[00:53] <Kaspi> nobody I guess
[01:08] <jelmer__> Kaspi: hi
[01:08] <jelmer__> Kaspi: I think most people are off enjoying their Christmas.
[01:08] <jelmer__> I'm afraid I don't have any experience with the http smart server
[01:09] <Kaspi> jelmer_: hey, yes, perhaps only me messing with computers even on Christmas :)
[01:30] <maxb> You'll need to be a lot more detailed to elicit help
[01:30] <maxb> No fcgi modul Well, install it then
[01:31] <maxb> deprecation warning? Only a warning, why does that matter?
[01:31] <maxb> mod_python? Don't use it, it's deprecated/
[01:31] <maxb> mod_wsgi? Probably a good method to choose. You'll need to explain where you ran into trouble
[01:51] <Kaspi> maxb: how to install the fcgi module? do you know? (on a linux actually)
[01:57] <maxb> apt-get install python-flup sounds promising
[01:58]  * maxb gets annoyed with mod_wsgi
[02:04] <maxb> Kaspi: Right, I've got mod_wsgi working just by following your linked documentation. My only problem was that when I got a path wrong, there was no logging whatsoever to help me figure this out
[02:05] <Kaspi> maxb: where did you put the apache configuration? into a virtual host?
[02:07] <Kaspi> and btw what if you had more than one repository on a virtual host? e. g. /code and /code2
[02:11] <maxb> Whether or not you put the config into a virtual host is dependent on how you want to fit the configuration into the rest of your Apache setup.
[02:12] <maxb> In the examples, /code is not a bzr repository, it is a tree which can contain any number of bzr repositories
[02:14] <Kaspi> ooh... aha!
[02:38] <Kaspi> maxb: well I have set it up for wsgi now and how to actually find out if things work?
[02:39] <Kaspi> I can't push to any repository, so I guess it doesn't work
[02:41] <Kaspi> does the wsgi script output any kind of html when the root directory is accessed for example?
[02:43] <maxb> no
[02:43] <maxb> and as for push, the example is in read-only mode
[02:44] <Kaspi> maxb: yes, I set that one to False
[02:44] <Kaspi> but anyways, when trying to push, this is print out: bzr: ERROR: Cannot lock LockDir(http://host/repo/.bzr/branch/lock): Transport operation not possible: http does not support mkdir()
[02:52] <Kaspi> can anyone see a problem? http://pastie.org/1405842
[02:59] <maxb> sounds rather like it's not talking over the smart server at all
[03:00] <maxb> Try using bzr+http:// not http://
[03:05] <Kaspi> maxb: this one says: "Generic bzr smart protocol error: Invalid http response for http://bzr.omicron.e0.cz/test/.bzr/smart: Unknown response code 403" (which means no access), but there's no /smart directory in that location
[03:07] <Kaspi> should there be a "smart" directory under .bzr/ ?
[03:10] <jelmer__> Kaspi: the configuration in apache should kick in and call bzr before it even checks to see if there's anything there
[03:21] <Kaspi> hm, I wonder what's wrong then
[04:03] <Kaspi> what is the enable_logging argument for? couldn't it help?
[04:54] <Kaspi> never mind, go to go now, I'll continue later with this.. thanks maxb and jelmer_ so far
[08:09] <aaronfay> I have a problem with a commit, I have a folder /plugins/ and I committed it and it's contents by accident, I really just wanted to commit the contents, so I uncommitted, reverted all the files.  Now when I add the child folder that I *do* want, it adds the original /plugins/ parent folder.  Any thoughts?
[08:09] <aaronfay> By 'contents' I mean 'content folder'
[08:09] <aaronfay> as in /plugins/some-plugin/
[08:15] <aaronfay> Mmm, appears it has to add the parent folder, OK, I'll work around it
[11:15] <Kaspi> hey!
[11:16] <Kaspi> can anyone see the problem http://pastie.org/1405842 which doesn't let you push to http://bzr.omicron.e0.cz/test ? (no authentication)
[12:48] <Kaspi> can anyone see the problem http://pastie.org/1405842 which doesn't let you push to http://bzr.omicron.e0.cz/test ? (no authentication)
[12:58] <jelmer_> Kaspi: where are you requiring authentication?
[12:59] <Kaspi> jelmer_: nowhere yet
[12:59] <Kaspi> (to test it out)
[13:00] <jelmer_> Kaspi: Then I don't understand your question.
[13:04] <Kaspi> jelmer_: just try, for example: bzr branch bzr+http://bzr.omicron.e0.cz/test
[13:05] <Kaspi> jelmer_: the things are set up, but you won't branch any revision because of 403
[13:06] <jelmer_> Kaspi: not sure, it's .bzr/smart specifically that's refused
[13:07] <jelmer_> is WSGI not allowed there for some reason?
[13:07] <Kaspi> jelmer_: yeah, and there's no "smart" directory under .bzr/
[13:08] <Kaspi> jelmer_: you can see the whole apache configuration in the pastie I post, the WSGI should be allowed in the whole virtualhost
[13:08] <Kaspi> jelmer_: WSGIScriptAliasMatch ^/.*/\.bzr/smart$ /srv/bzr/bzr.wsgi does the trick I guess
[13:10] <jelmer_> Kaspi: I'm not familiar with the ins and outs of mod_wsgi
[13:12] <Kaspi> jelmer_: it is set up as according to this http://doc.bazaar.canonical.com/bzr.2.2/en/user-guide/http_smart_server.html#mod-wsgi
[13:36] <Kaspi> I don't really get why it doesn't work :-/
[13:54] <jelmer_> Kaspi: I'm not sure either
[15:09] <Kaspi> can anyone see the problem http://pastie.org/1405842 which doesn't let you push to http://bzr.omicron.e0.cz/test ? (no authentication)
[16:06] <mgz> hm, jelmer has been squishing 2.7 things over xmas but missed the bug tag and discussion in here with vila the other day.
[16:09] <jelmer_> hi mgz
[16:11] <mgz> don't suppose launchpad has a mass-tag edit facility anywhere, seems everyone else decided on "python27" since we started filing ones under "python2.7"
[16:11] <mgz> jelmer_: I've put some comments on things you submitted.
[16:15] <jelmer_> mgz: thanks
[16:15] <jelmer_> mgz: fwiw bug 637674 doesn't have a python27 tag either..
[16:28] <mgz> okay, put up an mp fixing that.
[16:47] <jelmer_> cool
[16:48] <jelmer_> anyway, looks like we're in Germany now and roaming chargees are crazy
[16:48]  * jelmer_ -> offline
[16:52]  * Kaspi is praying for a working smart server
[16:53] <Kaspi> do you trust in a god?
[17:07]  * maxb thinks Kaspie should pray less and debug more
[17:07] <maxb> Hmm
[17:07] <maxb> Has anyone ever broached the idea of adding authentication to the smart server protocol?
[17:08] <maxb> I'm thinking of something much like how SASL-based auth was added to the svnserve protocol
[17:09] <Kaspi> maxbie how should I debug such a thingy?
[17:15] <maxb> Actually stating precisely how it doesn't work would be a good start.
[17:23] <Kaspi> maxb: you can try yourself: bzr branch bzr+http://bzr.omicron.e0.cz/test
[17:27] <Kaspi> it says just 403
[17:27] <Kaspi> for .bzr/smart
[17:29] <maxb> You've not granted access to the WSGI script directory
[17:33] <Kaspi> maxb: I've been told that the script must run out of a web root. Do you mean this kind of access?
[17:35] <maxb> There's a difference between being mapped anywhere into the URL namespace, and being granted access
[17:38] <maxb> Move the WSGIApplicationGroup out of the <Directory>
[17:38] <maxb> Change the AllowOverride All to AllowOverride None, unless you have a VERY good reason to allow that
[17:39] <maxb> Create a new <Directory /srv/bzr> block and move the Order and Allow directives there
[17:47] <maxb> I've just pushed fine to my local http smart server
[17:54] <Kaspi> maxb: you win the golden cookie, little corrections and it works!
[17:54] <Kaspi> maxb: thanks! :)
[17:55] <Kaspi> maxb: btw why not to allow overrides?
[17:56] <Kaspi> maxb: because people could push .htaccess files there?
[18:04] <lifeless> mgz: so
[18:04] <lifeless> mgz: I stlil don't understand why re.sub is a *problem* for you
[18:04] <lifeless> I understand it does \ escape sequence evaluation
[18:33] <maxb> Kaspi: yes, exactly (about the .htaccess)
[18:38] <Kaspi> maxb: yeah but the repositories are not creating a working tree by default (or they seem so), so if I won't ever run a bzr update server side, it should matter, should it?
[18:38] <Kaspi> shouldn't, should it? *
[18:55] <maxb> I suppose not, but it is still good practice to never enable AllowOveride unless you need to
[18:56] <maxb> and, in theory, someone could still upload a .htaccess file using the bzr vfs calls
[21:03] <mgz> lifeless: C:\temp has a backslash-t in it. see the example I pasted in the log the other day.
[21:03] <mgz> in the wider scheme of things, it's a problem because string interpolation on paths is a problem.
[21:04] <mgz> but changing from re.sub is just... well, I don't see what problem you have with str.replace is the point.
[21:04] <mgz> you appear to want a straight string substitution, and none of the fancy features regular expressions give.
[21:32] <lifeless> mgz: I'm likely to break out a full posix shell variable expansion implementation in python one of these days
[21:33] <lifeless> mgz: see for instance the bug on shell expansion support - and windows doesn't have a handy shell to delegate to
[22:03] <mgz> with all the related horrors of shell scripting?
[22:04] <mgz> unquoted parameters and security holes.
[22:05] <fullermd> Oh, don\\\\\\\\\'t exaggerate...
[22:09] <Noldorin> hello. how can i overwrite a remote branch when pushing?
[22:09] <Noldorin> it gives me complaints about the remote branch having diverged
[22:09] <Noldorin> (i don't care about the remote branch)
[22:12] <darren> hey
[22:13] <darren> if I run bzr status, it tells me a couple of files have been removed. But if I run bzr commit, it tells me PointlessCOmmit (No changes to commit).. any ideas what gives?
[22:14] <darren> specifically: http://pastebin.com/8FWi41dj
[22:30] <fullermd> Noldorin: push --overwrite
[22:30] <Noldorin> ah thanks
[22:30] <fullermd> darren: Dunno.  Maybe it some oddity with -x?  Try saving off the changes to settings and revert'ing it, then committing without -x.
[23:20] <darren> fullermd: weird, but true
[23:20] <darren> thanks!