[00:10] <poolie> ah
[00:11] <poolie> i see vila updated the wiki page to be pilot again this week
[00:20] <james_w> hi poolie
[00:22] <james_w> we should have a call this week
[00:27] <poolie> ok
[00:27] <poolie> can be in 5m if you like, or earlier tomorrow
[00:28] <poolie> are you still in ca?
[00:28] <james_w> I'm not
[00:32] <james_w> I'm happy to chat now though
[01:05] <james_w> night
[04:55] <poolie> spiv, want to approve or object to https://code.edge.launchpad.net/~mbp/bzr/progress/+merge/15997
[04:55] <poolie> it's pretty small
[04:59] <spiv> poolie: sure
[05:16] <poolie> thanks
[05:28] <wgrant> poolie: So the progress bar is not easily fixable? :(
[05:28] <poolie> in the sense of making it proportional to overall progress?
[05:29] <poolie> not _easily_
[05:29] <poolie> it is possible but it will be a bit street-by-street
[05:29] <wgrant> poolie: Or just volume downloaded, which is nearly correct and much easier.
[05:29] <poolie> i guess there is a risk with this patch the underlying overall progress will be worse
[05:29] <poolie> that would be easy to add
[05:33] <poolie> wgrant: in places where we do have an accurate overall count like "building files %d/%d" that will still show - and actually be more prominent
[05:38] <wgrant> poolie: How hard is that to do for the download involved in the initial checkout?
[05:38] <poolie> wgrant: maybe i should make a faq
[05:38] <poolie> there are a few things
[05:38] <poolie> the progress bar model is that there is a tree of tasks
[05:39] <poolie> and you compute the fractional completion across that tree
[05:39] <poolie> sometimes that tree structure doesn't correspond to what actually takes time
[05:39] <wgrant> Right.
[05:39] <poolie> 2- we try to do streaming operations that by design don't know up front how much work they're going to do
[05:40] <poolie> actually are you asking about showing the total download to date, or the total that will be downloaded?
[05:41] <wgrant> poolie: Both.
[05:41] <wgrant> poolie: Surely the server can tell you somehow how much there is to stream?
[05:42] <poolie> actually
[05:42] <poolie> do, or do-not
[06:08] <assad> i did a bzr  branch lp:~ but after doing the the fetching part it is not able to copy. how can i make it copy the fetched data i.e. complete the bzr process!
[06:09] <spiv> What do you mean by "it is not able to copy"?  Perhaps pastebin the error you are getting?
[07:48] <poolie> hi vila?
[08:04] <vila> hey  poolie
[08:04] <vila> hi all
[08:06] <poolie> hi there
[08:09] <poolie> still piloting this week?
[08:11] <vila> yup, there was no volunteer and it felt wrong to leave the patches without pilot :)
[08:11] <vila> I won't be able to do that next week though :)
[08:16] <poolie> ok
[08:16] <poolie> you don't _have_ to do it this week but it could be nice
[08:17] <vila> It's ok, there are still some rough edges in *my* workflow I'd like to polish :)
[08:19] <lifeless> I would like folk to clearly say how much time it took
[08:19] <lifeless> so that people that feel busy can assess whether they can do it
[08:19] <poolie> well
[08:20] <poolie> it may be a bit elastic
[08:20] <lifeless> sure
[08:20] <poolie> so we can try "i was really busy and could only give it 2h and ... (it still seemed useful | i was thrashing and it sucked | ...)"
[08:20] <lifeless> but data will help figure that out
[08:20] <lifeless> yes
[08:22] <vila> It can vary greatly depending on whether you "just" review or do some coding to help the patch succeeds
[08:24] <vila> but anyway, raw numbers says 14hours last week for me
[08:25] <vila> courtesy of the hamster time tracker applet (it rocks :)
[08:27] <lifeless> jwow
[08:56] <bialix> hello bzr
[09:23] <saedelaere> hi
[09:23] <saedelaere> i was asking this already a week ago, but i had to leave and i'am still unsure what to do.
[09:24] <saedelaere> i started using bazaar with the solo way described here
[09:24] <saedelaere> http://doc.bazaar.canonical.com/bzr.2.0/en/user-guide/solo_intro.html
[09:27] <saedelaere> now, i do want to release a stable version. in order to be able to continue development i need to make a new branch?!
[09:27] <saedelaere> so what i want to do is developing new code and have the original code of my stable release so i can fix errors if any occur.
[09:27] <saedelaere> i was reading about branches and so on, but i'Am not sure what to do, and a bit afraid i could mess up things...
[09:35] <saedelaere> ok for example i start with,
[09:35] <saedelaere> bzr branch bzr://foo-bar-baz stable.081
[09:35] <saedelaere> so this will create a new directory with the name stable.081
[09:35] <saedelaere> do i have to make this in the root directory of my main branch or i can i do it everywhere?
[09:35] <saedelaere> will the new branch and my main branch share history?
[09:36] <saedelaere> if it is important, the project is hosted by sourceforge.net
[09:46] <bialix> saedelaere: to ensure 2 branches will share history you need to put them into shared repository
[09:46] <bialix> otherwise 2 branches will be completely standalone
[09:47] <bialix> saedelaere: you can create as many branches as needed, you won't mess things
[09:48] <saedelaere> ok, this sounds good :)
[09:48] <saedelaere> the shared repo thing, is this recommended? or would you run different branches?
[09:48] <bialix> shared repo is recommended in most cases
[09:48] <bialix> it helps to reduce the size of storage for related branches
[09:49] <bialix> also creating copy of new branch inside shared repo is much quicker
[09:49] <bialix> because you don't need to copy entire history
[09:50] <saedelaere> ok then i have to reread the documentation about shared repos. but in general it is now problem to move my current project to a shared repo with the complete history?
[09:50] <saedelaere> s/now/no
[09:52] <saedelaere> ehm, when i started a project like this
[09:52] <saedelaere> http://doc.bazaar.canonical.com/bzr.2.0/en/user-guide/starting_a_project.html#starting-a-new-project
[09:52] <saedelaere> isn't this already a shared repository?
[09:53] <bialix> if you already executed init-repo command, then it is
[09:53] <bialix> you can check this easily with `bzr info` command
[09:53] <bialix> if it says about shared repo location, then your branch already uses shared repo
[09:54] <bialix> if you created shared repo after you created branch, you can reconfigure your branch to use shared repo, with `bzr reconfigure` command
[09:54] <saedelaere> so this doesn't look like a shared repo.
[09:54] <saedelaere> http://www.pastebin.org/64665
[09:55] <bialix> what says `bzr info bzr+ssh://christianrapp@tv-viewer.bzr.sourceforge.net/bzrroot/tv-viewer/` ?
[09:55] <bialix> your local branch is not used shared repo, yes
[09:58] <saedelaere> bialix: http://pastebin.org/64667
[09:58] <bialix> you have shared repo and the branch on the server
[09:59] <bialix> usually branches placed one level behind repo root
[09:59] <bialix> but your setup will work too
[09:59] <saedelaere> hmm it was no special setup :D
[10:00] <bialix> you just need to create new branches under tv-viewer.bzr.sourceforge.net/bzrroot/tv-viewer/
[10:04] <saedelaere> bialix you mean something like this?
[10:04] <saedelaere> bzr branch bzr://tv-viewer.bzr.sourceforge.net/bzrroot/tv-viewer/ bzr+ssh://christianrapp@tv-viewer.bzr.sourceforge.net/bzrroot/tv-viewer/stable.081
[10:04] <bialix> yep
[10:07] <saedelaere> bialix: http://pastebin.org/64668
[10:07] <saedelaere> this is it , right?
[10:07] <saedelaere> ha
[10:07] <bialix> saedelaere: perfectly
[10:07] <saedelaere> thank you very much for your time!
[10:07] <bialix> happy to help
[10:09] <saedelaere> cd ls
[10:09] <saedelaere> ups
[10:16] <saedelaere> bialix one more question. if i want to work on the code for the stable release is it better to do a checkout of the branch or do it with bzr branch?
[10:16] <bialix> it depends how you feel more comfortable
[10:17] <bialix> checkout will automatically copy all your new revisions on the server
[10:19] <bialix> with branch you need push your new revisions from time to time manually
[10:27] <saedelaere> i'Am currently searching for an alternative server where i could push my bazaar repos. i know of sourceforge and launchpad, is there a list somewhere of free alternatives?
[10:27] <bialix> also savannah
[10:27] <lifeless> and that .fr hosting service
[10:27] <bialix> but you can use any cheap web-hosting with ssh/ftp access
[10:27] <bob2> berlios
[10:28] <bialix> hmm, I did not know
[10:28] <bialix> hi all
[10:30] <lifeless> hi ;)
[10:31]  * bialix waves to lifeless
[10:36] <saedelaere> ok there are some alternatives out there. have to check them. thanks for all your help!
[15:02] <jam>  morning all
[15:18] <rubbs> morning jam
[15:44] <bialix> morning jam
[16:35] <vila> jam: morning !
[16:51] <jam> hey vila
[16:52] <vila> jam: you're still the RM for 2.0.3 right ?
[16:52] <jam> Supposedly
[16:52] <jam> I didn't do that good of a job on 2.0.2 / 2.1.0b3 :)
[16:52] <jam> since they *still* aren't officially announced
[16:52] <jam> but yeah, I should be doing the 2.0.3 / 2.1.0b4 prep work
[16:53] <vila> jam: I see, "procastinate long enough and it will disapear" ftw !
[16:53] <jam> or I get fired and somebody else does the work, yeah something like that
[16:53] <jam> but nobody really likes releasing
[16:54] <vila> and you don't seem to get fired any time soon :)
[16:54] <jam> at least, nobody is clamoring to do it
[16:55] <vila> is that the same nobody that wanted to be patch pilot this week ? :-)
[16:55] <jam> vila: yeah, nobody is pretty active lately
[16:56] <vila> nobody: take it easy, we don't want you to burn out !
[16:57] <vila> We should do more Call For Participation for both patch pilot and release manager (though PP may have higher requirements surprisingly...)
[16:58] <vila> And I've yet to be a real RM (that may be due to all my mistakes when I've been acting RM :)
[16:58] <rubbs> I'd help out, if I knew what I was doing ;) as it happens, I'm still bumbling around with doc patches. Nothing serious yet.
[16:59] <vila> doc is serious, you may do even more damage with doc than with code !
[16:59] <vila> haha, only half joking
[17:01] <rubbs> fair enough
[17:24] <jam> rubbs: "It is safe to rm -rf / at any time"
[17:24] <jam> would be a pretty nasty doc bug :)
[17:24] <jam> sorry
[17:24] <jam> "sudo rm -rf /"
[18:17] <hsn> who removed eclipse bzr plugin download from bazaar-vcs.org?
[18:18] <hsn> it was at http://bazaar-vcs.org/releases/3rd-party/bzr-eclipse/
[18:24] <mtaylor> arrrg. bzr-notify bubbles could really go away when I mouse over them.
[18:26] <jam> hsn: It was never 'actually' there, but we had a redirect to the actual place
[18:27] <jam> it got lost in the shuffle of transitioning to bazaar.canonical.com
[18:28] <jam> hsn: http://qbzreclipse.tribalo.com.ar is the actual hosting location
[18:31] <hsn> jam, it seems to be different plugin
[20:07] <jam> ah sorry, you wanted bzr-eclipse not qbzr-eclipse, I'm not sure where the official location is, or if it just hasn't updated in a long time
[21:28] <davidstrauss> What happens if you use the --directory option on a merge and there's a conflict?
[21:28] <davidstrauss> Or must --directory be a local tree?
[21:52] <lifeless> jam: hi
[21:53] <lifeless> jam: wondering if you'd like to review the subunit2junitxml patch I wrote for you :)
[21:57] <poolie> jam: http://manpages.ubuntu.com/manpages/jaunty/en/man7/tcp.7.html
[22:19] <jam> hi lifeless, I knew I'd see you around :)
[22:21] <lifeless> jam: :)
[22:21] <lifeless> personal itches getting scratched
[22:27] <lifeless> jam: jml is releasing testtools this lunchtime, and reviewing my subunit protocol branch. After which, we'll be able to land my subunit branch of bzr
[22:27] <lifeless> which has had only comments so far ;)
[22:28] <lifeless> but it makes selftest --parallel include the log file, etc.
[22:29] <jam> poolie: looking at the openssh source code, the only time SO_SNDBUF is mentioned is in a "get_snd_buf_size" it doesn't seem to ever get set.
[22:30] <jam> so a) there doesn't seem to be any way for us to set it, and b) ssh itself doesn't seem to set it to anything bigger than the 4k default
[22:32] <lifeless> LD_PRELOAD ? :P
[22:32] <jam> you have to set the option on the socket between the time when you allocate it and when you call connect or listen
[22:34] <lifeless> jam: yes, wrap socket()
[22:35] <lifeless> on b.l.n. we don't use openssh anyway
[22:35] <jam> sure
[22:35] <jam> paramiko doesn't have a hook point for it either
[22:35] <jam> I don't know about twisted
[22:35] <jam> I was hoping to test it via chinstrap
[22:35] <jam> but that looks to be difficult
[22:38] <jam> as I don't think I can just "bzr serve" and I can't get at the ssh socket options...
[22:40] <lifeless> use a linode or something
[22:40] <lifeless> or an ec2 instance
[22:40] <lifeless> (europe zone)
[22:41] <lifeless> but as I think I've said already, get it sorted without ssh first, as ssh is known to mess with this.
[22:41] <jam> Well, just getting it out of my home machine would be a benefit.
[22:41] <jam> Even if it was US
[22:41] <lifeless> we do have an allowance for clcoud experiments
[22:41] <jam> I had tried vila's babune, but his upload cap is lower than my download cap
[22:41] <jam> On the plus side, we fully utilize it without problems :)
[22:43] <tsmith> bzr on Win32 is *kicking my ass* today!
[22:43] <tsmith> when I do a $ bzr diff on a *single* file, either via the commandline or Tortoise, i get this:  bzr: ERROR: bzrlib.errors.NoSuchRevision: KnitPackRepository('file:///c/www/adportal/.bzr/repository/') has no revision
[22:44] <jam> tsmith: is this on a bzr-svn revision?
[22:44] <tsmith> yes
[22:44] <tsmith> jam: yes
[22:45] <jam> this is probably the "bzr diff doesn't work with ghosts" issue
[22:45] <jam> tsmith: bug #393837
[22:45] <tsmith> what's a ghost?
[22:46] <tsmith> something VERY peculiar happened to SVN... branches i created w/ bzr-svn *disappeared*
[22:46] <tsmith> and then this problem happened
[22:46] <tsmith> i lost like 4 weeks worth of work
[22:47] <jam> tsmith: a ghost is a revision which is mentioned but not actually present
[22:48] <jam> pushing from bzr => svn will tend to create them for merge revisions
[22:48] <tsmith> how do i recover it???
[22:48] <jam> because svn will not actually contain the merged revisions
[22:48] <jam> I'm not really sure what the issue is for branches disappearing
[22:48] <jam> did someone delete them ?
[22:48] <jam> Can you check the svn log directly?
[22:48] <jam> as for bzr
[22:48] <tsmith> no i mean i had ~40 comits in one then i came back 3 weeks later and only the initial "create branch" revision was there
[22:49] <jam> well, if someone is mutating your svn repo in destructive ways, it is going to cause problems for br
[22:49] <jam> bzr
[22:49] <tsmith> i was thinking hard drive failure, honestly
[22:50] <jam> so if the revisions were originally in bzr somewhere
[22:50] <jam> then 'bzr pull'ing from that bzr repo
[22:50] <jam> rather than via svn
[22:50] <jam> would be a way to fill in the ghosts
[22:51] <tsmith> i dont know
[22:51] <tsmith> how do i fix ghost revisions?
[22:52] <jam> tsmith: you 'bzr pull -r revid:XXXX' whatever revision was missing from wherever they actually exist
[22:53] <tsmith> revno: 156 [merge]
[22:53] <tsmith>  would that be a "ghost revision"?
[22:53] <jam> presumably someone pushed changes to svn
[22:53] <tsmith> how do i find out the missing revision?
[22:53] <jam> and then you pulled from svn rather than the original bzr branch
[22:53] <jam> tsmith: at the moment, by figuring out what revision is part of the traceback
[22:53] <tsmith> you're speaking greek
[22:54] <tsmith> do you mean the missing revision is mentioend in the crash debug message?
[22:55] <tsmith> http://pastebin.ca/1714989 <-- error
[22:55] <tsmith> revision tsmith@jwt.com-20091029161028-ublx1l0861lbmxtq
[22:55] <tsmith>  
[22:55] <tsmith> is *that* the "revision number"?
[22:56] <jam> tsmith: if you did the work locally in bzr, then you wouldn't have lots the work. I honestly don't have a good understanding of how your are interoperating with bzr and svn
[22:56] <tsmith> ok
[22:56] <tsmith> it's like this
[22:56] <jam> tsmith: that is the "revision id"
[22:56] <tsmith> i am programming a website
[22:56] <jam> and yes, that is what you are looking for
[22:56] <tsmith> i have several branches for each defect/feature
[22:56] <tsmith> when i program one, i branch the SVN repository,i commit locally, push to SVN branch, test on server
[22:57] <tsmith> then when it's time to do the next feature, I bzr switch to another SVN branch, commit locally, push to SVN branch, test on server
[22:57] <tsmith> i assumed SVN was storing copies of my work, so i didn't
[22:57] <tsmith> 3 weeks go by on different projects, come back, work is gone, had just reformatted my computer too
[22:58] <jam> tsmith: are you using a 'shared repository' locally?
[22:58] <tsmith> no via https:// over the internet
[22:59] <jam> tsmith: so, #1, you'll want to do "bzr init-repo ." above all of your feature branches
[22:59] <jam> that will make creating them much faster, etc, etc.
[22:59] <jam> then you can do
[22:59] <jam> bzr reconfigure --use-shared
[22:59] <jam> in each one
[22:59] <tsmith> i asked someone about that several weeks ago
[22:59] <jam> at that point, I *think* your ghosts will be gone
[22:59] <jam> as all the branches will talk to eachother and fill in appropriate history
[22:59] <jam> not positive, though
[23:00] <tsmith> did you know that cygwin bzr doesn't support https:// authentication?
[23:04] <jam> tsmith: It probably does if you find the python-ssl package for cygwin somehow
[23:05] <jam> personally, I find cygwin bzr to be crappy to start with, so I've avoided it entirely
[23:09] <jam> lifeless: well, there also seems to be: "They can        be    set    globally    with   the   /proc/sys/net/ipv4/tcp_wmem   and        /proc/sys/net/ipv4/tcp_rmem files" though that would mean having root on chinstrap, which isn't going to happen :)
[23:45] <poolie> james_w: hi, still around, want to chat?
[23:45] <james_w> hi poolie
[23:45] <james_w> not tonight really
[23:46] <poolie> ok
[23:46] <poolie> no problem
[23:47] <poolie> we could talk tonight my time if you like?
[23:47] <poolie> just suggest a time?
[23:47] <james_w> no, I've got a personal day tomorrow
[23:48] <james_w> so either tomorrow night or thursday morning your time is best for me