullermd> Maybe it was naughty?
[01:24] <thumper> there is no indication in the methods what it should contain
[04:57] <jam> thumper: I did the timeout stuff, I would happily merge a patch that set the default to None.
[04:57] <jam> IIRC, at the time I did it, I wanted to make sure I audited all the callers
[04:58] <jam> and there wasn't anywhere else in the code that was using it
[04:58] <jam> I didn't realize it was being poked at externally.
[06:39] <thumper> jam: :-) public api and all that
[07:43] <hongyun> Hi, I set the ssh through http_proxy by using corkscrew. And add Host *.launchpad.net
[07:43] <hongyun> ProxyCommand /usr/local/bin/corkscrew www-proxy.us.oracle.com 80 %h %p
[07:43] <hongyun> User carolliu to ~/.ssh/config file
[07:43] <hongyun> But I still can not push code to launchpad
[07:43] <hongyun> Can anyone help me to see the problem?
[08:16] <hongyun> Is there anyone meet this error?
[08:16] <hongyun> bzr: ERROR: Connection error: curl connection error (SSL certificate problem: self signed certificate in certificate chain)
[08:16] <hongyun> How to deal with it?
[08:23] <jam> hongyun: you can try running 'bzr XXX -Derror' to get a traceback, but I don't quite understand why curl is involved in the above chain of events.
[08:23] <jam> If you aren't logged into launchpad and are using 'lp:' urls, then we may be trying to connect to http://...launchpad.net to try and resolve the lp:
[08:23] <jam> so you could try using 'bzr+ssh://' directly.
[08:24] <hongyun> I try bzr+ssh://  ,but get an error
[08:24] <hongyun> bzr: ERROR: Don't know how to handle SSH connections. Please set BZR_SSH environment variable.
[08:24] <hongyun> I have already installed corkscrew
[08:25] <jam> hongyun: if you are getting that, then you don't have an 'ssh' executable in your path
[08:25] <hongyun> add Host *.launchpad.net  ProxyCommand /usr/local/bin/corkscrew www-proxy.us.oracle.com 80 %h %p User carolliu to ~/.ssh/config file
[08:25] <jam> which I think is necessary to get it working with corkscrew.
[08:25] <jam> Do you have 'openssh' installed
[08:26] <jam> can you run just plain 'ssh bazaar.launchpad.net' ?
[08:27] <hongyun> Displayed :  No shells on this server. Connection to bazaar.launchpad.net closed.
[08:28] <hongyun> jam, do u know what's the problem?
[08:29] <jam> hongyun: well that means ssh is working at least, you don't have a login shell on bazaar.launchpad.net
[08:29] <jam> but you are able to connect to it, and your login with ssh keys is working.
[08:29] <jam> hongyun: so try just doing "export BZR_SSH=openssh" and then see if you can do "bzr XXX bzr+ssh://bazaar.launchpad.net/...."
[08:33] <hongyun> only openssh can work?
[08:34] <jam> hongyun: bzr understands a few ssh providers, mostly we just need to know what the right parameters to pass to ssh are.
[08:35] <hongyun> Displayed: bzr: ERROR: Unrecognised value for BZR_SSH environment variable: ssh
[08:35] <jam> You can also do "BZR_SSH=sshcorp"
[08:35] <jam> hongyun: there are a few specific ones (openssh, and sshcorp), otherwise you need to pass the full path to the 'ssh' file, like: "BZR_SSH=/usr/bin/ssh"
[08:36] <hongyun> ok  I try it again
[08:40] <hongyun> jam, nothing diaplayed after I use push command till now
[08:41] <jam> hongyun: sounds like it is working, but I couldn't tell you for sure.
[08:41] <jam> You can run 'ps' and see what processes are running (ps -ef is what I usually use)
[08:46] <hongyun> other method can  prove bzr work correctly?
[08:51] <jam> hongyun: if you tell me where you are pushing things, I can check if the data is there.
[08:52] <hongyun> //bazaar.launchpad.net/~carolliu/carol-bzr-test/
[08:53] <hongyun> When I run bzr push bzr+ssh command , it always halt
[08:54] <jam> hongyun: there is a problem with the URL, as you need a project and a branch name, so something like bzr+ssh://bazaar.launchpad.net/~carolliu/+junk/carol-bzr-test
[08:55] <fullermd> You could also just 'bzr ping' the URL, and make sure you're at least getting end-to-end communications out of it.  The plugin's cheap to grab on older versions.
[08:58] <hongyun> Displayed error :bzr: ERROR: Unrecognised value for BZR_SSH environment variable: /usr/bin/ssh
[08:58] <hongyun> I am sure ssh on my machine can work
[09:00] <mgz> morning!
[09:01] <jam> hongyun: can you give your bzr version? "bzr --version"
[09:01] <jam> Supporting path-based ssh was a newer one.  Also 'ssh' information (ssh -v I believe)
[09:02] <hongyun> Bazaar (bzr) 2.3.3
[09:02] <hongyun> I install it on solaris
[09:03] <mgz> that has BZR_SSH settable to a specific path, but not the solaris ssh fix maybe?
[09:04] <hongyun> So I must install openssh?
[09:05] <jam> mgz: I don't see "solaris" mentioned in release notes past about 1.18, though it may not be mentioned in that form. Do you know what the patch is?
[09:06] <mgz> trying to find it in my mail archive but failing currently
[09:06] <hongyun> I can checkout code from launchpad.net . But can not push code
[09:08] <mgz> question 105156 has some info
[09:08] <mgz> and links bug 544786
[09:09] <mgz> which isn't fixed... so it seems either using sunssh as openssh or installing real openssh is what people have done in the past
[09:10] <jam> hongyun: according to https://answers.launchpad.net/bzr/+question/105156 the issue is that Solaris has different output for "ssh -v" but if you set "BZR_SSH=openssh" it will still work.
[09:10] <jam> We just fail to auto-detect the ssh vendor because of changes in Solaris.
[09:10] <hongyun> ok  I try
[09:10] <mgz> going by bug 374700 it would not be hard to contribute a branch fixing it
[09:10] <jam> mgz: we only need the vendor type to pass the right parameters, and if sunssh is a fork of openssh, it is likely to accept the same params.
[09:11] <mgz> jam: indeed.
[09:12] <hongyun> woo, It works . Thanks jam and mgz
[09:13] <jam> hongyun: can you add the output of "ssh -V" (or ssh -v) to bug #544786?
[09:13] <jam> It is hard for us to fix things like that on platforms we don't have access to.
[09:13] <hongyun> OK , no problem
[09:16] <Lo-lan-do> Hi all
[09:17] <Lo-lan-do> I'm working on adopting the loggerhead package, but I'm encountering problems.
[09:17] <hongyun> Only this displayed :
[09:17] <hongyun> ~$ ssh -V
[09:17] <hongyun> Sun_SSH_2.0, SSH protocols 1.5/2.0, OpenSSL 0x100000af
[09:17] <hongyun> Does it help?
[09:18] <Lo-lan-do> Loggerhead uses the YUI library, but the embedded copy is very old (3.0.0pre2 or something), and it should be ported to current YUI 3.5.
[09:18] <Lo-lan-do> Would anyone care to help?
[09:18] <mgz> Lo-lan-do: there may be some interest or more info in #launchpad on that front, though probably US timezones
[09:19] <Lo-lan-do> Apparently launchpad still uses the old one too.
[09:19] <Lo-lan-do> But I'll try there.
[09:19] <Lo-lan-do> Thanks :-)
[09:19] <mgz> launchpad have recently done various updating of yui libraries
[09:20] <mgz> and the main launchpad code was moved to 3.5 by rick_h
[10:11] <mira|AO> when I do a 'bzr dpush' into git, can the git side be updated later, when there are new commits and merges on the bzr side?
[10:11] <mira|AO> (assuming it doesn't *also* fail, like git cloning from bzr and bzr pushing into svn does…)
[10:13] <Lo-lan-do> You can bzr dpush again, yes
[10:13] <Lo-lan-do> I regularly do that :-)
[10:13] <mira|AO> ok
[10:14] <mira|AO> maybe this is the way out of my current problem then, unless someone steps up and fixes bzr-svn ☻
[10:15]  * mira|AO crosses fingers (it’s been “updating git map” for >2 hours already)
[10:34] <Lo-lan-do> loggerhead_1.19~bzr477-1_amd64.changes ACCEPTED into unstable
[10:35] <Lo-lan-do> (Without the YUI3 hackery, I suck at Javascript)
[10:36] <Lo-lan-do> mira|AO: You wouldn't have JS geeks at your place, would you?
[10:38] <mira|AO> nope, sorry
[10:39] <mira|AO> we're also all trying to scrape along
[10:40] <mira|AO> but if you /msg me some details, I can ask around on the dev list
[10:41] <mira|AO> bzr: ERROR: The file id "None" is not present in the tree <bzrlib.inventory.CHKInventory object at 0xd7d4550>.
[10:41] <mira|AO> ☹ so dpush also doesn't work
[10:42] <mira|AO> ok, last try… upgrading all svn related stuff (yuk 1.7)
[10:48] <mira|AO> ok, didn't work either with latest suckvürstchen/subvertpy
[10:48] <mira|AO> latest as in debian sid
[10:58] <jelmer> svn 1.7 changes the working copy API in a backwards incompatible (but ABI compatible) way
[11:09] <mira|AO> yeah
[11:09] <mira|AO> I'm mostly concerned about getting a bzr bound checkout pushed into either svn or git…
[11:11] <mira|AO> bbl, lunchbreak
[11:21] <jelmer> mira|AO: the exception you were seeing (The file id "None" ...) has to do with the mappings in bzr-svn, unrelated to the svn version
[12:09] <mira|AO> the 'file id is None' is actually in bzr-git
[12:09] <mira|AO> (both ways)
[14:06] <mathrick> hola
[14:06] <mathrick> I don't suppose bzr-lp supports multiple logins, does it?
[14:08] <mgz> it does, but with different bzr homes
[14:08] <mgz> so, you can have multiple configs and set BZR_HOME to change between them
[14:10] <maxb> Or you can just 'bzr lp-login foo' back and forth
[14:13] <mathrick> oh right, it uses the ssh key
[16:06]  * mirtwo tries: (gdb) b PyErr_NewSubversionException
[16:06] <mirtwo> maybe I can get a useful bt
[16:07] <mirtwo> oh well. I get a bt...
[16:08] <mirtwo>  http://sprunge.us/SOiJ
[16:09] <jlf> hi #bzr, is there a way i can find out what commit last affected a specified line in a specified file?
[16:09] <jelmer> jlf: hi
[16:09] <jelmer> jlf: 'bzr annotate FILE' should tell you
[16:10] <jlf> jelmer: ah, perfect.  thanks very much.
[16:11] <mathrick> pffft, "annotate" is PC-speak
[16:11] <mathrick> I always call it bzr blame
[16:14] <mirtwo> blame is git speak
[16:14] <mirtwo> oh no, svn speak
[16:14] <mirtwo> anyway, worse
[16:14] <mirtwo> can't decide which
[16:15] <mirtwo> hrm. putting debugging printfs into editor.c in subvertpy didn't help...
[16:15] <mirtwo> so I now got this backtrace, but... still no further
[16:15] <mirtwo> not easy to figure anything out like that
[16:20] <mathrick> mirtwo: I don't care if it's svn speak, it's accurate
[16:21] <mira|AO> I prefer to not blame people, it's the wrong state of mind in a team
[16:27] <jelmer> Lo-lan-do: thanks for taking over loggerhead!
[16:34] <Lo-lan-do> jelmer: Bleh :-)
[16:34] <Lo-lan-do> Is anyone active upstream, by the way?
[16:34] <Lo-lan-do> (So I could push my quilt patch their way)
[16:34] <jelmer> Lo-lan-do: lp:launchpad, you mean?
[16:35] <mira|AO> lo-lan-do: could be worse… look at what mess I got into, just for the forge…
[16:35] <jelmer> Lo-lan-do: you can propose it as a patch on Launchpad
[16:35] <mira|AO> hint: the mess uses wfSuppressWarnings() a lot
[16:35] <Lo-lan-do> mira|AO: Hahaha :-)