Team Chat Logs

April 24, 2010

2010 3
Mo Tu We Th Fr Sa Su
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30    
<< <   > >>

[00:34:39.549775]<davidfraser>Morning Hodgestar
[00:34:52.068744]<davidfraser>Looking forward to the sprint - will likely get there around 1-ish
[00:35:39.724814]<Hodgestar>Cool.
[02:01:54.730529]<Hodgestar>Off to run some errands before the sprint.
[03:25:32.972046]<Hodgestar>Step one is to create tickets for Walter's patches so they can be looked at in parallel.
[03:30:48.791019]<Hodgestar>http://bitten.edgewall.org/ticket/570 -- patch 1.
[03:33:15.817270]<Hodgestar>http://bitten.edgewall.org/ticket/571
[03:35:05.003855]<osimons>Hodgestar: name them .patch or .diff - that way they render in the Trac. .pat is not a recognized file ending
[03:35:32.334475]<Hodgestar>osimons: Will do. I was just uploading Walter's files.
[03:35:55.531990]<osimons>see that. cool. he had a batch stored then?
[03:37:17.342283]<techtonik>Would it be better to put them on http://codereview.appspot.com/
[03:37:19.386651]<techtonik>?
[03:37:24.650412]<Hodgestar>osimons: He did. :)
[03:37:53.658896]<Hodgestar>techtonik: Nah. They're really small.
[03:38:04.214778]<Hodgestar>And I'm hoping they won't require too much discussion.
[03:44:49.917247]<techtonik>Bug - http://bitten.edgewall.org/attachment/ticket/570/0001-Allow-timeout-parameter-to-be-set-on-sh-exec-python-.pat
[03:45:56.438959]<Hodgestar>Already renamed.
[03:45:57.267351]<Hodgestar>:)
[03:46:16.050850]<techtonik>I would like to propose moving "Build all revisions" options to "Configurations" or make it overridable.
[03:48:16.615427]<techtonik>For stable branches it is good to receive notifications about every failed revision, but for trunk/ it is better not to do this.
[03:48:31.815919]<Hodgestar>confluency, drnlm and jerith are at the sprint in Cape Town with me.
[03:49:13.543314]<Hodgestar>techtonik: What's wrong with the option present on the config page for the individual configs?
[03:50:02.281989]<techtonik>Hodgestar: there is no such option in b2
[03:56:52.816462]<Hodgestar>techtonik: Oh. I see -- it's only available as a master setting. Gotcha.
[03:57:30.575144]<techtonik>It is also an issue for HG style of development, when series of revisions are pushed to "trunk" at once and those often contain failed tests.
[03:58:11.701010]<Hodgestar>osimons: I'm moving the tickets I hoping we'll deal with today to Milestone 0.6. Currently this mostly affects the issues touched by Walter's patches.
[04:27:11.811401]<confluency>Is anyone currently busy on any of these?
[04:35:19.029413]<drnlm>I've been looking at #552
[04:35:46.011822]<drnlm>I fasr as I can see, there isn't any way to convince the subversion client to fail rather than warn in this case
[04:38:07.016851]<techtonik>File a bug to subversion?
[04:48:00.614246]<Hodgestar>Okay. I'm done with Walter's patches for the moment.
[04:48:14.017320]<Hodgestar>There is one more patch but it's bigger so I'll leave it for later.
[04:48:22.438410]<davidfraser>Hodgestar: OK, what shall I look at?
[04:48:37.876624]<Hodgestar>That's the next step -- finding issues for everyone to work on. :)
[04:48:55.223250]<Hodgestar>davidfraser: I was wondering if you want to work on your charting graphs patch and try get that in?
[04:49:04.911684]<davidfraser>Ah, good plan
[04:49:27.781865]<techtonik>What's the number?
[04:49:39.509937]<Hodgestar>davidfraser: It's quite a big change so I guess we'll want an okay from osimon's before final commit but the licensing issue is important for inclusion in distributions.
[04:49:52.273197]<techtonik>pacopablo: we need evil_twin autolinking here
[04:51:48.837502]<Hodgestar>techtonik: http://bitten.edgewall.org/ticket/426
[04:57:20.157911]<confluency>I'm going to look at 533/505/458/243.
[04:59:24.529251]<techtonik>Licensing for charting engine is an issue, but there should be an option to use Google Charts - they are awesome. XML charts screen is better than fplot. =/ It should be possible to compile flash files using MTASC until a better JS lib with compatible license is available.
[05:02:24.085014]<Hodgestar>http://ctpug.org.za/wiki/Meeting20100424 --> I've added a list of who is working on which issues currently.
[05:02:31.621826]<Hodgestar>If I've missed anyone, let me know.
[05:03:12.358866]<techtonik>Considering http://bitten.edgewall.org/ticket/505 - I can't run development code on my Debian server - only released versions from backports. It is more productive to point users to released versions or at least instruct how can they build them.
[05:03:55.723045]<techtonik>Is there an easy way to checkout and make bitten env running for testing?
[05:04:07.595752]<osimons>oh. got less time today than i thought, so just need and back out again now. i'll trust you judgement, Hodgestar & davidfraser. additional care on things that may break existing installs (like protocol changes) but otherwise i think most things are better off being worked on than not...
[05:04:24.657275]<davidfraser>osimons: Cool, thanks
[05:04:33.611313]<techtonik>How about release schedule?
[05:04:41.330510]<Hodgestar>osimons: One of Walter's patch has a database upgrade that adds a field to builds. That okay with you?
[05:04:46.801273]<techtonik>What bugs a critical to make a checkpoiynt?
[05:04:48.189751]<Hodgestar>osimons: It's an extra time field.
[05:04:56.861951]<osimons>Hodgestar: no idea :-)
[05:05:01.603030]<davidfraser>techtonik: If we get far enough today, then we should be able to release in the next week or so, methinks
[05:05:15.507004]<osimons>cool. good luck, all. later.
[05:05:21.382418]<Hodgestar>osimons: Bye! Thanks. :)
[05:05:22.258340]<techtonik>osimons: bye
[05:06:05.821179]<techtonik>It would be nice to see release criteria and potential bottlenecks/tricky stuff.
[05:07:09.950409]<Hodgestar>techtonik: For testing here I've made virtualenvs for us to use for development. Virtualenv has Trac, Bitten, mini-subversion repo, and a dummy Python project that has some tests and can trigger builds.
[05:07:50.790084]<techtonik>Do you have a python script that builds that virtualenv?
[05:07:52.007823]<Hodgestar>techtonik: Unfortunately porting it across machines requires some tweaking (virtualenv --relocatable was less relocatable than I'd hoped).
[05:09:30.005192]<techtonik>In any case it would be worthy to upload at least instructions to BittenDev and perhaps a mini-repository somewhere too.
[05:12:40.682716]<Hodgestar>I think a script that builds the env is a really cool idea.
[05:13:03.768302]<Hodgestar>techtonik: Do you feel like working on that?
[05:13:37.632400]<techtonik>If you send me some entrypoint information, then yes. =)
[05:21:51.677728]<Hodgestar>murphypa is also here with us.
[05:31:55.641291]<drnlm>I'm abandoning 552, as there doesn't look any way of solving that issue without adding some sort of parsing of subversion's checkout output
[05:40:20.045242]<Hodgestar>wbell: Hello! I've created tickets for your patches.
[05:40:32.328284]<Hodgestar>wbell: Or updated existing tickets as appropriate.
[05:40:39.371524]<wbell>Hello all!
[05:41:14.089012]<wbell>Thanks-- let me catch up on the progress you guys have made.
[05:41:24.401564]<wbell>So you're mostly working on 0.6.3?
[05:41:45.699140]<Hodgestar>drnlm: Are you going to investigate adding a svn:checkcheckout recipe command or are you moving on entirely?
[05:42:25.412764]<Hodgestar>wbell: The focus is mainly on the 0.6 Milestone issues -- and I guess we're aiming to release 0.6b3 sometime after the sprint.
[05:43:16.462629]<Hodgestar>wbell: The progress so far has mostly been in getting everyone here set up and organised with tasks. :)
[05:43:26.195969]<wbell>Hodgestar: this is in branches/0.6.x?
[05:43:37.521523]<wbell>Hodgestar: do you want a build setup for that?
[05:43:44.851976]<Hodgestar>wbell: trunk and 0.6.x are fairly identical at the moment.
[05:43:47.033176]<drnlm>Hodgestar: : I'm currently investigating a couple of ideas for a check script. I'll add anything interesting to the ticket and decide later
[05:43:59.521171]<Hodgestar>wbell: So I'll commit to trunk and backport to 0.6.x.
[05:44:09.154261]<Hodgestar>drnlm: Woot.
[05:44:27.562562]<wbell>Hodgestar: I can do the commits if you want.
[05:44:35.189575]<wbell>Might as well get my name on the records.
[05:44:35.727832]<Hodgestar>wbell: You're welcome to.
[05:44:43.451897]<Hodgestar>wbell: Lol. Woot.
[05:44:58.833822]<Hodgestar>wbell: Just let us know which tickets you're looking at so we don't duplicate effort.
[05:45:12.221912]<Hodgestar>wbell: And shout if you want anything backported.
[05:45:30.389020]<Hodgestar>I'm currently looking at #222.
[05:45:33.647428]<wbell>Hodgestar: sounds good.
[05:46:05.964845]<wbell>Hodgestar: the keepalive stuff I've done has an impact on #222.
[05:46:17.740640]<Hodgestar>http://ctpug.org.za/wiki/Meeting20100424 --> complete list of who is working on what.
[05:46:42.765977]<wbell>Hodgestar: the precursor to that patch makes it so a build only gets orphaned if the last step was before that timeout.
[05:46:45.216038]<Hodgestar>wbell: Ah.
[05:46:53.641021]<Hodgestar>wbell: I was hoping to leave the keepalive stuff for last.
[05:47:31.567283]<wbell>Hodgestar: seems reasonable. Prior to keepalives, what behavior would you want?
[05:48:10.497321]<Hodgestar>wbell: I'm not following the question?
[05:49:02.752356]<wbell>Hodgestar: Nevermind-- didn't see you applied my patch. I think that patch 7 fixes it.
[05:50:07.866919]<Hodgestar>Yep. Patch 7 is now linked to #222. I chose that ticket because it adds a new database field and I've worked on database upgrade stuff most.
[05:50:36.689562]<wbell>confluency: fixing #243 will make me very happy.
[05:51:03.451956]<confluency>Still reading through all the bits and pieces across all the bugs. :)
[05:54:59.137166]<Hodgestar>I think jerith is about to start work on #571.
[05:57:38.674516]<davidfraser>techtonik: There's a bastardized version of Hodgestar's bitten-sprint virtualenv at http://bitbucket.org/davidfraser/bitten-sprint - you may find r1 works better though :)
[06:02:44.449876]<frozty_sa>ohai from ctpug
[06:04:00.843136]<confluency>'sup
[06:04:02.868590]<wbell>New 0.6.x build is coming up http://bitten.edgewall.org/build/branches-0.6.x
[06:04:31.791819]<jerith>wbell: Is that your patch in #571?
[06:04:49.953166]<wbell>jerith: yep.
[06:05:12.743183]<jerith>Is there anything other than a review required?
[06:05:50.721457]<wbell>jerith: not that I know of. If people are cool with it, I'll push it into trunk and 0.6.x.
[06:06:48.576771]<Hodgestar>wbell: Don't forget to update the Milestone page (it has a list that osimons will eventually push into the change log).
[06:06:52.679774]<wbell>btw: It's been a long time since I've been on irc. Definitely let me know of any social missteps.
[06:06:58.870623]<jerith>I've just read over it and I'm not really familiar with the code, but nothing looks horribly broken in it.
[06:07:07.381367]<jerith>You have two tests with the same name, though.
[06:07:10.672696]<wbell>Hodgestar: good point. will do.
[06:07:16.576707]<wbell>jerith: will fix that.
[06:07:34.931890]<jerith>def test_onerror_defaults(self)
[06:09:12.587337]*jerith goes to look at recipe.py to see what happens in there.
[06:12:31.251495]<wbell>jerith: I've updated the patch-- nothing exciting:
[06:12:47.261907]<wbell>jerith: test_onerror_defaults -> test_onerror_override
[06:14:54.775892]<jerith>wbell: Is there anything else that we might want to set defaults for?
[06:15:12.468324]<Hodgestar>Oh noes. Failed builds. :)
[06:15:50.993398]<wbell>jerith: good question. I've needed the onerror one every bitten I've setup.
[06:16:16.874854]<wbell>jerith: I always want all the steps to run, but I want the overall build status to know if any of them failed.
[06:16:35.362797]<wbell>Hodgestar: yeah, one looks like it had an svn checkout issue. I've invalidated it.
[06:17:02.476141]<wbell>Hodgestar: At the bigger installation of bitten I have, we've had to wrap the svn checkout code in retries to handle those kinds of issues
[06:17:50.323598]<Hodgestar>wbell: Okay. Let me know if there's anything in the repo you need me to fix. :)
[06:18:18.049145]<wbell>Hodgestar: looks like things are good-- it's building a bunch of the old 0.6.x builds, some of which fail. Should be done in a few minutes.
[06:18:36.936782]<jerith>wbell: I'm wondering if we don't want a defaults dict rather than a param per default.
[06:19:10.388856]<wbell>jerith: I could go either way-- I figured the next param we added would force us to a dict.
[06:19:48.535881]*jerith shrugs.
[06:19:56.938037]<jerith>Seems fine the way it is, then.
[06:22:18.579230]<CIA-37>r830 by wbell in (5 files in 3 dirs): Allow ''timeout'' parameter to be set on sh:exec, python:exec and python:distutils commands, which kills the commands after the given number of ...
[06:24:40.056575]<techtonik>drnlm: About svn externals in 522 I've sent mail to subversion folks
[06:24:52.896700]<techtonik>the link in ticket.
[06:25:04.922174]<Hodgestar>wbell: It looks like the patch to #222 only updates last_activity when the build is initiated and when it stops. Am I missing something? If not, I'll try add something so that it's updated after each step is processed.
[06:25:23.448874]<wbell>jerith: cool. thanks.
[06:25:34.574252]<drnlm>techtonik: thanks
[06:26:03.353405]<wbell>Hodgestar: you are completely right.
[06:26:21.068587]<wbell>Hodgestar: must have missed that when I cut the patches apart. I can fix if you want.
[06:26:57.841459]<jerith>wbell: Just testing it now that I've patched my local copy.
[06:26:58.263600]<Hodgestar>wbell: I'm busy on it so I can do the modification (unless you're really looking for something to work on :).
[06:27:12.825299]<wbell>Hodgestar: nope. Go for it.
[06:35:31.485833]<CIA-37>r831 by wbell in branches/0.6.x/ (5 files in 3 dirs): Merge [830] from trunk.
[06:37:26.519058]<drnlm>OK, I've flogged #552 as far as I'm willing to, ticket updated with current notes
[06:37:47.440230]<Hodgestar>drnlm: Thanks.
[06:38:14.348275]<Hodgestar>wbell: Will you close http://bitten.edgewall.org/ticket/570 (assuming that's what r830 does)?
[06:38:14.915304]<CIA-37>r832 by wbell in (6 files in 4 dirs): Adding the ability to modify the default ''onerror'' property in the ''<build>'' element. If not specified, the behavior is unchanged; by default any ...
[06:38:49.639282]<wbell>Hodgestar: looking.
[06:39:18.352245]<CIA-37>r833 by wbell in branches/0.6.x/ (6 files in 4 dirs): Merge of [832] from trunk.
[06:39:18.655417]<wbell>Hodgestar: Ah. I didn't update my patch to auto-close your new tickets. Will close.
[06:39:19.943357]<confluency>Looks like half the 533 patch is already committed (the change in shtools.py).
[06:39:25.558355]<jerith>wbell: Just to clarify the different onerror behaviours: "fail" kills the build right there, "continue" marks it as failed but carries on building and "ignore" carries on as if the step succeeded?
[06:39:53.501866]<wbell>Jerith: correct.
[06:40:07.853245]<Hodgestar>confluency: Does the other half look sane? Any sign of why it wasn't committed?
[06:41:34.379715]<confluency>I'm looking at the repo now; the change may have been prompted by some other ticket.
[06:41:53.204186]<jerith>The old behaviour (as evidenced by a bug in the test config Hodgestar gave me) seems to treat anything that isn't "ignore" or "fail" as if it were "continue".
[06:42:55.127608]<Hodgestar>wbell: This isn't directly related to #222 but master.py only sets build.started = 0 on build cancellation but a bunch of other places also set build.stopped and build.last_activity to 0. Any idea what the preferred behaviour is?
[06:43:04.413139]<jerith>So "continue" apparently always worked, but only by accident. (Or very clever interactions between disconnected code paths.)
[06:43:24.318503]<wbell>jerith: yeah-- I had been using it for a while.
[06:43:31.799657]<wbell>jerith: I decided to make it concret.
[06:44:03.857484]<wbell>Hodgestar: There are about 4 places where people invalidate builds-- they really need to get consolidated. Any differences in them from what I know are just bugs.
[06:44:08.862831]<jerith>wbell: Your patch just makes it explicit and disallows random other values. (Which is a Good Thing. Tyops ftl.)
[06:46:13.052131]<wbell>Hodgestar: actually looking at my patch for #222, I set last_activity at the end of each step.
[06:46:25.190251]<wbell>Hodgestar: I think it's ok, no?
[06:47:12.201603]<wbell>jerith: yeah, it was a useful coincidence. I figure most people unfamiliar with the code would prefer something more explicit :)
[06:47:13.038228]<Hodgestar>wbell: Where does the setting happen?
[06:48:10.441052]<Hodgestar>wbell: Nm. I see my patch applying left out a hunk.
[06:48:16.248543]<wbell>Hodgestar: don't know what your repo looks like, but in the patch its: build.last_activity = step.stopped, when last_step is not set.
[06:48:43.946150]<Hodgestar>wbell: Oddly I had re-added the exact same characters. :)
[06:48:46.908365]<wbell>Hodgestar: it wouldn't surprise me if I missed something. I chopped up my commits to make them easier to read.
[06:48:56.456682]<drnlm>I've had I look at #529 - the patch looks sane to me, although it should probably treat the flag as a bool, not an int in the checks
[06:49:00.657698]<Hodgestar>wbell: Let me just read the upgrade stuff.
[06:49:14.232404]<wbell>Hodgestar: no rush.
[06:50:16.850453]<Hodgestar>drnlm: Does --no-auth-cache ignore the cache entirely or just not update it?
[06:51:16.871837]<techtonik>Bastardized version of virtualenv seems huge. I may run out of my last free 300mb soon.
[06:52:51.469231]<wbell>I can go either way on merging [834] to 0.6.x (assuming trunk is 0.7)
[06:52:53.337698]<wbell>Thoughts?
[06:53:02.388938]<CIA-37>r834 by wbell in bitten/ (8 files in 3 dirs): Charting fixes. Only show charts on configuration pages where there's a report for the chart's category in the range of active builds. Closes #454. ...
[06:53:42.317047]<Hodgestar>wbell: It's generally not safe to import from bitten.model in upgrades.py since the contents of bitten.model may change over time.
[06:54:00.498681]<wbell>Hodgestar: ah. Good point.
[06:54:15.316753]<drnlm>Hodgestar: I think it ignores it completely, will check on that
[06:54:31.232170]<wbell>Hodgestar: I was following the pattern there, probably on autopilot.
[06:54:39.522579]<wbell>Hodgestar: what's the recommended way to do that?
[06:54:42.503717]<Hodgestar>wbell: No problem -- you're not the first. :)
[06:55:03.828764]<wbell>Hodgestar: I'm trying to pick up good python habits from trac/bitten.
[06:56:04.475150]<Hodgestar>wbell: Pull in the actual scheme definition. I'm searching for the upgrade script method with a good example.
[06:56:18.585692]<wbell>Hodgestar: Do I just want to copy-paste the version of the build schema into that method to hardcode it for all time?
[06:57:11.431712]<confluency>OK, half of patch 533 was in patch 504, which was committed in an attempt to fix what looks like a similar issue.
[06:57:42.118575]<Hodgestar>wbell: If you can avoid including that whole thing that's useful, but yes, that's the general idea.
[06:58:47.349848]<Hodgestar>wbell: If you like you can put the schema definition outside the method and add a version number to the name.
[07:03:28.766851]<wbell>Hodgestar: can do. You find any good examples to follow
[07:04:14.225986]<Hodgestar>wbell: Maybe add_log_table().
[07:04:34.838105]<wbell>Hodgestar: ok. I'll match that.
[07:05:40.425368]<Hodgestar>wbell: One more thing -- give the temporary table a name unique to the upgrade (e.g. append '_v11') so that it's harder for temporary tables to clash names.
[07:06:28.157579]<Hodgestar>An existing upgrade example of that is 'old_log_v5'.
[07:06:53.212668]<Hodgestar>wbell: For 0.7 I will probably clear up all of upgrades.py -- it's just a scary thing to rewrite. :)
[07:07:30.458140]<wbell>Hodgestar: Will do. I don't envy you. Upgrade code is a fairly thankless task.
[07:10:40.818624]<confluency>I've added notes to 533 and 458; I'll try to find or make a testcase to see if this problem has actually been fixed now.
[07:14:34.861209]<wbell>Hodgestar: Updated patch to #222 for upgrades.
[07:15:34.350077]<drnlm>Hodgestar: scratch that, it just prevents the cache being updated, but it will still use entries in the cache if applicable
[07:15:37.505741]<Hodgestar>wbell: Great. I think it's basically ready to commit. I will just quickly run the new upgrade script through all the databases.
[07:15:52.220416]<wbell>Awesome. Let me know how it goes and I'll commit.
[07:15:54.148325]*jerith is looking at #466 now.
[07:16:24.366378]<Hodgestar>drnlm: That's actually better because it's nice to be able to add auth credentials to the cache on the build slave and then just have them used from that point on.
[07:19:19.157136]<Hodgestar>wbell: 2 hunks from master.py still don't apply cleanly (not a biggy, just a bit of a schlep :).
[07:21:02.544238]<Hodgestar>wbell: I would also update the comment above "if last_step:" to mention the new else part.
[07:22:07.338792]<CIA-37>r835 by wbell in branches/0.6.x/bitten/ (8 files in 3 dirs): Merge of [834] to 0.6
[07:29:28.415837]<CIA-37>r836 by wbell in bitten/ (slave.py master.py tests/master.py): Make all times be generated by the server. This eliminates a number of inconsistencies you'll see when the clocks on the slaves are skewed from the ...
[07:30:52.780703]<CIA-37>r837 by wbell in branches/0.6.x/bitten/ (slave.py master.py tests/master.py): Port of [836] to 0.6.x
[07:32:39.280355]<Hodgestar>wbell: The patch you uploaded to #222 doesn't appear to have the changes to upgrades.py (and fails my upgrade test :).
[07:33:02.800107]<wbell>Hodgestar: hmm. Let me look.
[07:33:36.537479]<wbell>Hodgestar: drat. one minute
[07:34:29.958541]<Hodgestar>wbell: Of course SQLite doesn't care about such minor problems as creating a table that already exists. :)
[07:36:52.374558]<wbell>Hodgestar: Ok. I actually uploaded a different patch this time :)
[07:37:06.320755]<wbell>Hodgestar: I certainly scrambled my db a bunch testing patches.
[07:44:55.170227]<Hodgestar>wbell: Still need to change old_build to old_build_v11 in the create table statement.
[07:46:10.226769]<wbell>Hodgestar: Really?
[07:46:44.149075]<wbell>Hodgestar: I guess you could be doing a bunch of updates at once and the "temporary" table would be scoped in all of them?
[07:46:52.461842]<wbell>Hodgestar: and thus maybe clash?
[07:47:02.882957]<Hodgestar>wbell: Yep,.
[07:52:25.351391]<wbell>Hodgestar: Good point.
[07:52:57.576531]<jerith>Yuck. The stuff that needs fixing for #466 lives in a template inside trac. :-/
[07:53:38.574394]<Hodgestar>wbell: I've got an update patch that passes all the database test. Will attach to #222 shortly.
[07:53:46.120388]<Hodgestar>wbell: Will add comment at the same time.
[07:53:54.737413]<Hodgestar>?
[07:54:50.037029]<wbell>Hodgestar: where's that from?
[07:55:28.686514]<Hodgestar>wbell: The update patch? I just wrote it. :)
[07:55:39.078278]<Hodgestar>s/te/ted/
[07:55:39.087411]<evil_twin>hodgestar meant: wbell: The updated patch? I just wrote it. :)
[07:55:45.619120]<wbell>Hodgestar: Ah. Nice.
[07:56:07.757429]<wbell>Hodgestar: is there an upgrade test suite?
[07:56:24.298623]<Hodgestar>wbell: There is tests/upgrade.py. :)
[07:56:35.601239]<Hodgestar>wbell: And I have a script that runs it under a variety of different databases.
[07:57:00.590778]<wbell>Hodgestar: Ah-- I've just been running the main test target, which presumably does tests/upgrades.py, but not different dbs.
[08:01:31.343138]<Hodgestar>wbell: Updated patched attached to #222.
[08:02:21.768477]<Hodgestar>Unfortunately testing against PostGreSQL and MySQL requires external setup so it's not easy to incorporate by default. I should add a my test scripts to the new dev section on the wiki.
[08:02:39.847704]<wbell>Hodgestar: I like it.
[08:02:46.486294]<wbell>Hodgestar: you want to commit?
[08:04:03.457230]<Hodgestar>wbell: You do it -- I just made some minor fixes. :)
[08:04:16.689559]<wbell>Hodgestar: will do. Thanks for the feedback.
[08:04:21.729048]<Hodgestar>wbell: Np. :)
[08:04:39.327799]<CIA-37>r838 by wbell in bitten/ (4 files in 3 dirs): Expand steps to allow in-progress steps. Some small additions to the BuildStep model, but most changes are in the ui to not assume steps are in their ...
[08:04:39.341664]<CIA-37>r839 by wbell in bitten/ (master.py tests/master.py): Master now creates next step as in progress when the previous step is complete. While a build is running, the currently running build step now shows ...
[08:12:11.711053]<CIA-37>r840 by wbell in bitten/ (7 files in 2 dirs): Add new last_activity field to build. I considered reusing stopped, but this seemed cleaner and more obvious, which seems like the right way to go. On ...
[08:16:01.282328]<Hodgestar>wbell: Will you backport those too?
[08:16:11.669272]<wbell>Hodgestar: Just done.
[08:16:14.700519]<Hodgestar>Woot.
[08:16:57.743835]<CIA-37>r841 by wbell in branches/0.6.x/bitten/ (10 files in 4 dirs): Port of [638], [639], [640] to 0.6.x
[08:17:22.223243]<wbell>Aargh. Build breakage.
[08:20:39.211572]<wbell>I think I just need to make the builds unique
[08:22:32.494180]<Hodgestar>Make the builds unique?
[08:23:52.279786]<wbell>Yeah-- the build ids in the master tests aren't unique.
[08:25:00.322855]<drnlm>I've filed 574 on the current status of the verbose flag, anyone keen on bikeshedding the correct fix?
[08:25:13.536346]<Hodgestar>wbell: Oh. So your tests re-use the existing database?
[08:25:45.637932]<wbell>Hodgestar: nope-- I may be wrong on that. Looking.
[08:28:18.520794]<wbell>Hodgestar: ahh, looks like a bad merge.
[08:28:30.835518]<Hodgestar>The Cape Town group will break for dinner in about 30 minutes (so around 16:00 GMT).
[08:28:36.478342]<Hodgestar>wbell: Eep.
[08:33:06.845679]<CIA-37>r842 by wbell in bitten/tests/master.py: Bad merge of patch from [839]
[08:33:33.121138]<wbell>Build should be better now.
[08:33:52.251658]<CIA-37>r843 by wbell in branches/0.6.x/bitten/tests/master.py: Bad merge of patch from [839]
[08:38:34.465108]<Hodgestar>Great. Test passes again locally (so now I can commit my tiny patch for #451).
[08:40:18.535665]<CIA-37>r844 by hodgestar in bitten/report/lint.py: Fix reference to report.category (Fixes #451. Patch from Roland Wilczek).
[08:41:47.266713]<Hodgestar>wbell: Might be worthwhile standardizing merge messages a bit (svn --log on the branch will show you the format osimons and I have been using).
[08:42:05.985075]<CIA-37>r845 by hodgestar in branches/ (0.6.x/bitten/report/lint.py 0.6.x): Merge of [844] from trunk.
[08:52:45.616842]<drnlm>Simple patch to 574 uploaded - Should I tweak the patch to 529 to behave similarly?
[08:56:54.956580]<Hodgestar>Let me go have a look.
[08:57:56.766875]<CIA-37>r846 by dfraser in bitten/ (17 files in 5 dirs): Reimplement charting using flot 0.6 - see #426
[08:57:56.938816]<Hodgestar>drnlm: Seems like a good idea.
[08:58:04.857317]<drnlm>updated patch to 529 added
[08:58:07.744507]<Hodgestar>dfraser: Hooray!
[08:58:23.337981]<Hodgestar>drnlm: Let me look at committing those in the 3 minutes before dinner. :)
[08:58:37.658308]<Hodgestar>Hopefull wbell's build slaves are still running.
[08:58:51.014737]<osimons>wow! well done all! just back, and look at all the stuff that has been done while i'm away?! i should be away more often.... :-)
[08:59:28.162594]<Hodgestar>osimons: Lol. Hopefully we didn't break anything.
[09:00:11.312429]<osimons>don't know yet - untested. see walter has merged without recoding, so i'll just clean that now. he just timed out on irc i see.
[09:00:30.530596]<Hodgestar>Merged without recoding?
[09:00:34.116564]<Hodgestar>Recording?
[09:00:58.009271]<osimons>svn records merge info when using 'merge' command
[09:01:18.016389]<Hodgestar>Ah. Maybe he's using svn 1.4?
[09:01:36.359143]<osimons>in 0.6 branch, then "svn mergeinfo --show-revs eligible ^/trunk" shows all his already done merges
[09:01:53.833757]<osimons>either that, or he merges outside svn by reapplying the patches.
[09:02:44.262022]<Hodgestar>osimons: That seems unusually perverse. :)
[09:03:02.379109]<Hodgestar>osimons: Although given that his diffs seem to come from git, maybe not that unlikely.
[09:03:20.890209]<osimons>Hodgestar: well, those that have used svn for some time has usually learned how to deal with its missing abilities in certain areas...
[09:03:35.708915]<osimons>- by using workflows outside svn tools...
[09:03:42.332395]<osimons>if he gets online when i'm not here, help me remember to tell him about svn 1.5 or higher, and the two commands for eligble + merging that he needs
[09:04:07.254998]<Hodgestar>osimons: Will try remember.
[09:04:38.482905]<osimons>good work though, i'm very impressed with this stunt and all the effort put in by everyone!
[09:05:09.380532]<techtonik>I may try import Bitten into Google Code HG for review.
[09:05:25.108184]*jerith is sad about #466.
[09:05:53.998086]<jerith>I was hoping to fix it, not decide that it's as fixed as it's likely to get.
[09:06:14.217887]<CIA-37>r847 by hodgestar in bitten/build/svntools.py: Make svn:checkout command's verbose flag behave more rationally (Fixes #574. Patch from Neil Muller).
[09:07:43.063689]<CIA-37>r848 by hodgestar in branches/ (0.6.x/bitten/build/svntools.py 0.6.x): Merge of [847] from trunk.
[09:12:04.559736]<CIA-37>r849 by osimons in branches/0.6.x: Record svn merge information for [830,832,834,836,838,839,840,842] that was merged in [831,833,835,837,841,843].
[09:13:25.371427]<techtonik>Stefan Sperling from Subversion asks if exit code of 1 when externals fetch fail will help us with #552
[09:14:27.999094]<Hodgestar>techtonik: Definitely.
[09:14:37.842364]<drnlm>techtonik: I saw that and have sent a reply - bitten checks the return code, so it should fix the problem
[09:19:01.864145]<Hodgestar>osimons: The WindowsError in svntools.py should probably be protected for use on Linux systems.
[09:19:25.598300]<CIA-37>r850 by hodgestar in bitten/build/svntools.py: Add a selection of authentication options to svn commands (--username, --password, --no-auth-cache). Fixes #529.
[09:21:30.421310]<osimons>Hodgestar: shouldn't really matter i think, seeing we're not importing just checking names?
[09:21:55.211625]<CIA-37>r851 by hodgestar in branches/ (0.6.x/bitten/build/svntools.py 0.6.x): Merge of [850] from trunk.
[09:22:09.355187]<osimons>ie. name of exception class? however, can't say for certain. feel free to make it explicit
[09:25:25.575231]<Hodgestar>WindowsError isn't defined at all on other platforms so if that clause is ever reached it'll generate a NameError.
[09:26:16.155053]<Hodgestar>Not a huge issue. I just saw it and thought I would make it less likely to be forgot.
[09:26:18.130761]<Hodgestar>Dinner time!
[09:26:46.973045]<Hodgestar>I'll be back later to wrap up lose ends and update the Milestone page.
[09:26:58.525504]<Hodgestar>Might be joined by a few die-hard sprinters. :)
[09:27:03.213928]<techtonik>I replied to Stefan too, because I am not subscribed and it would be nice to get notification if exit code will be added in next release. so we can reschedule #552 to 0.6.1 or later
[09:28:09.774214]<Hodgestar>If subversion are going to fix it I think we can just close #552 with a note saying to upgrade to the latest subversion.
[09:28:30.300375]<Hodgestar>But let's keep #552 open until we know for sure the change has made it into upstream.
[09:28:54.865878]<Hodgestar>Thanks for bugging the subversion people. :)
[09:30:01.528235]<techtonik>They didn't mind. =)
[09:30:26.800784]<Hodgestar>Now really off to dinner. :)
[10:17:42.936389]<CIA-37>r852 by dfraser in branches/ (18 files in 6 dirs): Merge of [846] from trunk.
[12:45:52.543165]<Hodgestar>Sprint over.
[12:46:00.455980]<Hodgestar>Thanks everyone for their help!
[12:46:52.838550]<Hodgestar>Builds still green. :)