/srv/irclogs.ubuntu.com/2009/06/11/#bzr.txt

kenichiLarstiQ: hey there, do you remember my weird negative revno issue from yesterday?  interesting fix: edit .bzr/branch/last-revision and fix the revno.  so far, that seems fix a copy of my borked repo, without taking nearly as long as 'bzr reconcile'00:09
igcmorning00:10
garyvdmMorning igc - how did it go on Tuesday?00:17
=== Kissaki is now known as Kissaki^0ff
garyvdmpoolie: Hi - The channel topic is still not right?00:20
igchi garyvdm. OK thanks. Tired though.00:21
lifelessgaryvdm: you can fix it :)00:26
garyvdmOk00:26
=== garyvdm changed the topic of #bzr to: Bazaar version control system | 1.15.1 released 09 June, 2009 | http://bazaar-vcs.org | http://irclogs.ubuntu.com/ | today's focus: https://bugs.edge.launchpad.net/bzr/+milestone/1.16
pooliehello jml00:47
pooliesounds good00:47
pooliejetlag-induced second breakfast has been served and eaten :)00:47
pooliehello gary00:47
pooliespiv, hello, do you know much about bug 385132?00:58
ubottuLaunchpad bug 385132 in bzr "bzr uses inaccessible, internal URL when pushing to server with host-relative default_stack_on url" [Critical,Triaged] https://launchpad.net/bugs/38513200:58
jmlpoolie: :D00:58
jmlpoolie: I could use a first breakfast :)00:58
poolieyou can do that first if you want :)00:58
spivpoolie: not a huge amount01:07
spivpoolie: I don't have any insights to add that aren't already in the bug comments01:08
jmlpoolie: hi01:10
jmlpoolie: let's talk :)01:10
lifelesspoolie: I love it when people comment on many year old bugs of mine ;)01:11
poolielifeless: which one?01:11
lifeless183701:11
lifelessproduct/bugs target to milestones page shows fixed bugs01:12
lifelessat least, I think I filed it01:12
poolieoh, yeah01:13
poolieyou did01:13
pooliejml: is that really 5s lag!01:15
poolieor call me at home01:15
jmlpoolie: sure.01:15
jmligc: does the branch that's in pqm now fix a particular 1.16 bug?01:32
igcjml: I'm making the requested tweaks from jam's review now01:33
jmligc: sweet01:34
igcjml: there's a catch though ...01:34
jmligc: but I'm actually just wondering if "Thu Jun 11 00:42:04 2009 UTC: Ian Clatworthy <ian.clatworthy@canonical.com>, '(igc) fix rule handling so that eol is optional' " is associated with a bug report...01:34
jmligc: oh, what is it?01:34
igcI'm due at the hospital in 25 minutes01:35
jml:(01:35
igcand I'm not going to get the tweaks finished by then01:35
igcso I'll push what I can do by then but not land it yet01:35
jmligc: ok. I'll see what we can do to get it landed.01:36
jml(no promises though, sorry).01:36
jmligc: for the record, is this for bug 362030 or bug 379370?01:37
ubottuLaunchpad bug 362030 in bzr "files whose content changed after EOL filtering erroneously marked as modified" [High,In progress] https://launchpad.net/bugs/36203001:37
ubottuLaunchpad bug 379370 in bzr "bzr: ERROR: Unknown eol value 'None'" [Medium,Fix committed] https://launchpad.net/bugs/37937001:37
igcjml: it still needs a test jam has requested added.01:37
igc36203001:37
jmlok.01:37
igcthe other one has been sent to pqm01:37
jmligc: thanks.01:39
igcjml, poolie: see https://code.edge.launchpad.net/~ian-clatworthy/bzr/eol-st-ci-fix/+merge/726901:40
igcthe first tweak has been pushed to https://code.edge.launchpad.net/~ian-clatworthy/bzr/eol-st-ci-fix/01:40
igcjml: the second tweak is an additional test which jam has basically written in his review01:41
igcjml: but it needs some work because the "set rules" bit isn't quite right01:41
igcjml: poolie might be able to assist you in tweaking it01:42
jmligc: ahh cool.01:42
igcjml: we *could* land the branch without the additional test (for now) but please don't do that unless jam or poolie agrees01:43
jmligc: of course :)01:43
igcjml: bzrlib/tests/workingtree_implementations/test_eol_conversion.py shows how to patch in custom rules01:44
igcyou can't use absolute paths but that doesn't matter01:45
igcthe important point is that the rules change at all01:45
igcjml: ok, got to run - back in a few hours01:45
jmligc: np. thanks for the help :)01:45
lifelesspoolie: bubg 24117001:46
lifelesspoolie: bug 241170. Why did you remove the milestone for when it was fixed?01:46
ubottuLaunchpad bug 241170 in bzr "Assertion error when trying to push" [High,Confirmed] https://launchpad.net/bugs/24117001:46
poolielifeless: the bug doesn't say it's fixed01:50
poolieit says it's still open01:50
lifelessoh. I'm pretty sure its lying01:50
lifelessbecause we adjusted get_parent_map in the manner proposed01:50
poolieit's not in NEWS eithr01:51
jmldid the Russian doc patch land?01:51
lifelesstrue, but if you look at remote.py01:51
pooliejml: i sent to pqm for the zdll fix and the smaller images01:52
jmlpoolie: thank you.01:52
jmlyes, the russian doc patch landed :)01:53
lifelesspoolie: lets ask spiv01:54
lifelessspiv: ping01:54
spivlifeless: pong01:57
lifelesssee scrollback01:57
spivI vaguely recall seeing that bug get fixed, but I'm not 100% sure.01:58
spivIt only affects/ed branch format 5.01:59
pooliei'll mark it fixed then02:00
pooliespiv, what was the other bug you were working on on tuesday? is that now landed?02:00
spivNo, although I'm about to submit a cheap fix for that.  I have a followup that removes all VFS RPCs from the "pull -r 123" case (done in an empty repo, where 123 is a revision found only in the stacked-on repo of a stacked branch).02:01
spiv(18 RPCs total, down from >40)02:01
spivpoolie: ah, I believe 241170 may be a dupe of 214894, which is fixed.02:05
spivOh, although the 241170 reporter said they were using 1.5, which apparently had the 214894 fix...02:06
jmlI just sent a status update email for the 1.16 releaese.02:06
jmlwould greatly appreciate it if you had a look (particularly spiv & lifeless)02:06
spivjml: is it possible to submit a lp merge directive for a specific revision (i.e. not tip) of a branch?02:22
jmlspiv: not through the UI.02:23
spivjml: if I do it through email will it DTRT?02:23
jmlspiv: it's possible that emailing a merge directive might do it.02:23
spivOk, experiment time!02:23
jmlspiv: I couldn't say for sure though.02:23
lifelessspiv: I think that lp is a subset of merge directives.02:23
lifelessspiv: I'm not sure where the subset lines are drawn though02:24
jmllifeless: I've got a few questions about this final_stack_pwd.02:29
jmllifeless: where should I add the tests?02:30
jmllifeless: TestSmartServerRequestBzrDirInitializeEx seems to suggest that test_bzrdir is the right place.02:31
jmlbut in test_bzrdir, I can't see any sign of using the smart server (particularly in TestRepositoryAcquisitionPolicy)02:32
lifelessare you going to be writing implementation or interface02:34
lifelesstests02:35
lifeless?02:35
jmllifeless: I don't know.02:35
lifelessso, I don't know where the tests should go then02:35
jmllifeless: I was under the impression that there were already tests for this area of functionality02:36
jmllifeless: where do they live?02:36
lifelessif you think you've found a misbehaving implementation where tests for the behaviour are either broadly irrelevant or hard to exercise on other implementations, then I'd write implementation tests02:36
lifelessif you can write the tests as generic exercise-of-interface then interface tests02:36
spivjml: hah, the answer is both yes and no: https://code.edge.launchpad.net/~spiv/bzr/stacking-friendly-revision-history-verb/+merge/731402:37
spivjml: the diff in my email is included and shows the intended content (only 4 revs worth)02:37
spivjml: but there's a generated diff from lp linked on that page with too much history, and lp is listing all the revs in that branch as unmerged in this proposal.02:37
jmlspiv: good to know :)02:38
lifelessbzrlib/tests/bzrdir_implementations/test_bzrdir.py +116602:38
lifelessjml: the interface is 'initialize_on_transport_ex'02:39
jmllifeless: thanks. I gather those tests run against every implementation?02:40
lifelessyes02:40
lifelessthey run against a current and an older protocol version of the smart server too02:40
jmlcool.02:41
jmlso if the bug is actually in SmartServerRequestBzrDirInitializeEx then a properly written test ought to trigger it?02:41
lifelesswell02:41
lifelessI think we should pin down the contract02:42
lifelessso yes02:42
lifelessunit tests for smart server verbs are in02:42
lifelessbzrlib/tests/test_remote.py +74402:42
jmltest_smart02:42
jmloh, test_remote02:42
lifelessthere's history here. Don't look under the curtain02:43
jmlok.02:43
spivWell, test_smart.py is unit tests for the server-side.  test_remote.py is unit tests for how the client-side Remote* classes interact with the network.02:43
jmllooking at BzrDir.initialize_on_transport_ex...02:44
jmlthe docstring says that stacked_on=None implies following the target's stacking policy.02:45
jmlI guess that means that stack_on_pwd is ignored in that case.02:45
lifelessspiv: and server side unit tests.02:45
lifelessspiv: they are there, existence proof :P02:45
lifelessjml: look at bzrlib/bzrdir.py +315002:46
lifelessjml: did that help?02:54
jmllifeless: probably :)02:55
jmllifeless: still paging all of this stuff in.02:55
spivlifeless: hmm, some tests there do create actual servers rather than using test doubles, but AFAICT the primary purpose of all the tests in test_remote is to test client-side behaviour.02:56
lifelessspiv: you're likely right02:57
eydaimoni've done a bzr revert to an earlier revision... how can I do a diff against the latest revision if I didn't know what the rev number is? I was thinking something like bzr diff -r head file03:00
pooliediff runs against the branch's tip03:01
poolieso just 'bzr diff' will show the difference03:02
eydaimonso bzr diff -r tip head?03:02
eydaimonoh ok03:02
eydaimonout of curiousity, is there a way to specify tip?03:03
poolie-r head: i think03:03
spivOr -r -103:03
pooliejml, i think i'ts ok to slep bug 284038 though i also think it's a mysql bug and deserves a little extra effort03:03
ubottuLaunchpad bug 284038 in bzr "push should warn about uncommitted changes" [Medium,Fix committed] https://launchpad.net/bugs/28403803:03
poolieoh of course03:03
eydaimon-1 works03:03
pooliejml, bug 380314 i think is critical to use with launchpad?03:04
ubottuLaunchpad bug 380314 in bzr "Branch.revision_history RPC fails on a stacked branch" [High,In progress] https://launchpad.net/bugs/38031403:04
spivIt breaks "bzr pull -r 123"03:04
pooliei may try bug 38545303:04
ubottuLaunchpad bug 385453 in bzr "make dist should fail if C files don't exist or can't be built" [Medium,Triaged] https://launchpad.net/bugs/38545303:04
spiv(on stacked branches)03:04
spivWhich is pretty serious, I think.  The fix is up for review now.03:04
jmlpoolie: re 284038, it'd definitely be great to get it in, yes.03:06
jmlpoolie: likewise for the 'make dist' bug, but imo format bump & reviewing existing patches (like spiv's) take precedence.03:07
spivjml: fwiw I don't have anything to add to your status update email (other than pointing out that my fix is now up for review, which you already know)03:08
pooliejml, spiv, how about bug 381329 - is that in trunk? in 1.15 (where it's targetted)03:08
ubottuLaunchpad bug 381329 in bzr "Some expected smart server errors cause client tracebacks" [Critical,Fix committed] https://launchpad.net/bugs/38132903:08
spivpoolie: yes, the 1.15 patch was a direct cherrypick off trunk.03:08
poolieso i'll mark it released in 1.16 too?03:13
pooliedone03:15
igcjml: back. I'll keep working on that tweak assuming no-one has picked it up03:47
jmligc: cool :) as far as I know, no one has.03:47
pooliejml: format added, am going to do some adhoc testing03:48
poolieincluding upgrading to it from dev703:48
pooliewill put it up soon03:48
jmlpoolie: thanks.03:48
lifelesspoolie http://cworth.org/intel/driver_stability/03:49
igcpoolie: be sure to check upgrading from dev6 too03:49
poolieyeah, and on a small one upgrading from something earlier03:49
lifelessjml: so tell me when you're  looking at stacking03:49
igcpoolie: there were some problems with that a few days ago03:49
igcwhich I think jelmer or jam might have fixed since03:50
pooliei might get some lunch first though03:51
jmllifeless: now :) http://paste.ubuntu.com/193156/ -- that's the test I just finished.03:51
lifelesspoolie: I've lost the bazaar-announce moderator password again.03:53
jmllifeless: any comments on that validity of that test?03:55
lifelessbad:  except errors.BzrError:03:56
lifelessjml: asking for 1.9 is problematic03:57
lifelessjml: I suggest being a little bit more complex:03:58
lifeless - make a control dir without specifying format, which will use the parameterised format03:58
lifeless - if that is a bzrdirmeta1 then hard code 1.903:58
lifeless - otherwise use whatever branch format that the test makes by default03:58
lifelessit looks broadly fine though03:59
jmllifeless: you mean a control dir at the 'stack-on' location?04:00
lifelessjml: or even a third location04:01
jmlok.04:01
jmlI'm not sure how to avoid the BzrError.04:02
lifelessits too generic04:02
lifelessits like catching Exception04:03
jmlthree of the formats simply raise a BzrError when calling set_default_stack_on.04:03
lifelessmmm04:03
jmlBzrError: Cannot set configuration in <bzrlib.bzrdir.BzrDir5 object at 0x6bb7350>04:03
lifelessoh04:03
jmlI guess I can check the str() of the exception.04:03
lifelessthen its fine04:03
lifelessbut please make it more specific04:04
lifelessdon't chain the calls so deep04:04
lifelessconfig = self.make_bzrdir('.').get_config()04:04
lifelesstry:04:04
lifeless    ...04:04
jml*nod*04:04
lifelessexcept BzrError:04:04
* spiv -> food04:21
GPHemsleyUgh... who do I have to poke to get an up-to-date Mac OS X package?04:26
GPHemsleyPreferably, a package that works04:26
GPHemsleyBecause my lack of a working subvertpy seems to be causing trouble04:27
GPHemsleyNot to mention that rebase is apparently out of date, too04:28
GPHemsleyLooking for 1.13.0 when 1.15.0 is required04:29
GPHemsleyWell, downgrading to 1.14.1 at least gets rid of the rebase error04:43
poolielifeless: you can call me about format naming if you want04:46
poolieor we can talk here04:46
lifelesspoolie: either is fine04:47
poolieif we're miscommunicating04:47
lifelesspoolie: I'm filing the bugs we talked about at the moment, just takes a while to make sure htey aren't dups04:47
pooliethere's no super rush04:47
pooliei am going to put this patch up in a bit, taking teh approach i outlined04:47
lifelessI think if we're making it supported in 1.16, we should put 1.16 in the version string04:48
lifelessmy key point is that I *know* have a code change which requires a string bump but no disk changes04:48
lifelesss/no disk/no other disk/04:48
pooliewhat is that change?04:49
lifelessfixing the rich root additions the streaming fetch code sends04:49
lifelessIts the one I mentioned in the thread04:49
pooliejml, both the smaller images and the zdll fix landed04:53
mwhudson.oO(will brisbane-core make a difference to how much i exceed my download cap each month...)04:54
lifelesspoolie: house phone is out of juice04:54
jmlpoolie: thanks.04:54
jmllifeless: fwiw, I'm now stumped on the actual fix to the stacking bug. I'm off to get lunch, should be back in ~20mins.04:55
jmlpoolie: I just have you down for the format name thing -- is that your understanding too?04:55
poolieyes04:56
jmlcool.04:56
lifelesspoolie: I'm assuming that was you ringing04:57
poolielifeless: so to me it doesn't make any sense to say it's the same format on disk but it needs a new number04:57
poolieit was04:57
lifelesspoolie: other suggestions to stop broken code writing to this format appreciated04:57
poolieis there a bug for this?04:58
lifeless13:41 < lifeless> poolie: I'm filing the bugs we talked about at the moment, just takes a while to make sure htey aren't dups04:58
lifeless(no not yet)04:58
poolieby which i mean, what broken code are you talking about?04:58
lifelessthere was but I think it got closed when 1/2 the code got fixed04:58
lifelessbecause the test suite will fail when we fix the dependent issues04:59
poolieif using 1.16 to fetch into this format will lose data then that's a blocker for doing this release at all04:59
poolieif this release is ok, but you want to change the way things are represented in future, then i'd say that really is a change to the ondisk format05:00
lifelessIt won't, because IDS will take over. It will just take hours to pull over the network.05:00
lifelesspoolie: its neither of those cases.05:00
igcjml: those tweaks are now done. Just running the test suite and I'll submit it to pqm05:02
jmligc: thanks.05:13
GPHemsleyHow do you change repo formats?05:13
GPHemsleyI forget the command05:13
jmlbzr upgrade05:14
GPHemsleyhmm... I should've thought of that05:14
GPHemsleybut thanks :)05:15
lifelesshttps://bugs.edge.launchpad.net/bzr/+bug/37473505:48
ubottuLaunchpad bug 374735 in bzr "Plan and UI for upgrading multiple stacked branches " [Critical,Triaged]05:48
jml:(06:12
pooliejml, lifeless : https://code.edge.launchpad.net/~mbp/bzr/385103-format-name/+merge/732106:16
pooliecomments on NEWS especially welcome06:16
* jml waits for the diff to generate06:17
lifelesspoolie: 2a - not 2s-rich-root?06:17
lifelesss/s/a06:17
lifelesspoolie: not every format is listed by bzr init --help IIRC06:18
lifelesspoolie: its controlled by data on the formats06:18
lifelesspoolie: IMBW06:18
poolies/every/lots of/ :)06:19
bob2is there some way to get a full diff of a merge req on lp?06:19
poolieand i think we need to get away from having -rich-root  in the name06:19
pooliewe need to do it sometime06:19
poolienow's a good time06:19
pooliebob2 there's a link to download it06:19
lifelesspoolie: I think we should for 2.0 release06:19
lifelesspoolie: I think you should try working with a bzr tree in 2.0 and see what users get told about the incompatibilities before deciding that we're ready to make it less scary06:22
pooliei'm upgrading now06:23
lifelessin particular interacting with the rest of the community06:24
pooliei'm happy to tweak that text to make it more constructively scarey06:24
lifelessI don't think people will read the text06:24
poolieor the short format description06:24
lifelessindeed06:25
lifelessoften people will just see 'run bzr upgrade --FOO'06:25
lifelessbased on previous observation.06:25
lifelessIts safer to say 'this is safe to do with [x] conditions' when the string is scary, than is is to say 'this is scary unless [x] conditions' when the string looks safe.06:26
pooliejml, thumper, do either of you want to talk?06:26
poolielifeless: possibly we should say interactively "do you really want an experimental format"?06:27
pooliewe could print a warning but if you just did an upgrade it may be too late06:27
jmlpoolie: yes. :)06:27
pooliei'm open to changing the ui name to say 2a-beta06:27
bialixgaryvdm: still here?06:29
garyvdmbialix: Yhea - but not for long - I've been up all night :-)06:30
lifelesspoolie: I don't think that is enough. Also showing a interactive question would seem to me to be further away from what you seem to want than 2.0-rich-root would be.06:30
lifelesspoolie: I'm not sayin the disk format needs to mention rich roots.06:30
bialixgaryvdm: almost the same06:30
bialixGary, I'm planning to make release today06:31
bialixit's ok to using current trunk?06:31
lifelesspoolie: fwiw bug 368921 was the IDS version of the bug with the network server06:31
ubottuLaunchpad bug 368921 in bzr "rich root upgrade adds inconsistent parents/rich root parents are not evaluated correctly." [Critical,Triaged] https://launchpad.net/bugs/36892106:31
garyvdmbialix: Cool - I did see you mail. Yhea - there are no bugs/regressions I'm am aware of.06:32
garyvdmbialix: I've got some stuff I'm holding off till you release.06:32
bialixgayvdm: I have one question about next one release. It seems we will release next release in sync with bzr 2.006:32
bialixgaryvdm: I can use separate release branch, so you don't need to hold off06:33
pooliehello gary, bialix06:33
bialixI need only finish new Inno Setup based installer06:33
bialixpoolie: good day06:33
garyvdmbialix: Don't worry - It probably not even ready for trunk yet.06:33
garyvdmHi poolie06:34
bialixgaryvdm, poolie, luks, igc: I'm thinking about calling next qbzr release as 1.006:34
bialixthis is very interesting opportunity06:34
igcbialix: as in the one after today's?06:35
bialixigc: the one in next month06:35
igcbialix: right, so qbzr 1.0 ships with bzr 2.006:35
bialixyes, exactly06:35
bialixrecently jelmer asked about releasing bzr-svn as 1.006:35
igcbialix: that would be good I think06:36
garyvdmbialix: I think we should wait untill qmain is finished.06:36
bialixgaryvdm: do you see bzr-explorer?06:36
igcbialix: Javier will hopefully be adding some more q* commands in coming weeks, e.g. qsend, qexport06:36
garyvdmbialix: yes - and I have to chatting alot to igc about it.06:37
bialixGary, I know we have a lot of unfinished stuff06:37
igcbialix: and I have plans for putting some together over nights/weekends too, e.g. qbind06:37
garyvdmbialix: we should probably test if bug 385550 also affects the bzr installer.06:37
ubottuLaunchpad bug 385550 in qbzr "Unexpected error when uninstalling QBZR " [Critical,Confirmed] https://launchpad.net/bugs/38555006:37
bialixgaryvdm: no06:38
bialixbzr.exe installer uses Inno Setup.06:38
bialixI wrote it06:38
bialixqbzr installer wrote luks06:38
igcbialix: any chance qversion will make the cut this time around?06:38
bialixI never used NSIS myself06:38
garyvdmbialix: Ok then it really makes sense to use Inno.06:38
bialixigc: there is some layout problems with your dialog06:38
bialixI'll plan to fix it after 0.1106:39
bialixmaybe tomorrow06:39
igcbialix: ok06:39
igcbialix: also, thanks for the qview to qviewer change06:40
igcbialix: do we need to tweak TortoiseBzr accordingly?06:40
bialixigc: what's your vision of annotate in explorer? this command require filename as argument06:40
bialixigc: I've checked tbzr trunk: qviewer is not used there06:40
igcbialix: I'll probably drop it from the menu in the short term ...06:41
bialixit seems last time Mark Hammond touches the tbzr this winter. and then disaooear06:41
igcand expect people to reach it via qbrowse say06:41
bialixpoolie: does Mark was contractor?06:41
pygibialix: yes, he was a contractor06:41
garyvdmigc, bialix: or qlog06:41
pygibut he's god a full time job now afaik, so he isn't as available06:41
igcgaryvdm: right06:42
garyvdmigc: My 2c about qversion: it would be nice if bzrlib could give us rst which we could convert to html and display in a QTextDocument06:42
bialixgaryvdm: browse see,s more appropriate06:42
bialixseems06:42
garyvdmigc: and bzr cli should also use the rst converted into plain text?06:42
bialixgaryvdm: no06:42
pygioh noes, garyvdm with his qlog :P06:42
bialixno need for rst right now06:43
garyvdmHi pygi06:43
garyvdmbialix, igc: Ok but I would like to see qbzr more common with bzr cli.06:43
bialixso, without Mark it seems like tbzr will stalled06:43
garyvdmjust my 2c..06:44
bialixgaryvdm: https://bugs.launchpad.net/qbzr/+bug/27384706:44
ubottuLaunchpad bug 273847 in qbzr "[todo] simple and lightweight rst2html engine needed for qhelp" [Low,Confirmed]06:44
bialixhave not ime for this06:44
garyvdmbialix: I did start doing some work on tbzr - but you convinced me to refocus on qmain :-)06:45
bialixgaryvdm: there is a lot of hard work for tbzr, especially because it affects entire windows performance06:45
garyvdmYhea06:45
bialixwe can implement qmain much faster06:46
garyvdmAnd we can do much more interesting things with qmain...06:46
bialixbut I see bzr-explorer as qmain 0.5 ;-)06:46
bialixexactly06:46
bialixI'd like to implement projects support06:46
igcbialix, garyvdm: my main desire for tbzr is to tweak the menu to reflect the one we're using in Explorer & qbzr-eclipse06:47
bialixok, so today it's too early think about 1.0 date06:47
bialixnp06:47
bialixpygi: what's wrong with either gary or qlog? ;-)06:48
garyvdmI think so06:48
pooliei think it would be cool to do a 1.0 at the same time06:48
pooliei'm using it a bit; i haven't dived straight in as much as ian has06:48
pooliei probably should06:48
fullermdHm.  I like how selftest has a --parallel option that takes various values, and doesn't give any hint as to what the possible values are.06:49
garyvdmGary will show anybody that give him 2sec qlog's features :-)06:49
pooliespeaking of which, hopefully he'll be around moer in future06:49
igcpoolie: have you tried explorer yet?06:49
poolieno, good idea though06:49
bialixigc, btw we have nice working tree widget that could be used in bzr-explorer to show status06:50
lifelesspoolie: so actual review comment added06:50
igcpoolie: it's making progress most weekends06:50
lifelesspoolie: but my prior comments still stand06:50
bialixgaryvdm: maybe if explorer will be pretty usable we can talk about calling qbzr+explorer as qbzr 1.006:50
bialixgaryvdm: one more question, if you can06:51
* igc returns to reviewing jam's iter-nodes patch before he falls aslepp06:52
* igc afk06:52
garyvdmbialix ?06:53
pygibialix: nothing, I like joking with him xD06:53
pooliejml, filex bug 38181406:54
ubottuLaunchpad bug 381814 in netbeans "Netbeans 6.7 in Karmic" [Undecided,New] https://launchpad.net/bugs/38181406:54
bialixgaryvdm: can we chat one day when we both will have enough free time and not be up all the night about alog internals? maybe on some weekend? I have troubles to understand all details, but I'd like to. I have some ideas about adding unit tests to qlog internal engine06:54
poolieor not06:54
pooliebug 38581406:54
ubottuLaunchpad bug 385814 in launchpad-code "it would be nice to show targeted bugs in code review listing" [Undecided,New] https://launchpad.net/bugs/38581406:54
poolieonly wishlist i think06:54
jmlpoolie: yeah, except for reasons I don't understand, we don't use wishlist, just low.06:55
bialixgaryvdm: I'm mostly asking beforehand because it's hard for me to catch you in IRC06:55
pooliei didn't file it06:55
pooliei mean i didn't rank it, i'll let that team do it06:55
jmloh thanks :)06:55
garyvdmbialix: Ok - I started with tests for qlog at uds - there is just so much to test.06:55
bialixI need to understand your design first06:56
garyvdmI'll keep chipping away at it.06:56
bialixthere is too little comments06:56
garyvdm:-(06:56
AfCAm I correct in understanding that there's no release of bzr-gtk that works with bzr 1.15.1?06:56
AfC(if so, pity)06:56
bialixgaryvdm: mostly it's my problem, in my head06:57
bialixso I need some initial guidance06:57
bialixe.g. what is msri means, etc.06:57
garyvdmbialix: ok - questions at the moment?06:57
bialixperhaps it's very obvious for you06:57
bialixno06:57
garyvdmOk - there is a comment for that - I think06:57
* garyvdm checks06:58
bialixone more question:06:58
bialixyou recently changed signals for finsihed/failed/errors06:58
bialixbut all dialogs now require uodate to use new signals06:58
bialixupdate06:58
bialixIMO there is need for another refactoring06:59
bialixGAry, we need to chat more often06:59
bialixok, this was last question06:59
garyvdmHmm - can't find a comment about msri.07:00
bialixme too07:00
lifelesshttps://bugs.edge.launchpad.net/launchpad-code/+bug/38581507:00
ubottuLaunchpad bug 385815 in launchpad-code "need a way to upgrade a project past the rich root border" [Undecided,New]07:00
lifelesspoolie: ^07:00
garyvdmOk it stands for merge_sorted_revisions_index07:00
* bialix notes07:00
garyvdmThere are 2 types of indexes - before its filtered - msri and after - just called index07:01
bialixso msri represent a graph?07:02
pooliejml, igc points out that i also need to specially test upgrading dev6 to dev707:02
pooliejelmer may have a patch for this07:02
garyvdman index for a revision in merge_sorted_revisions07:02
AfCI guess we'll have to downgrade to bzr 1.14.1. Again.07:02
bialixAfC: qbzr has compatible releases07:03
=== mneptok is now known as mnepton
AfCbialix: that is profoundly fascinating.07:04
poolieigc could you read just the news in https://code.edge.launchpad.net/~mbp/bzr/385103-format-name/+merge/732107:04
=== mnepton is now known as mneptok
bialix:-P07:04
poolieoeuthaouth07:04
hchen59Hi, one question: I downloaded .tar.gz file from one project in http://bzr.linuxfoundation.org/unofficial/, extracted to local disk, I found no files except .bzr directory. I tried several bzr commands, while failed to extract version from local bzr repository( some big files in .bzr/repository/indices, I think they are local repository). So which command I should use to extract files? Thanks07:13
bialixbzr co07:13
hchen59[root@hchen59 ispras-moblin-misc-test]# bzr co07:15
hchen59bzr: ERROR: File exists: u'/home/git/ispras-lsb/ispras-moblin-misc-test/.bzr': [Errno 17] File exists: '/home/git/ispras-lsb/ispras-moblin-misc-test/.bzr'07:15
bialixwhat tell you `bzr info`?07:16
hchen59[root@hchen59 ispras-moblin-misc-test]# bzr info07:17
hchen59Standalone tree (format: pack-0.92)07:17
hchen59Location: branch root: .07:17
lifelesshchen59: run 'bzr checkout .'07:17
hchen59Related branches: parent branch: /home/git/ispras-lsb/misc-test07:17
jmlspiv: btw, what's the status of that RPC bug?07:18
hchen59[root@hchen59 ispras-moblin-misc-test]# bzr checkout .07:18
hchen59bzr: ERROR: File exists: u'/home/git/tmp/ispras-moblin-misc-test/.bzr': [Errno 17] File exists: '/home/git/tmp/ispras-moblin-misc-test/.bzr'07:18
bialixhchen59: what `bzr status` say?07:18
spivjml: it's a small patch (100 line diff), still waiting for review.07:19
jmlspiv: ta.07:20
hchen59bzr status => list a lot of files: " removed: ....."07:20
jmlspiv: poolie offered to review it on our recent phone call.07:21
jmlhchen59: 'bzr revert'07:21
poolieigc, wow, that's pretty cool!!07:21
poolieigc: what does 'settings/ignores' do?07:21
pooliei thought it'd change the .bzrignore file but apparently not?07:22
igcpoolie: edits the per-user ignore file07:22
lifelesstotally freezing cold07:22
igcpoolie: editing of branch-specific config files is coming this weekend07:22
poolie:)07:22
jmllifeless: I've got further, but I'm now blocked again.07:22
igcpoolie: along with a nicer status display07:22
lifelessjml: talk to me07:22
hchen59great... bzr revert works, files extracted. While it would revert to version before ? how to check current bzr version?07:23
poolieit'd be nice to be able to see unmodified files too07:23
poolieas a way to kick off editing them07:23
igcpoolie: right07:23
jmllifeless: fetch lp:~jml/bzr/default-stacking-bug-385132 and take a look at the diff.07:23
lifelesshchen59: it looks like whoever made the tarball was a little unused to bzr07:23
jml(or give me a moment to do paste shenanigans)07:24
lifelessI have it07:24
jmlhttp://paste.ubuntu.com/193242/ -- the failure I now get07:24
jmlI've made one of the two blackbox tests pass.07:24
igcpoolie: try https://code.edge.launchpad.net/~jameinel/bzr/1.16-chkmap-updates for a preview of the status display coming07:24
igcbialix: ^^^07:24
bialixigc ?07:25
igcpoolie: sorry, wrong paste ...07:25
hchen59ok, i see, for current bzr revision, I just bzr log, and check top one, right?07:25
lifeless             final_stack_pwd = response[9] or None07:25
lifeless+            if final_stack_pwd:07:25
lifeless+                final_stack_pwd = transport.abspath(final_stack_pwd)07:25
lifeless             remote_repo = remote.RemoteRepository(repo_bzr, repo_format)07:25
igcbialix, poolie: lp:~ian-clatworthy/bzr-explorer/better-wt-view07:25
jmlyeah, that won't work if final_stack_pwd is an absolute url, IIUC.07:25
lifelessabspath('') != '', I think07:25
bialixok!07:25
jmllifeless: that's why it's guarded?07:26
igcbialix: ui not hooked up to a model yet07:26
fullermdWhat's the difference between a selftest FAIL and ERROR?07:26
jmlfullermd: FAIL = failed assertion07:26
jmlfullermd: ERROR = unexpected error07:26
lifelessjml: paging in07:26
jmllifeless: ta07:27
lifelessjml: I think we serialise None as '', so thats why07:27
fullermdSo either one is bad.  Lovely.07:27
hchen59Thanks guys07:27
jmllifeless: bool('') is False, so the line above will guarantee it's None, and the if will make sure abspath is called only if it's a non-empty string.07:27
lifelessjml: so07:29
lifelessIs 'extra' well, extra?07:29
jmllifeless: I'm not sure what extra is there for. It's part of the smart server test fixture, afaict.07:30
lifelessright07:30
lifelessso the path shouldn't be showing07:30
igcpoolie: so the most important part of explorer so far is the "Bazaar" menu - that's what's going in qbzr-eclipse and (hopefully) TortoiseBzr soon07:30
jmlit's in the test because the external URL is bzr://localhost:NNNN/extra/stack-on07:30
igcpoolie: and one other thing, if you have bzr-gtk installed, try ...07:31
igcbzr explorer --gtk07:31
lifelessjml: no07:31
igcthat runs the gtk applets instead of the qbzr ones07:31
lifeless+        self.make_branch('stack-on', format='1.9')07:31
spivThe smart server test support intentionally sets up a server with 'extra' as part of the path the flush out some path translation issues.07:31
lifelessThe set_default_stack_on is setting a bad path07:31
lifelessit should set '/stack-on'07:32
lifelessjml: or listen to spiv :)07:32
lifelessjml: specifically, "chroot-93074512:///extra/stack-on/"  is not valid in the chroot07:32
jmlhttp://paste.ubuntu.com/193245/07:33
jmlthere's the error without it.07:33
lifelessjml: thats better; thats a client side issue07:33
lifelessjml: so lp sets abs paths?07:34
spivjml: you may find -Dhpss -Eallow_debug makes the test output more helpful (or merely more verbose)07:34
lifelessspiv: IIRC server/extra == '/' in the chroot ?07:34
spivlifeless: right07:35
lifelessjml: so the http://paste.ubuntu.com/193245/ case is more correct07:35
jmlhttp://paste.ubuntu.com/193248/ -- more debug output07:35
lifelessjml: it means that path translation is not being done correctly.07:36
jmllifeless: ok.07:37
lifelessjml: specifically, '/foo/bar/' in the chroot needs to be turned into '../foo/bar' or something similar07:37
bialixigc: nice07:37
lifelessserver side07:37
lifelessbecause / is not really an absolute url07:37
bialixigc: new wt view is nice07:37
igcbialix: thanks07:37
lifelessjml: there is code to do this, let me dig a sec07:37
jmllifeless: at the moment I'm doing:07:38
jml            final_stack_pwd = urlutils.relative_url(07:38
jml                target_transport.base, repository_policy._stack_on_pwd)07:38
spivSo the server says the final_stack_pwd in the result is /stack-on, where the right answer would be /extra/stack-on (or perhaps a path relative to the path of the request).07:38
lifelessspiv: extra isn't known by the server07:38
lifelessspiv: has to be relative07:38
spivSure.07:38
lifelessspiv: its an aspect /of/ the server. Think http mapped servers07:39
spivBtw, SmartServerRequest.translate_client_path has one direction of this.07:39
lifelessrepo_relpath07:40
jmlspiv: how do I dump stuff into the debug log?07:40
lifelessbzrlib/smart/bzrdir.py +8407:40
lifelessjml: debug(...)07:41
lifelessor mutter I think07:41
spivRequests actually know the /extra, IIRC, it's stashed in self._root_client_path or somesuch.  But as lifeless is saying you can only adjust for paths inside requests, if the request itself was addressed to one HTTP URL by the client then the HTTP server did magic before handing it off to bzr we don't necessarily know about that...07:41
spiv(Although I believe HTTP servers can pass that info along to the bzr-wsgi glue, perhaps not heavily tested though)07:42
spivjml: I generally use (bzrlib.trace.) mutter07:42
fullermdLast time I looked at setting up bzr+http, it occured to me that I was gonna need a crappile more path translation smarts than I had the slightest idea how to put in place   :|07:42
lifelessfullermd: it is documented07:42
fullermdOnly AIR in a way that makes sense if all requests are rooted at the same place.07:43
lifelessspiv: wsgi can do it but its fragile and best to avoid IMO07:43
jml2.896  target_transport.base: chroot-96675856:///to/07:43
jml2.896  repository_policy._stack_on_pwd:chroot-96675856:///07:43
jml2.897     result:   ('.', 'no', 'no', 'yes', 'Bazaar RepositoryFormatKnitPack6 (bzr 1.9)\n', 'Bazaar-NG meta directory, format 1\n', 'Bazaar-NG meta directory, format 1\n', 'True', '/stack-on', '..', '')07:43
spivlifeless: I meant specifically the bzrlib's wsgi app, rather than wsgi intrisically.07:43
lifelessspiv: so did I07:43
lifelessspiv: :P07:43
fullermdBut when /x and /y aren't peers...07:44
* fullermd watches selftest wind down.07:44
lifelessfullermd: there is a limit to magic, obviously.07:44
=== Kissaki^0ff is now known as Kissaki
jmllifeless: so you can see that's returning a relative stack_on_pwd07:44
fullermdYah.  But withotu breaking that limit, it won't do what I want, so it doesn't run.07:44
lifelessjml: no07:44
lifelessjml: the leading / is the problem07:44
jmlbut the stack_on.startswith('/)07:45
lifelessjml: you need the code from _repo_relpath07:45
lifelessjml: oh yes, indexed variables suck yada yada07:45
fullermdWell, that was "fun"...07:45
fullermdFAILED (failures=12, errors=84, known_failure_count=12)07:45
fullermd2039 tests skipped07:45
lifelessjml: anyhow; the way those two variables interact isn't something I totally get. I didn't write it ;)07:45
jmllifeless: me neither. ISTR that 'stack_on' is what's actually in the config file & 'stack_on_pwd' is how it should be interpreted.07:46
lifelessjml: I think so07:46
lifelessanyway07:46
lifelesswith a smart server a non-abs url isn't usable07:47
lifelessand a root relpath fragment isn't usable either07:47
lifelessso I think07:47
lifelessif stack_on and stack_on[0] == '/': # fix up07:47
lifeless    segments = ['..' * len(stack_on.split('/'))07:48
lifeless    stack_on = '/'.join(segments + stack_on.split('/'))07:48
lifelessvaguely07:48
jmland what for stack_on_pwd?07:48
lifelessdunno07:49
lifelesssee what happens first07:49
jmlsnrk07:49
jmlok07:49
lifelessmy emulator is too small07:49
jmlhttp://paste.ubuntu.com/193256/07:52
jmlI think maybe all of the ../ should go into stack_on_pwd07:53
lifelessjml: looks like we need to say that if stack_on == '/', stack_on_pwd should be ''07:53
lifelessor '.'07:53
lifelessjml: no, because if stack_on == '/'+ anyything, stack_on_pwd is ignored07:54
jmllifeless: stack_on_pwd = '' -> http://paste.ubuntu.com/193258/07:55
jmlstack_on_pwd = '.' -> http://paste.ubuntu.com/193259/07:55
lifelessstack_on has too many ..'s07:56
lifelessuhm07:56
igcjml, jelmer: see http://bundlebuggy.aaronbentley.com/project/bzr/request/%3C20090605165828.GA4605%40vernstok.nl%3E07:57
lifelessperhaps stack_on_pwd should be used to normalise stack_on somewhat07:57
lifelessthat is, the ..'s from stack_on_pwd shrink stack_on07:57
jmllifeless: something like this? http://paste.ubuntu.com/193260/07:59
jml(it doesn't work of course. sanity checking)08:00
lifelessno08:01
lifelessyou're replacing the pwd with the ..'s from the stack_on08:01
lifelessI mean adjust by08:01
lifelessso segments = ['..'] * (len() - stack_on_pwd.count('..')) #sketch08:01
lifelessthats clearly wrong; presumably we have a path joining function somewhere08:02
spiv(not following closely, but there's urlutils.pathjoin)08:03
lifelessjml: or the stack_on version we had before08:03
lifelessjml: with stack_on = pathjoin(stack_on_pwd, stack_on)08:04
lifelessand stack_on_pwd = '.'08:04
jmlso, stack_on_pwd = '.', and then calculate how to get from target_transport.base to /foo/bar/baz as a relative path.08:11
lifelessjml: no. I think you've gottenconfused.08:12
lifelessjml: we have two inputs08:12
lifelessstack_on and stack_on_pwd08:12
lifelesswe know that combining them in the server chroot works08:12
lifelesswe want two outputs08:12
jmlso far I am with you.08:13
lifelesssuch that combining those two and the clients url for the bzrdir we created will will work from the client08:13
lifelessthere are several options here08:14
lifelessI need to call it 'done' for the day; possibly the most simple thing is to say08:14
lifeless'/' stack_on paths are not usable in the smart server. When we see one we will take the transport to the thing we stacked on and transform it as per _repo_relpath08:15
lifelessso toss out all the data08:15
pooliejml, are you referring to spiv's patch for bug 380314?08:18
ubottuLaunchpad bug 380314 in bzr "Branch.revision_history RPC fails on a stacked branch" [High,In progress] https://launchpad.net/bugs/38031408:18
jmlpoolie: that one, yes.08:18
spivpoolie: diff viewable at https://code.edge.launchpad.net/~spiv/bzr/stacking-friendly-revision-history-verb/+merge/731408:19
=== abentley1 is now known as abentley
jmligc: did you end up doing your eol tweak?08:20
lifelessjml: anyhow, there are several ways you could do it. I don't have a preference.08:20
jmllifeless: ok, thanks.08:20
jmllifeless: I'll figure something out.08:20
igcjml: yes. It's landed I believe08:20
* igc checks08:20
* jml checks too08:20
mwhudsonjml: how do you get -D hpss like stuff in the selftest's .bzr.log ?08:21
mwhudsoni thought it was -E hpss08:21
jmligc: should I mark https://bugs.edge.launchpad.net/bzr/+bug/362030 as Fix Released then?08:21
mwhudsonbut that doesn't seem to be it08:21
ubottuLaunchpad bug 362030 in bzr "files whose content changed after EOL filtering erroneously marked as modified" [High,In progress]08:21
jmlmwhudson: -Dhpss selftest -Eallow_debug08:21
mwhudsonah08:21
mwhudsonta08:22
igcjml: just did it - thanks for reminding me08:24
jmligc: cheers.08:24
pooliejml, spiv, review done08:27
poolieand S says "get ready!"08:27
poolieso i will08:27
spivpoolie: thanks08:27
jmlpoolie: thanks!08:27
jmlpoolie: have a good evening :)08:28
jmljust requested bzr branch be mirrored immediately08:28
jmlyay08:28
poolie2a format sent to pqm08:29
igcnight all - good luck jml with the release08:29
jmligc: g'night.08:29
jmligc: thanks for your help :)08:29
pooliethe launchpad workflow for bugs and code reviews is getting really nice08:30
poolieit could be even better but it's pretty helpful08:30
jmlyeah.08:30
spivjml: have replied to poolie's review, will do the minor tweaks and send to PQM08:34
jmlspiv: music to my frosty ears08:35
lifelesspoolie: I'm still really concerned that people will try 2 and then be wedged irrecoverably08:36
lifelessvila: jml: re bialix's reported problem can you clarify.08:38
jmllifeless: running the script client side only affects the upload copy of the branch08:39
spmspiv: pqm is still working for you then?08:39
jmllifeless: it won't affect the mirrored copy.08:39
lifelessjml: but the mirrored copy is only accessed over http right?08:39
spivspm: we will find out shortly!08:40
jmllifeless: it's also accessible over bzr+ssh & sftp08:40
lifelesssftp like http is VFS anyway08:40
jmllifeless: where you'll get the mirrored copy iff you lack write access to the branch.08:40
lifelessok08:40
lifelessso this raises the 'when is lp bulk fixing things'08:41
lifelessquestion08:41
jmlyes.08:41
lifelessok08:42
lifelesswell, i'm going to ring poolie quickly and then -> gone08:42
jmlok.08:42
jmlthanks for the help w/ the release.08:42
lifelessnp08:42
lifelessjml: are you cuttting the release tonight?08:48
jmllifeless: yes.08:48
lifelessjml: or do you think it will be tomorrow? I have a change I want to make that poolie and I don't quite agree on.08:48
spivspm: my merge request just came and went in a flash...08:49
jmllifeless: it's got to be tonight, I'm afraid.08:49
jmlor rather.08:50
jmlit's got to be the 11th.08:50
lifelesswhat happens on the 12th?08:54
lifelessanyhow, you're doing rc1 right?08:55
fullermdA plague of locusts o'er the land.08:55
lifelessjml: ^08:55
jmllifeless: yes, rc108:56
lifelesswe can always change it between rc and release08:56
lifelessso don't block on this08:56
jmllifeless: great news :)08:56
tmetroWhat's the best way to get in touch with the maintainer of the bzr packages on backports.org? I suspect the Debian packaging tools view version 1.5 as newer than 1.13. Although it looks like 1.13 has been around since April, so I would have expected someone to notice this before now.09:05
lifelesstmetro: dpkg considers 1.5 < 1.1309:05
lifelesstmetro: there used to be contact details on backports.org09:05
lifelessa bug tracker of some sort09:05
tmetroHmmm...I have 1.5 installed, even though I have backports as a repository. It doesn't think there is a newer version. Trying to explicitly install bzr=1.13.1-1~bpo50+1 results in:09:07
tmetrobzrtools: Depends: bzr (< 1.6~) but 1.13.1-1~bpo50+1 is to be installed.09:07
tmetroOK, I can check backports.org, if they're the responsible party for creating the builds, and not just hosting them.09:08
lifelesstmetro: they certainly are.09:08
tmetrook, thanks.09:08
fullermdtmetro: Er, that's telling you your installed bzrtools won't work with 1.13.109:08
lifelessyou will need to install bzrtools *and* bzr at the same time because bzrtools is version locked to a specific bzr release.09:08
tmetroBut it should pull in bzrtools automatically...regardless, something is broken if I have to specify an explicit version. (Unless I added an apt config that I'm forgetting about.)09:10
lifelesstmetro: IIRC with backports you have to list each thing you want09:10
lifelessso you'd need to add bzr and bzrtools to your config09:10
tmetroAh, I see, they only have 1.5 of bzrtools on backports.org.09:11
tmetrohttp://backports.org/debian/pool/main/b/bzrtools/09:11
tmetroI'll send them an email...thanks.09:12
spivspm: yep, lp: URL in merge request still vanishing without a trace.  I'm going to workaround it with http:// again.09:14
jmlok. I think I've got this critical bug sorted.10:02
jmlnow checking that it fixes the original problem.10:03
awilkinsDoes .bzrrules work for in-tree content filtering rules?10:03
awilkinsAnd I'm having "trouble" with eol filtering10:05
spivjml: my branch landed, btw10:13
jmlspiv: cool!10:13
* jml updates the bug 10:13
jmlspiv: I have a patch you can review. I'm blocked on testing it against Launchpad though. :(10:14
jmlvila: hello10:23
=== abentley1 is now known as abentley
awilkinsOk, correct me if I'm wrong .. If I have a folder that contains a mix of LF-ended and CRLF-ended (*.cs) files, and I add a rule [name *.cs], then after a touch, on a Windows working tree, the result should be that i) bzr reports that LF files created as LF and saved as CRLF since checkout (but no other change) are unchanged 2) LF files that have not been re-EOLed have been changed10:35
awilkinsNot sure about the inverse though.10:36
awilkinsDammit, I hate tools that switch EOLs on you10:37
awilkinsOn linux shouldn't it take files that were created CRLF, check them out as LF.. then what. The file hasn't changed (according to the filter rule, if it's "Native") but should the stored version get flipped to LFs or stay as CRLF?10:39
jmlawilkins: sorry, I don't really have a clue about that.10:47
awilkins(the hating-tools-that switch EOl is directed at SharpDevelop, not Bazaar)10:47
jmlspiv: so this fix affects both the client and the server.10:47
awilkinsjml: I'm trying to figure it out10:48
jmlspiv: 1.15 is completely broken wrt Launchpad's use of stacking.10:48
stbuehlerbzr needed nearly 30 seconds to push one commit to svn; imho that is not really acceptable: http://paste.lighttpd.net/18710:57
stbuehlerany chances this will be fixed?10:57
jmlspiv: https://code.launchpad.net/~jml/bzr/default-stacking-bug-385132/+merge/733011:02
jml(or any other bzr-core member)11:02
bialixhi, anybody can suggest me the way to find unused imports in python code?11:05
bialixwill be great if this method will look into bzr lazy_imports too11:06
spivjml: glancing now11:06
jmlspiv: ta.11:06
jmlbialix: pyflakes11:06
jmlbialix: there's a branch of pyflakes that is bzr lazy_imports aware11:07
spivjml: It's not immediately clear to me that given "full_path = self._root_client_path + final_stack[1:]", that calculating the relative_url of that against target_transport.base is sane.11:07
spivjml: it certainly feels odd...11:08
jmlbialix: if you use emacs, you can also hook things up so that pyflakes runs all the time (using flymake).11:08
jmlspiv: what did you expect?11:08
spivjml: hmm, so s/full_path/client_path/ for clarity, I think11:08
bialixno, I'm not emacs user11:08
spivjml: but, the target_transport.base is presumably ignorant of any _root_client_path11:08
spivjml: and may be pointing at a chroot or something, right?11:09
spivjml: I would understand relative_url(_root_client_path, client_path)11:09
bialixI guess this is the branch you mentioned https://code.launchpad.net/~mwhudson/pyflakes/support-lazy-imports11:09
jmlbialix: yeah, I guess so too :)11:11
jmlbialix: I don't actually use that one.11:11
jmlspiv: changed.11:11
spivjml: but I don't see why relative_url([backing_transport_url], [a_path_adjusted_for_clients_view]) makes sense.11:11
spivThis is potentially a failing of me rather than the code, but you'll need to convince me :)11:12
jmlI've changed it to  relative_url(_root_client_path, client_path)11:12
spivjml: cool.  does it still work? :)11:12
awilkinsErm, I think eol support is broken, but that could just be me11:12
jmlyes.11:12
spivEven better!11:12
jmlawilkins: it's possible. We landed some patches about that today.11:13
awilkinsFiles that you commit on Linux with LF endings don't expand to CRLF on windows, even with eol=naitve11:13
jmlawilkins: bug 362030 and bug 37937011:13
ubottuLaunchpad bug 362030 in bzr "files whose content changed after EOL filtering erroneously marked as modified" [High,Fix released] https://launchpad.net/bugs/36203011:13
ubottuLaunchpad bug 379370 in bzr "bzr: ERROR: Unknown eol value 'None'" [Medium,Fix released] https://launchpad.net/bugs/37937011:13
awilkinsBut files with CRLF on Windows with eol = native do collapse to LF11:13
spivjml: omit self.reset_smart_call_log() calls if you aren't going to inspect the call log in the test11:14
jmlspiv: done.11:14
bialixpyflakes has no help?11:18
bialixI need just to run it and specify modules? or I can pass the entire package11:19
jmlbialix: you pass it filenames, I believe.11:20
spivjml: reviewed (approve; no other issues)11:20
jmlspiv: thanks. any thoughts on the fact that bzr 1.15 will break on pushing new branches to Launchpad?11:20
spivjml: I don't have any great ideas about what to do for old clients.11:20
lifelessjml: are there client side changes needed?11:21
jmllifeless: yes.11:21
bialixcool11:21
lifelessjml: thats surprising to me.11:21
lifelessif 1.15 is broken, 1.15.211:21
bialixpyflakes even suggest modules to make them lazy11:21
bialixwow11:21
lifelessor bump the server verbs if needed.11:21
lifelessI'm not here.11:22
* lifeless hides11:22
jmlbialix: #divmod if ever you have detailed pyflakes questions.11:22
spivLaunchpad (and potentially bzr core) could peek at the 'Software version' header and reply with UnknownMethod for that verb...11:22
bialixjml: I guess suggestions about lazy makes sense only for bzr code11:22
bialixe.g.11:22
jmllifeless, spiv: I think this is something that we can resolve over the next week before the final release.11:23
spivjml: +111:23
bialixC:\work\Bazaar\repos\qbzr-repo-1.9\trunk\lib>pyflakes commands.py11:23
bialixcommands.py:21: import of 'os' could be lazy11:23
bialixcommands.py:22: import of 'sys' could be lazy11:23
bialixcommands.py:23: import of 'ui' could be lazy11:23
bialixcommands.py:25: 'get_cmd_object' imported but unused11:23
bialixcommands.py:25: 'register_command' imported but unused11:23
bialixcommands.py:32: 'QtCore' imported but unused11:23
bialixmwhudson: hi11:23
lifelessheh. note that making things lazy that will be needed for most commands is unneeded overhead.11:23
jmlbialix: the lazy branch might actually be up for review on divmod's tracker.11:23
* jml shrugs11:23
bialixso twisted uses lazy imports too?11:24
spivNo.11:24
spiv(But maybe they should!)11:24
bialixjml: pyflakes even works for the packages! wow11:25
bialixthx!!!11:25
spivbialix: pyflakes is pretty excellent :)11:26
bialixyeah!11:26
bialixthis is what I need for so long11:27
jmlbialix: also, since it's really fast, it's something that you might be able to always have running & integrate with your editor.11:29
bialixor as part of test suite11:30
jmlI guess.11:31
jmlit still has false positives.11:31
jmlfor Launchpad, we have a wrapper around bzr send that does a pyflakes check on all changed files & includes that in the body of the email.11:31
jmlso they can be caught at review time.11:32
jmlthat seems to work ok.11:32
awilkinsHmm, the eol filtering applies to initial checkouts, but not to subsequent updates11:32
awilkinsPlus I've now noticed the "you checked it out fresh but everything is changed" bug11:32
fullermdawilkins: That (initial but not update) behavior is one of the things I noticed playing with keywords that just stopped it dead.11:33
awilkinsIs there a bug for that?11:33
fullermdDunno.11:33
awilkinsIt's also not clear to me what retro-active behaviour should be on e.g. - files that were created CRLF but are now content filtered and should be LF in storage11:34
awilkins(created with no filter)11:34
fullermdI didn't file one.  It seems like such an absolute showstopper that I assumed it had to be "incomplete", not "unknown bug".11:34
awilkinsigc: Ping?11:34
fullermdHe left about 3 hours ago.11:35
awilkinsDarn11:35
* awilkins forsees an "eol" tag for the bug tracker11:36
fullermdWell, since I saw it with keywords, I'd assume it's more "filter" than "eol".11:36
bialixmere mortals most likely find eol, though11:39
awilkinsAnd EOL and ignore are the only filters as yet11:40
awilkinsWell, in the default11:40
spivfullermd, awilkins: I think igc would appreciate bug reports11:42
* awilkins is filling out one and also composing overall mail of all bugs he is aware of with feature11:42
fullermdMy experience (and memory of it, come to that) is stale.  Bug reports of "I sorta remember way back when X happened" are rarely appreciated   :p11:49
jmlnoooo, one test fails.11:54
jmlhttp://paste.ubuntu.com/193374/11:55
GPHemsleyIs there a way to uninit?11:59
jmlGPHemsley: what precisely do you mean?12:02
GPHemsleyI did an init12:03
GPHemsleyand now I want to undo it12:03
jmlGPHemsley: remove the .bzr file from the directory inited.12:03
jmlGPHemsley: beware, you'll lose any history you've made.12:03
GPHemsleyno command, then?12:04
GPHemsleyno history12:04
jmlGPHemsley: no command.12:05
GPHemsleynow, would you recommand creating a trunk/ directory?12:06
jmlbzr init trunk12:06
GPHemsleyOK, but I'm asking if you think that's a good idea?12:06
jmlGPHemsley: it really depends on the broader context.12:07
jmlGPHemsley: a lot of the time I do this:12:07
jmlbzr init-repo <project-name>12:07
jmlbzr init <project-name>/trunk12:07
GPHemsleyany way to directly convert an init to an init-repo?12:08
fullermdNo more than there are ways to directly convert a submarine to a skyscraper.12:08
GPHemsleyfair enough12:08
jmlGPHemsley: there are things you can do, but they are all a pain, and rarely worth it.12:09
jmlCertainly not if this is for a new project.12:09
GPHemsleyk12:09
fullermdThey're totally different things; it doesn't make sense to convert something that you'd init into something that you'd init-repo.12:09
GPHemsleycd <project-name>; bzr init-repo .12:11
GPHemsleyis this same thing, right?12:11
GPHemsley(as the first command above)12:11
jmlyep.12:11
GPHemsleyand then I can cd trunk and bzr init .12:12
GPHemsley?12:12
jmlfor reference, when you do 'init-repo', you're saying "this directory is going to be a shared repository; a bucket where all of the branches beneath it store their data"12:12
GPHemsleyright12:13
jmlGPHemsley: or just bzr init trunk12:13
GPHemsleyI get this straight eventually12:13
GPHemsleys/I/I'll/12:13
GPHemsleywell, trunk already exists12:13
GPHemsleybut no matter, it's done12:13
jmlGPHemsley: when you say "bzr init foo", you're saying "I'm starting a new branch -- a new line of development --called 'foo'"12:13
GPHemsleyoh12:13
GPHemsleyhmm12:13
jmlyou tend to need at least one of these per project :)12:14
GPHemsleyyeah12:14
GPHemsley:)12:14
GPHemsleycan bzr store symlinks12:49
GPHemsley?12:49
=== mrevell is now known as mrevell-lunch
spivGPHemsley: yes12:50
GPHemsleycool12:50
Peng_But can it handle them decently when you try to check out on a platform that doesn't support them?13:04
* Peng_ goes /away again.13:05
=== Kissaki is now known as Kissaki^0ff
=== Kissaki^0ff is now known as Kissaki
jmlpoolie: looks like your format rename branch didn't land.13:26
awilkinsPeng_: You need a plugin for them on Win3213:38
Peng_awilkins: Oh, nice.13:38
awilkinsPeng_: But AFAIK NTFS does support symlinks with varying degrees of support13:38
awilkinsPeng_: I think the support gets a bit better in Vista and 713:39
awilkinsPeng_: On XP it's limited to "juncton points" which only link folders on the same physical volume13:40
awilkinsVista and up support file links, cross volume links, and network symlinks (but only on servers that also support them)13:41
awilkinsAnd by default non-admin users can't create them... nice13:42
Peng_I see. Thanks for the information. -- Eh, non-admin users can't create any symlinks, or just weird ones?13:42
awilkinsPeng_: If you are a non-admin or non-elevated admin you can't create a symlink in the default config13:42
awilkinsBut you can change the policy13:42
Peng_...That's dumb.13:43
Peng_Maybe it's to prevent symlink attacks?13:43
awilkinsPeng_: Who knows why... perhaps only the ACL for the link is considered when accessing the file, so it could be a potential hack to create links to system files and write viruses into them13:44
awilkinsPeng_: But without testing it....13:46
fullermdIt's tough to get right.  They only had, what, 25 years of prior art to look at for it...13:46
awilkinsShouldn't be too hard to write into the Python library to use it though (but you'd have to throw an error about setting the policy)13:47
* awilkins requires tea13:48
=== mrevell-lunch is now known as mrevell
=== thekorn_ is now known as thekorn
* jml halts14:04
awilkinsDoes the py2exe build process strip the libraries it puts in library.zip?14:32
ddaamostly, library.zip contains pyc files, so stripping does not aplly14:33
ddaaif there are any .dll, they'll (presumably) be in whatever state they were found on the system.14:33
awilkinsddaa: I don't mean it in the "strip stuff out of libraries" sense, sorry, I meant, does it ignore libraries for which their are no dependencies on the code14:33
awilkinsE.g. - if it doesn't depend on SimpleXMLRPCServer, it won't be in library.zip14:34
awilkins(I think this may be the case)14:34
ddaayes it does14:34
awilkinsAha14:34
awilkinsThat explains why bzr-eclipse doesn't work anymore14:34
awilkinsCAn you just plonk the .py file in the lib folder and have it work?14:34
ddaayou need to specify some badly documented option to have setup.py include the libs you want in the zip14:35
* awilkins tries the plonk-the-lib-in method14:35
ddaathat will work14:35
ddaabut it's ugly14:35
awilkinsI'm not sure it does.. unless SimpleXMLRPCServer now has missing dependencies14:36
bialixyou can put missing py files into library.zip14:37
awilkinsbialix: Ah, but not straight into /lib ?14:37
bialixnope14:37
ddaaoh, misread what you meant :)14:37
bialixawilkins: we using another approach in QBzr14:37
awilkinsthe _lib folder?14:38
bialixyes14:38
awilkinsOh wait, that's gtk14:38
awilkinsAh, as well14:38
bialixso any plugin can support all required but missing libs14:38
bialixalthough it's a bit boring14:38
awilkinsI'm jsut trying again, I don't think it killed the process that it had previously spawned14:38
bialixawilkins: are you using bzr-svn?14:39
awilkinsYes14:39
bialixthen you can't rebuild bzr.exe without all svn dlls14:40
awilkinsbialix: Yes, I had a win32 build env for bzr going before and it was a real pain to set up14:40
bialixor may be can14:40
awilkinsbialix: Since then it's all gone kaput14:41
bialixI'm build bzr.exe for myself w/o bzr-svn14:41
awilkinsbialix: You should be able to rebuild the exe as long as you have the built extensions (with a little finagling and judicious touching)14:41
bialixawilikins: in fact I think one can use compiled bzr-svn from official installer14:42
awilkinsbialix: Does the python-flavoured one have the bzr-svn extensions in it yet or does it just not bundle them at all now?14:42
bialixpython-based installer installs only bzr+bzrlib+docs, nothing more14:43
bialixso you agin have to install bzr-svn separately14:43
* bialix never had enough patience to learn bzr-svn build process14:46
awilkinsIt's not fun14:46
awilkinsWell, not on windows14:46
bialixone man gave me instructions14:46
awilkinsOn Linux it's doss easy14:46
bialixapt-get14:46
bialix?14:47
awilkinsEven building from scratch14:47
awilkinsYou just apt-get a few dev packages14:47
bialixI'm not sure it's entirely Windows problem. Some credits should go to svn developers14:48
awilkinsWell, there is that14:48
awilkinsI'm sure glad they have those prebuilt dev libs for Windows because the thought of actually building it makes the blood run cold14:48
* bialix nods14:49
tedgIs there a way to regenerate the indicies?14:51
tedgI'm getting this error: bzr: ERROR: Error in data for index GraphIndex('file:///home/ted/Development/inkscape.newbzr/.bzr/repository/indices/1d2b8fb37e32a1e94a733fbc3d195756.tix').14:52
awilkinsDoes py2exe deliberately not compress the content of library.zip ?14:53
bialixawilkins: it's controlled by options in setup.py14:54
bialixwhen I wrote py2exe support in 2006 I've decided to not compress it14:54
awilkinsI wonder if it might improve startup time to compress it (reduced disk IO vs CPU time)  .. I wouldn't like to guess without testing it14:55
bialixas I vaguely remembered Mark Hammond suggest to actually compress it.14:56
bialixI see what you mean14:56
bialixthe difference will be really small14:56
awilkinsOh yes indeed14:56
bialixtens of milliseconds14:56
bialixin the end windows has to read 1 file14:57
bialixand will ikely this file could be cached by disk subsystem14:57
awilkinsWould pyflakes tell me what bzr-elcipse is dependant on14:57
bialixno14:57
bialixbut you can use modulefinder.py module from bzr.dev/tools14:58
awilkinsGroovy14:58
bialixerr14:58
bialixpackage_mf.py14:58
awilkinsAh, it's all java14:58
awilkinsAnd makes XMLRPC calls14:58
bialixbang14:59
awilkinsSo I guess I just need to check what xmloutput and SimpleXMLRPCServer use14:59
bialixat least14:59
bialixyes,14:59
bialixyou don't have python traceback from java, do you?14:59
awilkinsbialix: It spawns a process and talks to it over a socket, so no15:00
awilkinsbialix: Unless you trapped the xml server to do it15:00
bialixit was half/joke15:01
* bialix tried to run eclipse and look at qbzr-eclipse. it looks very decent15:01
bialixperhaps not too serious like bzr-eclipse15:02
awilkinsIt's probably best I stick with it, all my users are on it15:02
awilkinsFor some reason, people get antsy about having new software installed15:02
bialixI'm trying to imagine "ansty"15:04
* bialix cannot15:04
awilkinsI think it's "the feeling of discomfort ones gets from having ants in ones pants"15:04
bialixit was second half/joke15:05
bialixdon't mind15:05
awilkinsDon't know how Russian you are :-)15:05
bialixmay be I'm just need to sleep a bit more15:06
bialixbut yes, I'm15:06
Spabbyhi guys anyone awake?15:07
* bialix half awake15:08
Spabby:D15:08
SpabbyI'm wondering if it's possible to use bazaar to run the  way I want, my friend Google hasn't been much help15:08
bialixyou want asking about bzr-svn?15:08
bialixwell, bzr can't make sandwich for you15:09
bialixso it's hard to suggest you something without more details15:09
SpabbyI want to have a central repo that myself and a colleague can merge to, but I want what I merge to be visible as a proper website, Ie I want to checkout a branch, edit the files, then merge back to the central server, at which point the changes are reflected on the website15:10
Spabbysorry I was typing a long sentence ;)15:10
SpabbyI'm not particularly concerned about the sandwiches15:10
bialixit depends how you configure your server and bzr repo15:11
bialixlook at plugins: push-and-update andd/or upload15:11
bialixor use hook on smart server15:12
Spabbythere are plugins that will push and upload to a server?15:12
bialix2 all: this question asked very frequently. Does bzr has some sort of FAQ?15:12
SpabbyI've read the FAQ15:13
pygibialix: lies, it can make a sandwich :(15:13
bialixpush-and-update IMO15:13
pygithe book claims it can :(15:13
Spabbyand i've googled like mad15:13
SpabbyI'm not looking for easy answers, I'm just looking for some help and direction15:13
pygiSpabby: loggerhead perhaps?15:14
bialixpush-and-update:15:14
pygiif I got your requirements right15:14
bialixProvide the "push-and-update" command to automate the running of 'ssh remote bzr update path' after pushing to sftp://remote/path. This allows us to keep a remote working tree up to date.15:14
* pygi just came in :p15:14
bialixSpabby: does it relevant for you?15:14
Spabbyyes I think so15:14
Spabbyif I'm reading it correctly, after the changes have been pushed, I can trigger an update on another path15:15
bialixor, if you really need webbrowser for your repository -- loggerhead15:15
Spabbyso if I checkout the bzr branch to the webroot15:15
Spabbywhen I push I can trigger an update on that branch?15:15
bialixif you have smart server running on your server (e.g. bzr+ssh)15:15
SpabbyI'm not stupid, honest :D15:15
bialixme too15:16
bialixI'm trying to help15:16
bialixI'm just bad in English15:16
Spabbyno you are helping15:16
Spabbyand for that I thank you15:16
SpabbyI'm reading on the push-update plugin15:16
bialixI mean bzr FAQ need more love15:19
Spabbyok15:19
SpabbyI think I need to buy a good book :D15:19
bialixor just ask here15:22
* awilkins should write the Bazaar book and make oodles of cash15:24
awilkinsNot that I'm especially knowledgeable, just especially fond of cash15:24
SpabbyI need to understand the basics I think15:24
awilkinsAre you experienced with SVN or CVS?15:25
awilkins(or other VCS systems with central repo)?15:25
Spabbyfor someone who is fairly knowledgable in a wide sphere of IT based stuff, revision control evades me15:25
SpabbyI used svn for about a year but I kept breaking it15:25
Spabbyand I never really understood it15:25
Spabbywell15:25
SpabbyI understood how to update my working copy, and checkin changes15:26
Spabbybut not how it all worked15:26
awilkinsSVN is just another filesystem with the time dimension included15:26
awilkinsYou can go back, create absurb paradoxes, etc15:26
* bialix likes paradoxes15:26
Spabbymy problem with svn is that if I deleted a local file I seemed to completely break the repository15:27
awilkinsThat is oddd15:27
Spabbyor at least my ability to checkin to the repo15:27
ddaaI would like to be able not to make it sound like a putdown, but you are clearly confused about something.15:27
ddaaBut we'd need a lot more information to able to identify what you are confused about.15:28
SpabbyI'm confused about a lot, it's not a criticsm15:29
SpabbyI guess that the bzr repo is a database of files, and how those files have changed (at a broad level)15:29
Spabbyand I understand how I can use bzr to manage my revisions on  local level by using commit15:30
Spabbybut  I struggle to understand how it works in a centralized environment15:30
awilkinsSpabby: In a centralized environment you would use bound branches15:31
awilkinsSpabby: The commit occurs either locally_and_centrally or just centrally15:31
Spabbyright15:31
SpabbyI guess the process we want to use is to commit locally during development of a branch, and then centrally when you are ready to publish your changes to the project15:32
awilkinsSpabby: The Bazaar development branch method is an excellent example15:32
Spabbyand I guess not understanding it makes that very difficult for me to visualise how bzr would handle that15:32
awilkinsSpabby: All commits to the main trunk of development are made by the patch queue manager15:32
awilkinsSpabby: And only after they pass testing.15:32
Spabbyis that the "Gatekeeper" from the bzr manual?15:33
awilkinsOne manifestation of a Gatekeeper, yes15:33
Spabbyok15:33
bialixthe great and terrific one15:33
awilkinsDo you have qbzr?15:33
awilkinsOr bzr-gtk?15:33
SpabbyI'm going to read the development branch method15:33
awilkinsJust looking at the revision graph for bzr.dev would be very informative15:34
Spabbyim using cl bzr on Centos ATM to try and setup the central location15:34
awilkinscl bzr ?15:34
SpabbyI seem to have created the branch and commited the initial files to it15:34
Spabbycommand line15:35
awilkinsAh15:35
awilkinsRight, what serving methods do you have available?15:35
Spabbyhttp, sftp, ftp, or any other that can be installed if it is better15:35
awilkins(and which bzr version are you using? Last I looked the CentOS packages were quite old)15:35
mattlhey, what does this mean?15:35
mattlbzr: ERROR: Revision {daniel@daniel-watkins.co.uk-20090412154020-0fk0vbvg644oe8zo} not present in "254@1c61c228-c9dc-448a-9a81-c9cbcf67beca:trunk%2Fweb%2Fdata%2FServer.php".15:36
awilkinsIf you and all users have ssh access to server than that would work well15:36
Takmattl: bzr-svn?15:36
awilkinsOtherwise, http only works both ways with effort15:36
Spabbyyes we have ssh access15:36
mattlTak: this was an svn that was converted to bzr.15:36
Spabbythis is a development server hosted on our lan15:36
Spabbyso ssh would be fine15:36
awilkinsSpabby: Use bzr+ssh then, that's probably the least-setup way to do it15:36
Spabbyhow do I find out version number sorry awilkins15:37
awilkins`bzr version`15:37
ddaamattl: that means some revision in the svn repo was committed with bzr, but it's missing from your bzr repo (because bzr commit on svn repo do not store the parent revisions of a merge in the svn repo)15:37
Spabbyah15:37
Spabbyyou'd think they would include that in the default command list15:37
Spabbyim running 1.3.115:37
mattlddaa: so, how do we fix this?15:37
awilkinsSpabby: That's the old version I was referring to15:37
Spabbyok15:37
ddaamattl: bzr is normally about to deal with this, they are called ghost revisions15:37
Spabbyi'll try and get an update15:38
ddaamattl: so this is a bug that this is causing a crash15:38
awilkinsSpabby: If you have root it's not hard to build from sources15:38
mattlddaa: is there a way i can skip these revisions?15:38
Spabbyok15:38
SpabbyI will do that now :D15:38
awilkinsSpabby: But things will still work at the moment15:38
ddaamattl: as I told you, it's a bug that the presence of ghost revisions is causing any problem at all.15:38
Spabbyah ok15:38
Spabbyso I have a repo setup15:39
awilkinsYour client machine should be able to `bzr checkout bzr+ssh://server/full/path/to/branch15:39
ddaamattl: you cannot "skip them", it's just not the way bzr-svn works.15:39
Spabbyat least if I do "bzr info" I see the branch15:39
mattlddaa: right, so... should i try a later version of bzr, or is my project repo screwed?15:39
Spabbyhow do I check that out to my local pc?15:39
awilkins`bzr checkout bzr+ssh://server/full/path/to/branch`15:39
ddaamattl: if you can identify the bzr branch that was merged into svn prior to the import you are using, you can fill the ghost revisions using "bzr fetch-ghosts"15:39
ddaamattl: you should ALWAYS try the latest version of bzr and bzr-svn when you have any sort of problem.15:40
awilkinsSpabby: That produces a bound checkout - commits will be mirrored to the server15:40
awilkinsSpabby: But all the revisions are also held locally15:40
Spabbyand if there are conflicts I will be prompted to merge?15:40
awilkinsSpabby: Yes15:40
Spabbyexcellent15:40
awilkinsSpabby: Or update15:40
SpabbyI thank you15:40
ddaamattl: as I told you, your repo is not screwed, the presence of ghosts revision is not a problem, bzr and bzr-svn SHOULD be able to deal with them. If you still have this problem using the latest bzr, you should file a bug so bzr can be fixed.15:41
ddaaactually, ghost revisions is a feature15:42
ddaaas it allows round-tripping of bzr commits through a svn repo.15:42
Spabbyawilkins that checked out great15:45
Spabbythanks again15:45
awilkinsSpabby: You're welcome. Soon you'll progress to multiple branches in a no-trees repo and switch between them like a gazelle15:45
Spabbythat worked great15:50
Spabbybut when I commit my changes to the central, is there any way I can force an update on the branch there?15:51
awilkinsSpabby: THere are plugins15:53
awilkinsSpabby: Or you could use a server side hook15:53
Spabbyserver side hook sounds intriging15:53
bialixbut anyway you need plugin on server side to hook it in15:56
SpabbyI'm guessing push and commit are not the same thing15:57
awilkinsNope15:57
=== herb__ is now known as herb
Spabbysorry to perster you awilkins, im using tortoiseBzr on my local machine, any idea how I can get it to use the plugin automirror?16:04
bialixSpabby: so now you're on Windows? great16:09
Spabby:D16:09
Spabbyonly temporarily16:09
Spabbyactaully scratch that16:09
Spabbyboth of us will be working on linux boxes16:09
bialixyou need to get the copy of this plugin16:10
bialixso why you using tvzr?16:10
SpabbyI'm working on my windows box atm because I have 1m other things going on16:12
Spabbybut when we start developing this app it will be coded on linux boxes16:12
bialixget the plugin with command: bzr get lp:bzr-automirror automirror16:12
SpabbyI think that's a job for tomorrow :D thanks for all your help again16:13
bialixput automirror directory (with plugin) to ~/.bazaar/plugins on Linux or %APPDATA%\bazaar\2.0\plugins on Windows16:13
bialixand read the help on this plugin16:13
Spabbythanks all, have a good evening16:14
bialixthat's basically all16:14
bialixbzr need magic button to automagically install plugins16:14
awilkinsI find that `bzr branch lp:<plugin-name> <plugin>`  usually works nicely16:15
bialixhow you're explain usually where one should put <plugin> on Windows?16:17
awilkinsIn powershell $env:appdata\bazaar\2.0\plugins16:17
awilkinsIn cmd.exe  %appdata$\bazaar\2.0\plugins instead16:18
awilkinsAck %appdata%16:18
awilkinsNot $16:18
bialixanyway ~/.bazaar/plugin is way too much shorter16:19
awilkinsYes16:20
awilkinsHey, powershell supports ~ for %userprofile%16:21
awilkinsWhy did they put BZR_HOME in %appdata% ?16:21
awilkins I didn't know it supported ~16:21
awilkinsDammit, that would be much easier16:22
awilkinsAnd why the fricking 2.016:22
bialixhehe16:23
bialixI know it's jam invent it16:23
bialixbut now it's too late to change16:24
bialixbut hey!16:24
bialixwe should have bzr 2.0 next month!16:24
bialixso16:24
awilkinsI'd write a change-the-location-of-BZR_HOME plugin16:24
bialixjam invented time mechine in 200616:24
awilkinsBut it wouldn't be able to load itself16:24
bialixso you need to write it virus-like16:25
bialixso it will reborn itself after total windows reinstall16:25
bialix:-)16:25
bialixmore seriously: I'm always using BZR_PLUGIN_PATH settings16:26
RockyRoadHi :)16:46
RockyRoadWould someone know what would happen (I mean, would it be safe) if I deleted, from the repository, a branch taking part in the history of another ?16:46
Peng_RockyRoad: Don't go deleting .bzr/repository data that's necessary. .bzr/branch doesn't matter.16:46
Peng_RockyRoad: (Unless you're using stacked branches, obviously.)16:46
RockyRoadI didn't yet understand stacked branches16:47
Peng_RockyRoad: When using stacking, it says "look at branch X for all the other data" instead of storing it all locally.16:48
RockyRoadHow can I check it is safe16:48
Peng_RockyRoad: "bzr info", I suppose.16:48
hsn_is there SLES prebuilt package?16:48
* Peng_ /away!16:49
Peng_RockyRoad: On disk, branches (.bzr/branch) are simply pointers to data in the repository (.bzr/repository). Deleting them is perfectly safe. However, when using stacked branches, you may have several different repositories holding a few revisions each, so it may not be safe.16:50
RockyRoadI'm working on lp:planet-drupal16:51
Peng_Soo?16:51
RockyRoad(just to let you check picture if you wanted) I think lp:~m-baert/planet-drupal/planet-6x is left behind after a sequence of merges16:52
RockyRoadbut if the others refer to it, they would break16:53
RockyRoadI've rebuilt all the bzr history from multiple branches16:54
RockyRoadI think I'd better check ids in bzr log xmloutput16:57
LarstiQhsn_: I think there have been, I don't know who is responsible for them17:09
jamawilkins, bialix: We used "APPDATA" because that is where you are "supposed" to put application specific information on Windows17:35
jamWe used "Bazaar\X.X" because that is also what you are recommended to do17:35
jamAnd we used 2.0 because at that time17:35
jambzr was likely to be released as Bazaar 2.0 rather than our final choice of releasing it as bzr 1.017:36
jamBut as Alexander says, it will be there in another month or so :)17:36
beunoso18:19
beunoafter upgrading about 30 branches at the office18:19
beunoall of them upgraded fine18:19
beunothey're all faster18:19
beunobut about half of them are the same size in bbc than in 1.918:19
visik7is there a way to get mail notification on commit ?18:23
visik7I want to develop a protocol for bazaar18:25
* SamB wonders why aptitude can't show changelogs for packages from PPAs18:26
SamBhmm ... "bzr trees http://libarchive.googlecode.com/svn/" is failing for me using bzr 1.15+4422+110 ...18:31
SamB bzr trees http://libarchive.googlecode.com/svn/18:31
SamBbzr: ERROR: Transport operation not possible: Transport <bzrlib.transport.http._urllib.HttpTransport_urllib url=http://libarchive.googlecode.com/svn/> has not implemented list_dir (but must claim to be listable to trigger this error).18:31
SamB... oops, I missed the "%"18:32
wlawlessIs there a good site to visit if I am having problems with bzr-svn on Windows?18:32
SamBwlawless: talk to jelmer18:33
SamB... who is conveniently missing18:34
SamBwhat kind of problems?18:34
wlawlessa bunch of 'procedure entry point missing' errors when I try to do anything that involves the svn repository (checkout, update, commit)18:35
SamB("talk to jelmer" is my advice for most situations where a bzr-svn bug is suspected ;-)18:35
SamBwlawless: could you paste one ?18:35
wlawless"The procedure entry point svn_client_revprop_set2 could not be located in the dynamic link library libsvn_client-1.dll"18:36
wlawlessplus about 8 more, every time I run a command18:36
SamBhow'd you instal bzr-svn ?18:39
SamBdid libsvn_client come with it ?18:39
wlawlessI used the default installer for windows from here: http://bazaar-vcs.org/Download18:40
wlawlessthe instructions says bzr-svn is included18:40
SamBdid you try googling for the error message yet/18:41
SamBer, ?18:41
SamBhmm, well, I can't find a thing18:42
schmichaelis there any reason not to always use bzr merge --pull?18:42
wlawlessI did, not much luck18:42
SamB... next question: search for files named "libsvn_client-1.dll"18:42
SamBon your hard drive18:42
wlawlessnot that I know of, I'm new to bazaar, so was just following the tutorials I could find18:43
wlawlessthe only instance I could find was in c:\program files\Subversion18:43
wlawlessactually, nevermind, there is one in c:\program files\bazaar as well18:44
wlawlesswould bzr-svn care what version of the svn client I have on my machine?18:45
bialixgaryvdm: good evening18:59
garyvdmHi bialix18:59
bialixI'm packaging release right now19:00
bialixgaryvdm: done, trunk is open for 0.1219:28
garyvdmCool19:29
garyvdmI'm adding some comments to loggraphprovider19:30
bialixthank you19:30
bialixnow the most boring part:mark bugs as fix released19:31
wlawlessdoes the latest bzr-svn work against subversion 1.6 repositories?19:36
garyvdmbialix: let me help with that.19:42
bialixok19:42
bialixgaryvdm: we talked about bug statuses some months ago. you can use fix released when you push your fix to trunk19:43
bialixI've got too much karma from closing bugs19:44
bialixthanks, done now19:45
garyvdmDo you get karma for changing from fix committed to fix released?19:46
bialixyep19:46
zu22can someone please help me with bzr import19:46
zu22this is driving me crazy!19:46
bialixand I don't think it's quite rigt19:46
zu22i followed all documented steps i could find and it still fails19:46
zu22anyone here familiar with bzr's fast-import?19:47
garyvdmbialix: I think the bug assigne should get the karma19:47
bialixme a bit19:47
zu22i am using darcs fast-export and then bzr fast-import19:47
zu22bialix: ok cool one sec i will get a pastebin for you to see19:47
garyvdmzu22 - can you pastebin some code.19:47
zu22sure one sec19:47
zu22gotta fire up firefox, slow computer :)19:47
bialixgaryvdm: after each release I see karma jump19:48
bialixyou can simply explain19:48
garyvdmIt's rewarding you greatly for a boring job ;-)19:48
bialixha ha19:49
bialixIIUC bzr 1.16 will release brisbane-core19:50
bialixwe need to try conversion of our trunk to bbc (locally)19:50
bialixgaryvdm: you don't have a chance to test brisbane core yet?19:51
garyvdmI've got a bbc branch of bzr.dev that I have been using to test some parts of qbzr against. But I have not tested upgrading the qbzr branch.19:53
garyvdmI'll brb19:53
zu22ok got it19:53
zu22bialix and garyvdm you guys still here?19:53
zu22sorry for the delay!19:53
zu22http://zu22.pastebin.com/f145ddea919:53
zu22that shows what is going on19:53
bialixcool, you have your own pastebin19:54
zu22heh ya19:55
bialixzu22: bzr: ERROR: No WorkingTree exists for19:55
bialixyou should run command: bzr co .19:55
zu22bialix: so what am i doing wrong?19:55
bialixor simply bzr co19:55
bialixrtfm?19:55
bialixrun bzr co19:56
zu22i am converting a darcs repo to a bzr19:56
zu22ok19:56
zu22i just did "bzr co" in /var/www/darcs/netrek-client-brmh.bzr and it said:19:57
zu22bzr: ERROR: Not a branch: "/var/www/darcs/netrek-client-brmh.bzr/.bzr/branch/".19:57
zu22it seems either darcs export is failing or bzr import or both :(19:57
zu22any more ideas?19:57
* bialix re-read paste one more time19:58
zu22lines 31 and 32 is where conversion should happen19:58
bialixok, let's try that without cd to .bzr dir and ls it, there is nothing very interesting19:59
bialixcd trunk19:59
bialixbzr info -v20:00
bialixI need output of info command20:00
bialixzu22: bzr info -v in trunk after import20:01
zu22ok i will get that20:02
zu22http://zu22.pastebin.com/f6f26e4c520:03
bialixzu22: now: cd trunk; bzr info -v20:04
bialixdoes your darcs repo has only 2 commits?20:04
zu22ok20:04
zu22there is no trunk directory20:05
zu22there is "branch-lock" and "repository" directories20:05
zu22my darcs repo should have 8 patches in it20:05
zu22but the bzr conversion should import ALL my source code20:05
bialixdon't don't don't don't cd to .bzr/ dir please20:05
zu22the original code + my code in the darcs repo20:05
zu22it doesn't import any code at all20:06
zu22ok20:06
bialixI see 2 revisions in your trunk20:06
zu22you can pull my darcs repo and see yourself:20:06
bialixC:\work\Bazaar\repos>bzr log http://www.jesujuva.org/darcs/netrek-client-brmh.bzr/trunk/ --line20:06
bialix2: netrek 2009-06-10 decimal precision conformance patch + compile fixes + updated ChangeLog20:06
bialix1: netrek 2009-06-10 Import old source code.20:06
zu22darcs get http://www.jesujuva.org/darcs/netrek-client-brmh/20:06
bialixI don't have darcs installed, sorry20:06
zu22yes20:06
zu22ok that is correct 2 big patches20:06
bialixcan you do following:20:07
bialixbzr info -v /var/www/darcs/netrek-client-brmh.bzr/trunk20:07
zu22i will do it now nad pastebin20:08
zu22one second20:08
zu22http://zu22.pastebin.com/f1b2199520:10
bialixdoes location /var/www/darcs/netrek-client-brmh.bzr/trunk writable for you?20:11
bialixthis is valid branch. you need to run in /var/www/darcs/netrek-client-brmh.bzr/trunk either `bz revert` or `bzr update`20:12
zu22it has permissions: drwxr-xr-x20:13
bialixone of these comands will restore your working files20:13
zu22oh ok20:13
bialixso it's read-only for you20:13
zu22then i will be able see all my code good20:13
bialixdo you have writable location in your session20:13
bialix?20:13
zu22but i run all my commands as root20:13
zu22yes root can write it :)20:13
zu22let me try bzr revert20:13
bialixyou can get a fresh copy20:14
zu22ok how do i do that?20:14
bialixbzr branch /var/www/darcs/netrek-client-brmh.bzr/trunk ~/trunk.bzr20:14
bialixlast part of command -- is path where new copy will be created20:14
bialixbzr branch -h | less20:15
bialixthis say you more20:15
bialixyou can browse your trunk with command: bzr log /var/www/darcs/netrek-client-brmh.bzr/trunk20:15
zu22ah20:16
bialixcheck your patches with command: bzr diff /var/www/darcs/netrek-client-brmh.bzr/trunk -c120:16
bialixbzr diff /var/www/darcs/netrek-client-brmh.bzr/trunk -c220:16
zu22cool bzr has diff!20:16
bialixso, I;d say conversion was successful20:16
bialixrun this: bzr help command|less20:16
bialixI recommend you to read this: http://doc.bazaar-vcs.org/bzr.dev/en/mini-tutorial/index.html20:17
zu22bialix: the 'bzr update' worked, now i must find out how to get my bzr repo to import into launchpad.net :)20:18
bialixhttp://doc.bazaar-vcs.org/bzr.dev/en/mini-tutorial/index.html#publishing-your-branch-with-launchpad20:18
zu22thank you bialix20:19
bialixnp20:19
zu22bialix: you are Ukraine?20:19
bialixbut it seems you need to read this mini-tutorial20:19
zu22yes20:19
bialixyes, from Ukraine20:19
zu22good20:19
zu22i remember your leader20:19
zu22he was poisoned :(20:20
zu22is he ok now?20:20
bialixdon't worry about him20:20
bialixhe's not very good leader20:20
bialixwhy you migrate from darcs?20:21
zu22our project switch to bzr20:21
zu22but i prefer darcs :)020:21
zu22so i will keep using darcs and import to project using bzr20:21
bialixI guess fast-import/fast-export should work for you20:21
zu22yeah i just wish is a nice GUI to do all this20:22
zu22so i must not fiddle with so many command line options, too easy to make mistake hehe20:22
bialixGUI for fast-import?20:23
zu22for both :)20:26
bialixI'd write a simple script or Makefile20:26
zu22good idea20:28
zu22bialix: do you have your own website?20:28
bialixmore or less20:28
bialixwhy you asking?20:28
mib_2l987k8uHow do I completely remove a file from history?20:31
mib_2l987k8uA colleague added a directory of temp files by mistake, and they waste space.20:32
Peng_mib_2l987k8u: Mostly you don't.20:32
bialixhow many revisions you have after this mistake?20:32
mib_2l987k8uAbout two20:32
mib_2l987k8u>1, if that's what you mean :)20:32
bialixthen create fresh copy of your branch before mistake20:33
bialixand replay all good commits in new branch20:33
bialixthen kill old branch20:33
mib_2l987k8uBut I cannot (as in Git) just remove the files from each commit?20:33
bialixyou can try this with fastexport/fast-import20:34
mib_2l987k8uOK, thanks20:34
bialixbut if you have only one revision on top, I'd better replay20:34
mib_2l987k8uWith uncommit?20:34
bialixno20:34
bialixwith rebase plugin20:34
bialixlike in git20:35
bialixrebase plugin has handy command: replay20:35
mib_2l987k8uAh20:35
mib_2l987k8uOK, I'll RTFM for the plugin20:35
mib_2l987k8uThanks20:35
bialixnp20:35
bialixbut to save the space you need to get fresh copy and delete old branch20:35
bialixgaryvdm: https://bugs.launchpad.net/qbzr/+bug/32859820:36
ubottuLaunchpad bug 328598 in qbzr "Create a revisions selector widget. Use in commands like push/pull." [Medium,In progress]20:36
mib_2l987k8uAh20:37
mib_2l987k8uThat's the main thing I'm trying to gain: space20:37
mib_2l987k8uSince the temp files are just text, but lots of them20:37
bialixbzr has no garbage collector20:37
mib_2l987k8uHmm20:37
garyvdmbialix: There are 2 out standing issues that I ran into - that would need lots of codeing20:37
mib_2l987k8uIf I revert, just make the final changes, then commit and repack, would I regain the space?20:38
bialixmib_.... no20:38
mib_2l987k8uDamn20:38
bialixuncommitted revisions is not garbage collected20:38
mib_2l987k8uMaybe I should just bend over and take the wasted space, and hope bzr packs well enough as it is20:38
bialixgaryvdm: :-(20:39
garyvdm<mib_2l987k8u: if you uncommit, and then push somewhere - that uncommitted rev won't be pushed.20:39
mib_2l987k8uThat would be possible.20:40
bialixis there short path? may be to limit its functionality a bit?20:40
hsn_you know if bzr-trac works with Trac 0.11.4? i am getting some exceptions20:40
garyvdmbialix: I'ts usable as it is. May be we should look at merging it as it is.20:40
bialixhow to test it?20:41
garyvdmbialix: qpull20:41
garyvdmqpush20:41
bialixok, will look now20:41
garyvdmqmerge - i think20:41
bialixhsn_: jelmer or ddaa might know20:42
garyvdmbialix: One of the issues is that you can't select a revision range, to say do a cherry pick merge.20:42
bialixthis is bearable20:43
bialixknown bug is not bug anymore, but feature20:43
bialixhaaa!20:43
bialixErrorFromSmartServer: Error received from smart server: ('error', "'AbsentContentFactory' object has no attribute 'g20:43
bialixet_bytes_as'")20:43
bialixfrom your branch20:44
bialixcan you try to fix it?20:44
garyvdmbialix: how?20:44
bialixwait a sec20:44
bialixhttps://bugs.launchpad.net/launchpad-code/+bug/35403620:45
ubottuLaunchpad bug 354036 in bzr "ErrorFromSmartServer - AbsentContentFactory object has no attribute 'get_bytes_as' exception while pulling from Launchpad" [Undecided,Confirmed]20:45
bialixrun the script from this bug report as described20:45
garyvdmon local branch, and then push?20:46
bialixno20:47
bialixRun it as 'fix-branch.py bzr+ssh://bazaar.launchpad.net/~user/project/branch'.20:47
garyvdmok20:47
meoblast001hi20:47
meoblast001i'm trying to help someone set up his new computer with his old bzr email line20:48
meoblast001how can  be sure i have the exact same line as the old20:48
bialixbzr whoami20:48
mib_2l987k8ubzr whoami on his machine to set the new one20:48
mib_2l987k8ucheck the old logs for his old one20:48
mib_2l987k8ubzr whoami "Sam Adams <good@beer.com>"20:48
mib_2l987k8uOr whichever beer your friend prefers20:49
meoblast001i don't know what his old one was though20:49
meoblast001how can i find the logs20:49
mib_2l987k8uRun bzr log on the old repo20:49
mib_2l987k8uOn any machine with bzr installed20:49
rockstarmeoblast001, bzr log on the old branches20:49
meoblast001ok20:50
garyvdmbialix: I ran it and got http://paste.ubuntu.com/193697/20:52
garyvdmI don't know if that is good or bad?20:53
meoblast001rockstar: thanks20:53
bialixgaryvdm: I guess now it fixed20:53
* bialix checking20:53
bialixgaryvdm: now this branch is ok, thanks20:58
garyvdmcool20:59
bialixgaryvdm: so.. your widget works21:04
bialixit's strange it highlight only one column instead of all row21:04
garyvdmThat the other issue that I can't fix.21:04
bialixmay be we really need to land it and start dogfoding21:04
garyvdmGoing offline for 1 min21:05
tmetroFYI, it looks like the Ubuntu Intrepid PPA:21:05
tmetrodeb http://ppa.launchpad.net/bzr/ppa/ubuntu intrepid main21:05
tmetrowill successfully work as a stand-in for backports.org, which currently is missing bzrtools, for Debian stable.21:05
bialixgaryvdm: but it produce tracebacks sometime21:06
bialixe.g. in qtag21:06
bialixor in qpush when branch is not selected (empty string)21:07
garyvdmbialix: Yes are rough edges like that21:07
garyvdmBut the big issue is the it sometimes selects only one column21:07
bialixit's not sometimes for me21:07
bialixit's all the time21:08
garyvdmAnd I can't enable multiple selection.21:08
garyvdmYes - i't all the time on windows - some time on linux21:08
bialixyou don't need multiple selection for pull/push21:08
garyvdmThe fix for both issues is the same.21:08
bialixfor qmerge -- maybe just provide 2 revision selectors?21:08
garyvdmI need to rewrite large bits of QComboBox21:09
bialixmaybe bug with column somehow related to custom painting?21:09
bialixI assume you draw entire widget21:10
garyvdmNo - QComboBox insists that you can select cels21:10
garyvdmAnd insists that you can only select 1 item21:10
* Nafallo wonders if bzr import have changed recently21:11
garyvdmI've tried all sorts of hacks to get around that to no avail.21:11
Nafalloit's either that or my memory not working anymore :-P21:11
bialixgaryvdm: ar eyou using "highlighted" signal?21:12
garyvdmbrb21:12
bialixgaryvdm: it fails in strange way when I've tried to invoke qdiff from qlog21:14
bialixso, it seems is not ready yet to land21:14
garyvdmbialix: sorry - some unknown process keeps downloading something  - and I work out what it is.21:15
bialixyou hunting on virus?21:15
garyvdmbialix: Yhea - I need to go through it with a fine comb. There have been a number of merges with trunk that were quite confusing.21:16
garyvdmI don't know21:16
garyvdmAnd netstat is confusing21:16
bialixare you on Windows right now?21:16
garyvdmNo - ubuntu21:17
bialixoh, ok21:17
garyvdmOk - going to restart to see if that helps.21:18
bialixgaryvdm: I'm going offline. Do you have questions for me?21:23
garyvdmSeems to be fixed now21:23
garyvdmbialix: No - I'm working on making one widget for browse, commit, add, revert and qmain21:24
bialixcool21:24
bialixI will start to review patches from igc and javier21:24
bialixtomorrow21:24
garyvdmWill have better performance and will fix bug 25892921:25
ubottuLaunchpad bug 258929 in qbzr "Commit dialog does not show contents of directory that is not versioned only the directory itself" [Low,Confirmed] https://launchpad.net/bugs/25892921:25
garyvdmcool21:25
bialixI guess this is need to be done at wt_list level?21:25
bialixbtw21:25
garyvdmYes21:25
bialixtoday I've tested pyflakes21:25
garyvdmbut I have not get there today21:25
garyvdmSorry - I have not got there yet21:26
bialixwe have a lot of unused imports21:26
bialixmay be we need from time to time clean it21:26
bialixhttps://code.launchpad.net/~mwhudson/pyflakes/lazy-import-support21:26
bialixgaryvdm: ^21:26
garyvdmcool21:27
bialixno21:27
bialixthis one https://code.launchpad.net/~mwhudson/pyflakes/support-lazy-imports21:27
bialixjml/spiv can advertise it21:27
garyvdmOk - I'll give that a try - I allways forget to check for unnecessary imports.21:28
bialixme too21:28
bialixbut this tool is damn good21:28
bialixthe branch from mwhudson knows about bzr lazy imports21:28
bialixit's incredible21:29
mwhudson:)21:29
bialix:-)21:29
mwhudsonit's a bit trigger-happy about suggesting imports be lazy21:29
bialixyeah, I'm talking about pyflakes21:29
bialixyeah, it was so incredible cool21:29
bialixthe best thing about pyflakes: it does not have extra dependencies21:31
bialixe.g. I don;t need to install twisted to run pyflakes21:31
bialixI like it21:31
bialixbtw Gary, sometimes it's interesting to see unversioned (but to be added file) in the qdiff21:33
bialixe.g. if we select it to add before commit21:33
mwhudsonit's really fast too, that's important21:33
mwhudsonyou can just run it all the time21:33
bialixmwhudson: yeah, it's really fast21:34
mwhudson(unlike pylint)21:34
bialixit's even faster than bzr21:34
* Tak head explode21:42
* bialix z-z-z21:42
jdubanyone about who can help with a weird merging problem with bzr-svn?21:47
kenichihello, i'm trying to understand the way post_change_branch_tip hooks get called.  if i'm committing from a checkout to a local filesystem branch, would it make sense for this hook to be called twice?22:46
SamBwho should I talk to about bzr-bisect? it apparantly doesn't use launchpad for bugs ...22:49
garyvdmkenichi: I think if you commit to a heavyweight checkout - it changes the tip of the local branch (the checkout), and the bound branch.23:06
garyvdmSo post_change_branch_tip probably gets called for each of the branches23:06
garyvdm*I think*23:06
kenichigaryvdm: ah i see, i think that makes sense.  i will try with a lightweight to verify.  thanks!23:08
kenichigaryvdm: yes, verified only one call with a lightweight.23:09
kirklanddoesn't seem that i can add an "-sa" on the end of a "bzr bd -S"23:27
kirklandsuggestions?23:28
kirklandi found:   bzr bd --builder "debuild -S -sa" --source23:32
kirklandthat might do it for me23:32
lifelessI think that is what I used for a similar thing23:34
lifelessI can dig up my scripts if that doesn't work (I was building in a pbuilder myself)23:34
kirklandlifeless: cheers23:34
Gnx-still the same question, is there a way to teach bzr that something is a binary file and should not be merged?23:46
james_wkirkland: "bzr bd -S -- -sa"23:49
james_wanything after "--" is passed through to debuild verbatim23:49
james_wanything that starts with a "-" at least23:50
james_wI'd like to improve that a bit, but I'm not sure what the best plan of action is23:50
lifelessGnx-: no; in fact some binaries can be merged very well, and there are text files that don't merge properly. So it is something we would like to address.23:53
Gnx-lifeless: well, I have an sqlite.db and it currently goes wonk from bzr23:53
lifelesswhat do you mean by wonk?23:54
Gnx-well bzr makes it .THIS .OTHER23:54
Gnx-the behaviour I'd want would be that it would overwrite to whatever is in the last commit23:55
lifelessGnx-: just run 'bzr revert sqlitedbfilename'23:55
lifelessGnx-: I realise its not fully automatic, but it should work.23:55
Gnx-I meant like bzr update -> database gets overwritten from the update source23:56
lifelessGnx-: yes23:56
lifelessGnx-: bzr update + bzr revert databasefilename == database is overwritten from the update source23:57
Gnx-ah23:57
Gnx-okay23:57
lifelessGnx-: isn't that what you are asking for?23:57
Gnx-but I thought revert after update gets you the version from before the update23:58
lifelessrevert gets you the current basis23:58
lifelessupdate changes the basis23:58
lifelessafter a *merge* revert would get you what you had before the merge23:58
Gnx-ah okay, thanks, bzr is kind of MyFirstVCS ;)23:59
lifelessfor a merge you would do 'bzr revert -r branch:mergesource databasefilename'23:59
lifelessor, more simply23:59
lifelesscp dbfilename.OTHER dbfilename23:59

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!