[00:17] <stewart> hi! i'm using bzrlib to write a script that iterates through a sequence of revisions and applies them to a different tree. I'm using show_diff_trees to extract a patch, but I need to skip over merge revisions so I don't end up trying to do work twice. how do i work out if a revision is a merge revision?
[00:17] <spiv> stewart: look at the parents of the revision
[00:17] <spiv> merge revisions have >1 parent revision
[00:18] <lifeless> stewart: you could use use bzr-rewrite ;)
[00:18] <stewart> spiv, ahh..
[00:18] <stewart> lifeless, haven't been able to get it even to remotely work for my branches.
[00:18] <stewart> lifeless, plus, i need to apply transformations :)
[00:18] <lifeless> stewart: its designed to do that
[00:18] <stewart> lifeless, because regex over patches is an excellent idea :)
[00:18] <lifeless> its not entirely fleshed out, but thats the intent
[00:19] <spiv> But I agree with lifeless that this is something that bzr-rewrite ought to support.
[00:19] <spiv> (even though I can totally believe it's not quite there yet)
[00:21] <stewart> i couldn't get it to apply just a few revisions :)
[00:25] <spiv> stewart: you tried 'bzr replay -r N..M -d TARGET FROM_BRANCH'?
[00:26] <stewart> spiv, IIRC (was a few weeks ago), yes. needs mapping of paths thoguh.
[00:27] <stewart> oh wow... i'm now blowing up when iterating through revisions that have been merged from another repository... doh.
[00:29] <spiv> Hmm, in theory that should already support arbitrary transformations too, via --merge-type, if you implement a plugin that provides the transformation you want as a new merge type...
[00:29] <spiv> (A fairly cumbersome way to do it, obviously)
[07:06] <vila> hi all
[07:06] <fullermd> What?!  You just said that yesterday!
[07:07] <vila> hmm, then today is another day. QED.
[07:08] <fullermd> I can't deal with another day.  I haven't even reconciled myself to yesterday yet.
[07:16] <jelmer> 'morning vila, fullermd
[07:16] <jelmer> fullermd: hey, aren't you in the US?
[07:16] <jelmer> hmm, maybe that was another fullermd
[07:16] <fullermd> Oh god, there are two of me?!
[07:16] <fullermd> I am, yes.
[07:17] <jelmer> fullermd: isn't it the middle of the night for you then?
[07:18] <fullermd> Well, not for _me_.  Maybe for the other losers in my TZ, but that's their own fault for not synchronizing with the One True Time (i.e., mine).
[07:18] <vila> jelmer: that's because you're often up at this hour :) But fullermd don't sleep anyway
[07:18] <vila> doesn't, grr ttoyyops
[07:18] <fullermd> Sleep is for wimps.
[07:18] <fullermd> Happy, healthy, well-rested wimps.  But wimps nonetheless.
[07:19] <jelmer> vila: Heh, perhaps :-) I guess it's not yet midnight on the west coast
[07:19] <fullermd> I'm in the same TZ as jam.  0119 now.
[07:19]  * jelmer is slowly transforming from a student into a civilian
[07:19] <fullermd> Way better time to get work done than at 1319.  People call me then.
[07:20] <jelmer> heh, fair enough :-)
[07:20] <vila> fullermd: you'd better unplug the phone once and for all
[07:20] <fullermd> I tried that once.  Eventually somebody actually physically came by.  That's way worse  ;)
[07:21] <jelmer> vila: bzr 2.3b3 uploaded to unstable
[07:21] <poolie> hi jelmer, vila
[07:21] <vila> jelmer: cool
[07:21] <vila> poolie: hey !
[07:21] <jelmer> 'evening poolie!
[07:23] <GaryvdM> Morning all
[07:23] <vila> . o O ( GaryvdM is following fullermd into the never-sleeping land...)
[07:23] <vila> GaryvdM: hey !
[07:23] <GaryvdM> Hey
[07:24] <fullermd> Individually, we are merely groggy and harassed.  But together, we are a force to b...zzzzzzzz...
[07:25] <GaryvdM> My sleep habbits are much better than when I was at the ice rink. I often woked till 6am. Now I only work to 1:30 am..
[08:12] <vila> ping LOSA, any news or feedback about rt #41340
[08:23] <vila> jelmer: I can't see 2.3b3 with `rmadison -u debian bzr` . Is there some lag or something failed ?
[08:25] <jelmer> vila: it usually takes a while before it's processed. I believe the publishing happens 4 times a day at the moment.
[08:25] <vila> jelmer: ok, so lag it is, thanks
[08:25] <bob2> iirc rmadison depends on the mirror pulse
[08:25] <jelmer> vila: http://packages.qa.debian.org/b/bzr.html is up to date
[09:22] <vila> ping LOSA, any news or feedback about rt #41340
[14:38] <GaryvdM> jelmer_ : I'm getting this error with my bzr 2.3b3 build: http://pastebin.ubuntu.com/528719/  I'm not sure where to start looking for the problem. Any pointers?
[14:39] <jelmer_> GaryvdM: no idea, sorry
[14:39] <GaryvdM> Ok
[16:19] <GaryvdM> jelmer_: I figured out the issue. The installer was including a "SHFOLDER.dll". Removing this fixed it. Really odd.
[16:30] <jelmer_> GaryvdM, hmmok
[16:31] <jam> GaryvdM: fairly often you can get system dlls that aren't properly marked, and thus py2exe doesn't filter them out automatically
[16:31] <jam> there is a decent sized list of ones we've found so far :)
[16:35] <vila> GaryvdM: approved
[16:35] <GaryvdM> Hi jam
[16:35] <GaryvdM> vila: Thanks - I'll land
[16:53] <GaryvdM> vila: I'm trying to run bzr selftest with my windows installer, but it seems to just hang.
[16:54] <GaryvdM> I'm not sure if it is something wrong with my installer, or an existing problem.
[16:54] <vila> GaryvdM: unheard of, we are at 4 or 5 failures on babune
[16:55] <vila> GaryvdM: try with -v to see which one is hanging
[16:56] <GaryvdM> vila: ok
[16:56] <vila> GaryvdM: and --no-plugins :)
[17:02] <GaryvdM> vila: blackbox.test_breakin.TestBreakin.test_breakin_harder
[17:02]  * GaryvdM goes to read test
[17:06] <vila> GaryvdM: forget it, it's blacklisted on babune
[17:07] <GaryvdM> Oh
[17:07] <GaryvdM> vila: how can I see you blacklist?
[17:07] <vila> GaryvdM: or rather, try running with -x 'TestBreakin'
[17:08] <vila> GaryvdM: and then file  a bug about it
[17:08] <vila> or check for duplicate, I've lost steam fighting its failures :-/
[17:09] <GaryvdM> vila: ok
[17:23] <jam> GaryvdM: AIUI, the issue is that signals and threads don't get along, and our test suite still leaks threads
[17:23] <jam> if you run the breakin tests by themselves, they work fine
[17:24] <GaryvdM> Oh
[17:24] <jam> trying to confirm now
[17:25] <jam> Ran 3 tests in 6.681s
[17:28] <jam> GaryvdM: so they work here, but I won't guarantee they work w the installer, etc. I also just ran "bzr selftest -s bb" and it seems to be passing
[17:28] <jam> so I don't really know how to get the bad interaction that causes them to hang and fail
[17:28] <jam> it may also be one of the "you have to have more than one processor" side effects, etc.
[17:29] <jam> I got the win32 test suite to pass without skipping on multi-processor machines a while ago (could have bit rotted), but vila was convinced we needed it to always pass on single-cpu machines, so put in the extra effort for that
[17:29] <jam> and multi-core machines seem to have fewer bugs wrt threading (at least fewer deadlocks, etc)
[17:30] <vila> jam: meh, we're not supposed to leak threads anymore (except for the paramiko ones and even there..)
[17:30] <jam> vila: good to hear
[17:31] <jam> though still "bzr selftest -s bb" passes cleanly here (or everything has passed in the last 4m which should include breakin tests)
[17:31] <vila> that's the problem with transient failures in tests, they pass in some places...
[17:36] <jam> vila: IME breakin fails reliably for some people, but succeeds reliably for me
[17:36] <jam> FAILED (errors=1, known_failure_count=2)
[17:36] <jam> 135 tests skipped
[17:36] <vila> jam: but whether or not multi-core machines has less bugs, it would be nice to have TestBreakin passing everywhere since it's our guarantee that C-\ works
[17:37] <jam> And a failure while reporting about a missing feature...
[17:37] <vila> jam: file a bug, these are hard to track otherwise
[17:37] <jam> vila: I wouldn't be surprised if it is a testtools issue, I'm still at 0.9.3
[17:38] <jam> I don't really feel like fighting with that
[17:38] <vila> jam: don't fight, file :)
[17:47] <jam> filed bug #673128
[17:54] <sakura13> good evening
[17:55] <sakura13> i have a problem with bazaar how can i setup a webserver for it that my workers can use the same files for our project
[17:56] <GaryvdM> sakura13: hi. What OS? Why does is have to be a webserver, and not some other file share?
[17:57] <Peng> Assuming nothing has gone horribly wrong in the six months I haven't been paying attention: You don't have to do much setup. Just dump the .bzr directory somewhere a web server can serve the files.
[17:57] <sakura13> GaryvdM:  wait a sec my project leaders comes in few secs
[17:57] <Peng> You *can* set up the smart server program, but it's not necessary. It just makes things faster.
[17:58] <sakura13> Peng: moment pls project and server owner comes in few secs
[17:58] <GaryvdM> Peng: non smart http is read only.
[17:58] <Peng> GaryvdM: Oh, err, good point. Peng hasn't slept much.
[17:59] <Peng> Then, add SSH or SFTP for that! :P
[18:03] <GaryvdM> vila: oh boy - I've got some bug logging to do. 5345/28465 tests run, and allready 104 fails
[18:04] <sakura13> Swonline: hi
[18:04] <92AABIV95> Hi, is anyone aware of a command or plugin that will let me list revisions with or by their size? Basically, I'm trying to identify a revision that introduced a really large amount of stuff that was subsequently deleted so I can attempt to get rid of them.
[18:06] <Swonline> How do I set up a server?
[18:07] <vila> GaryvdM: if you're running from bzr.exe you may be the first one :-/
[18:07] <vila> GaryvdM: don't let it come in the way of releasing, I don't think these are *new* failures
[18:07] <sakura13> vila: GaryvdM its my project leader he want setup the server for our project with bazaar
[18:08] <GaryvdM> Swonline, sakura13: A bzr smart server is not needed to host a bazaar branch. What OS is you server?
[18:08] <vila> GaryvdM: but it will be good to fix them nevertheless (probably by skipping)
[18:09] <Swonline> windows
[18:09] <GaryvdM> Swonline, sakura13: The easiest is to create a window file share.
[18:10] <Swonline> how to?
[18:10] <sakura13> GaryvdM: hmm did it worked without server that we can make a networkhardrive
[18:10] <sakura13> GaryvdM: on our pc
[18:10] <GaryvdM> Swonline, sakura13: bzr init \\server\share\proj-name
[18:12] <sakura13> i use atm bazaar explorer
[18:13] <sakura13> ok if i made a branch
[18:13] <sakura13> and upload it with ftp to my server
[18:13] <sakura13> and give the workers the path to the branch
[18:13] <sakura13> will bazaar open it can work with it
[18:13] <sakura13> ?
[18:15] <Peng> FTP is a rather awful protocol.
[18:15] <Peng> Well, plus the implementations tend to be rather awful.
[18:15] <sakura13> yes i know but i dont have ssh acces
[18:15] <GaryvdM> sakura13: Yes - you can upload your branch to the server with bzr push ftp://server/yourbranchname (or a windows file share with bzr push \\server\share\proj-name)
[18:16] <Peng> My condolences.
[18:16] <sakura13> GaryvdM: hmm and what type of branch did we need to share files and so on
[18:16] <sakura13> GaryvdM:  i have used before p4 thats my problem :)
[18:17] <GaryvdM> What type of branch: Any bzr branch
[18:17] <sakura13> GaryvdM:  i have here different work modules colcated, feature, plain and shared branch
[18:19] <GaryvdM> sakura13: You probably want a shared repository
[18:19] <GaryvdM> sakura13: But plain branch is  fine
[18:20] <sakura13> GaryvdM:  and what is the diffrent of a branch and a plain
[18:20] <Peng> Branches in shared repositories share their data. It's more efficient.
[18:21] <Peng> But there's no functional difference. It's just faster and less disk-intensive.
[18:21] <Peng> Well, plus it makes complicated auth a pain since everybody needs to be able to write to the repo...
[18:21] <sakura13> becuase we must have merge files, work on the same files and so n
[18:23] <GaryvdM> sakura13: Can I recommend that you go though http://doc.bazaar.canonical.com/latest/en/mini-tutorial/ - I'm sure things will be a lot clearer afterwards.
[18:23] <sakura13> damn if i open my ftp bzr it crash
[18:23] <sakura13> -.-
[18:24] <Peng> What? If bzr crashes, paste the traceback at http://paste.ubuntu.com/ . Or the URL if you have that helper thingy enabled.
[18:25] <sakura13> i dont use command line on windows
[18:25] <sakura13> i use atm the gui
[18:25] <sakura13> and dont see any debug
[18:26] <sakura13> but i try now with bazzar on cmd i hate windows cmd but ok
[18:27] <sakura13> hmm but one thing i dont understand
[18:27] <sakura13> if i change some files with bazaar
[18:27] <sakura13> to example i put a new file on it
[18:27] <sakura13> ahh its ok
[18:29] <GaryvdM> sakura13: http://doc.bazaar.canonical.com/explorer/en/guide/processes/starting_a_project.html <- explains the different branch types you saw.
[18:31] <GaryvdM> If you select "Feature branches" it creates a shared repo, and a trunk branch for you. (I just learnt that...)
[18:32] <GaryvdM> So that's the recommended.
[18:32] <GaryvdM> You would then create you feature branches in the shares repo.
[18:37] <GaryvdM> Night all.
[18:46] <sakura13> k thx
[19:10] <eridu> I have gpg_signing_command = false in my ~/bazaar/bazaar.conf, and gpg-signing still works. what's going on? I'm using v2.2.1 (in Ubuntu 10.10)
[19:11] <eridu> (I know signing works because I see the "You need a password..." message in my terminal when I run bzr ci)
[19:15] <bob2> branch-specific bazar.conf?
[19:15] <eridu> nope, global
[19:16] <eridu> it was my fault
[19:16] <eridu> I set gpg_signing_commmand by accident
[19:17] <eridu> what's especially ironic is that I tested setting gpg_signing_asdfcommand to see if bzr would give me an "invalid configuration key"-type error, and it didn't, but I still was change-blind
[19:37] <poolie> hi all, hi bob2
[19:37] <bob2> 'morning
[19:57] <Silasle> Is there some guide to register you whit bzr (on launchpad)?
[19:58] <maxb> `bzr help launchpad-login`
[19:59] <Silasle> And then?
[20:00] <maxb> and then what?
[20:00] <Silasle> Dont i need some ssh keys?
[20:01] <maxb> yes
[20:02] <Silasle> How do i create them
[20:04] <Peng> THere's nothing LP-specific about creating SSH keys, although I think they have some guides anyway.
[20:04] <maxb> I have no idea where there might be a pleasant tutorial for this sort of stuff, sorry. I learn it so long ago that it's fairly alien to explain for me
[20:04] <Silasle> Ok, i found the guide
[20:05] <Silasle> launchpad-login and ssh keys is everything i need?
[20:07] <Silasle> Ok, thanks
[20:31] <jam> poolie: hey, are you online already? /wave
[20:32] <poolie> hi jam, i am
[20:33] <mgz> jam, is the bug 673128 about the UnicodeDecodeError? I marked it as a dupe of the testtools bug I fixed to start with, but I'm not sure you didn't mean something else.
[20:34] <jam> mgz: Well, I got the failure during the "XX tests skipped due to ..." section
[20:34] <jam> not while actually running the tests
[20:34] <jam> It may be due to testtools
[20:34] <jam> vila said "don't think, just file"
[20:34] <mgz> I agree with the sentiment, I'm just having trouble understanding the bug. :)
[20:37] <mgz> that traceback certainly doesn't come from the printing-skip-reasons phase. perhaps some kind of stream buffering thing?
[20:37] <jam> mgz: it would be a very strange buffering.
[20:37] <mgz> oh, wait, I know
[20:37] <jam> It definitely happens after the "10 tests skipped by 'foo'" has been reported
[20:37] <mgz> it's just bzrlib being helpful and printing out the error that broke it at the end
[20:37] <jam> but before the missing "unicode" feature is reported (or something like that)
[20:38] <mgz> for whatever reason you didn't get any more skip things listed.
[20:42] <mgz> anyway, upgrade and bug with bother you no more.
[20:42] <mgz> and you'll also be able to write tests that deal with unicode things.
[20:46] <spiv> Silasle: https://help.launchpad.net/YourAccount/CreatingAnSSHKeyPair
[20:47] <Silasle> spiv, Thanks but i'm already done :)
[21:04] <poolie> mkanat, hi?
[21:04] <poolie> hello emmajane!
[21:05] <emmajane> poolie, ola!
[21:40] <lifeless> sladen: actually, hi here.
[21:43] <sladen> lifeless: and here!
[21:44] <lifeless> hi
[21:44] <lifeless> so you seem to be filing a large number of bugs which seem like first-impressions issues, not functional problems.
[21:45] <lifeless> and particularly biased towards folk with existing muscle memory
[21:46] <lifeless> this doesn't seem like a particularly useful way to address whatever use case you're working on : consider what would happen if you filed bugs on dpkg that it doesn't support apts options, or on rpm that it doesn't support dpkg's options.
[21:46] <lifeless> sladen: so I thought I'd try to engage you in a higher level discussion
[21:47] <sladen> lifeless: yup, I've been chating to poolie in the background over this
[21:47] <poolie> sladen, perhaps we should just talk here
[21:48] <poolie> lifeless, yes i was just making the same point
[21:49] <poolie> people who want _exactly_ the git ui, byte for byte and bug for bug, are unlikely to ever be satisfied by bzr
[21:50] <poolie> on the other hand bugs about things that are missing features, or that are inconsistent within bzr on its own right, are worth noting
[21:50] <sladen> the general background is being (forced?) to start using bzr where for I had been using git for the last few
[21:51] <poolie> if i can draw another analogy
[21:51] <poolie> people file a bunch of bugs about ubuntu being different to mac os, or to windows
[21:51] <poolie> there may be a good point behind them but they are not generally very productive bugs ime
[21:52] <sladen> so "bzr commit -a" falls into that category
[21:53] <sladen> others, like bzr diff blocking whilst writing a bzr commit are workflow blockers
[21:53] <sladen> and still others (bzr pager by default) are out-standing issues from before I ever used git
[21:54] <sladen> it's just that (in the latter case), having used something else, it confirms the original issue
[21:55] <lifeless> sladen: so some of this is perspective
[21:55] <lifeless> sladen: for instance, the pager thing really should be fixed in your shell.
[21:55] <lifeless> sladen: its nuts to change every single tool to do its own pagination
[21:56] <sladen> lifeless: and that may ever well be a good point
[21:57] <sladen> lifeless: (but doesn't help get the reported issue fixed)
[21:57] <sladen> s/ever/very/
[22:00] <lifeless> sladen: well, part of it is whether its a good idea or not
[22:01] <lifeless> sladen: the reported issue can be viewed in a few lights - concretely, I /loathe/ gits auto pagination stuff.
[22:01] <lifeless> it invariably gets in my way when I use git.
[22:01] <sladen> lifeless: yup, and you're a poweruser, and can disable it
[22:01] <sladen> lifeless: it's not a use-case I'm interested in
[22:01] <lifeless> sladen: non powerusers can use GUI's
[22:01] <lifeless> sladen: if you want to get into an argument on that angle
[22:01] <lifeless> but I don't think an argument is a good way to move forward
[22:02] <lifeless> we want bzr to be joyful to use for as many people as possible
[22:02] <sladen> lifeless: yup, and there are plenty of choices for dvcs so people vote with their feet
[22:02] <sladen> lifeless: the question is, why did we, with a one year headstart, end up with less user-base than other dvcsen
[22:03] <lifeless> sladen: s/one year/2 months/
[22:03] <poolie> sladen, that's more of what I would call a 'beer question' than a 9am in the morning question
[22:03] <poolie> i realize you're in a different tz
[22:04] <poolie> at the moment i'm more interested in what we can best do this week and in the next few months
[22:04] <poolie> taking into account mistakes we might have made in the past
[22:04] <poolie> one of them was making it too hard to get patches in, and i think that is now a lot better (i'd welcome evidence to the contrary)
[22:05] <sladen> I don't think there are mistakes, were are where we are because of how things unfolded, and wouldn't be otherwise
[22:05] <sladen> just as, on the desktop we are where we are
[22:05] <poolie> another was not being systematic about performance, and that is somewhat better now, though not ideal
[22:06] <sladen> speed and illusions of interactivity are not the same---an iPhone draws pretty zooming windows with the GPU while the application takes 1-2 seconds to starts up on the CPU
[22:07] <poolie> sure
[22:07] <sladen> the conversation (feels) uncomfortably confrontation---which is why I'm being hesitant
[22:08] <poolie> bzr with progress bars on certainly feels different to without
[22:08] <sladen> it's not something I wish for, but it 9am or midnight
[22:08] <poolie> (though again, the progress bars are not perfect in either coverage or implementation)
[22:08] <poolie> no, me either, and i don't want to be defensive
[22:10] <lifeless> one thing I think in particular rubs about bugs that reference the git UI
[22:10] <lifeless> nearly -everyone- I know, including git afficiondos, dislikes the git UI
[22:11] <lifeless> I'd much rather see 'Foo was confusing, I looked on the web, in the help, and finally found an answer in the corner of ...' - thats a symptom we can definitely work on to improve.
[22:11] <lifeless> Adding to that that git does X, svn does Y, hg does Z can be good inspiration to fix it.
[22:12] <lifeless> but the actual /problem/ encountered is squarely and clearly focused on defects in bzr, not on comparisons which are rather more subjective.
[22:20] <sladen> lifeless: I agree, I've tried to use the wording "other dvcs" and not being specific about it
[22:20] <sladen> lifeless: this issues are all (I think) about usability---if usability is accepted as a defect in bzr, then that matches the spirit in which the bugs were reported
[22:21] <lifeless> sladen: usability is very much a defect
[22:21] <lifeless> blah, take that in the spirit intended
[22:22] <poolie> i think the reports are fine
[22:22] <poolie> in spirit
[22:23] <poolie> i agree with robert that saying "there is this problem, other people have done X" is kind of grounding it better
[22:23] <poolie> responses to bugs that suggest workarounds often make me uncomfortable
[22:23] <poolie> it's one thing to say "thanks, that is a bug, in the interim until we fix it you can do Y"
[22:23] <poolie> it's much less good to give the impression that because a workaround is possible we're denying there's a bug
[22:23] <sladen> +1
[22:36] <peitschie> (mornin sladen, poolie, lifeless and all you other lurkers :)  )
[22:37] <sladen> good morning poolie
[22:37] <poolie> hi there
[22:37] <sladen> peitschie even
[22:37] <peitschie> :)
[22:40] <jbowtie> Hello, also welcome back, poolie
[22:44] <peitschie> hi jbowtie
[22:44] <jbowtie> My email inbox is unusually full of bug activity this week.
[22:46] <jbowtie> Oh, right, UDS.
[22:49] <jbowtie> poolie, did you finally hire a BSE? (noticed that ad is gone from IRC topic) Congrats.
[22:49] <poolie> jbowtie, we did
[22:50] <poolie> i'll send a mail
[23:15] <jbowtie> I have to say the new release of bzr-tfs seems to be much better integrated, those per-foreign-vcs tests are very helpful.
[23:23] <poolie> excellent
[23:25] <maxb> hmm. the clearing of progress displays seems to have regressed again
[23:26] <maxb> "Pulling /home/maxb/wc/bzr/udd/trunk from bzr+ssh://bazaar.launchpad.net/~udd/udd/import-scripts/shing stream" says my multi-pull