Team Chat Logs

April 28, 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    
<< <   > >>

[01:20:36.976897]<Hodgestar>osimons: I think #411 can be pushed back to 0.7 or 0.6.1 (or at least I got the impression from wbell that that was what the two of you decided about the keepalives).
[01:22:21.535096]<osimons>Hi Hodgestar!
[01:22:57.171403]<osimons>about #411, i haven't looked at it in details yet.
[01:24:06.944080]<osimons>however, i think that 0.6.x should be just "harmless" improvements to tools++, and that anything requiring changes to db, master-slave protocol should really stay unchanged in 0.6 branch => also incentive to get 0.7 better and sooner...
[01:25:19.482838]<Hodgestar>Fine by me.
[01:25:47.063412]<osimons>so either we get that in by 0.6 or we just release it for 0.7
[01:26:28.404686]<Hodgestar>I also don't mind if it goes into 0.6.
[01:26:34.073389]<osimons>Hodgestar: trying a bit to avoid the looong branch cycles of Trac where too many things happen in stable branches
[01:26:46.236250]<Hodgestar>Agreed.
[01:27:11.763973]<Hodgestar>Once we have 0.6 then 0.6.x should be mainly bugfixes and minor enhancements.
[01:27:21.306331]<osimons>same with me - 0.6 if possible, 0.7 if we can't make it before release. i'll make a note on the ticket
[01:27:22.143048]<osimons>yes
[01:27:28.301495]<osimons>(in an ideal world....)
[01:28:49.939684]<Hodgestar>So I guess this means #411 stays as a 0.6 ticket for now.
[01:38:07.040483]<davidfraser>osimons: Did you see my comments on #536 - it seems to me that the slowness is from trying to get version history on a node in mercurial - even though we're really just wanting the latest revision for the node
[01:41:19.460468]<osimons>davidfraser: yeah, saw that yesterday. i'll add my 2c.
[01:41:35.700917]<davidfraser>osimons: Thanks
[01:51:12.012415]<Hodgestar>davidfraser: My plan was to just copy the JSON method from Trac 0.12 into bitten/jsonutils.py or something.
[01:51:40.177424]<Hodgestar>The Trac code is only 15 lines or so (I didn't look at how big javascript_escape is).
[01:53:15.789592]<davidfraser>Hodgestar: Yes, just doing that now
[01:53:25.422183]<Hodgestar>davidfraser: Woot.
[01:53:59.978401]<davidfraser>javascript_escape is in Trac 0.11 (but not Trac 0.10), although 0.12 has a version with better faster escaping - I think it's fine to rely on the trac version
[02:01:41.859350]<osimons>davidfraser: generally we should do whenever possible - not a big fan of duplicating code that then strays out of sync...
[02:03:03.481826]<osimons>davidfraser: also, for the json we make it a 'compat' style module so that we try to reuse the trac one directly and if not define the one we copied. and, mark it as compat to be removed once minimum Trac version is 0.12 (which it likely will be once we start working on multirepos stuff in trunk)
[02:03:04.947819]<davidfraser>osimons/Hodgestar: Hmmm, but javscript_escape is only from Trac 0.11.3 - one option would be to default to the Trac versions rather than trying to import the python json method ourselves
[02:03:15.759285]<davidfraser>osimons: You beat me by one second
[02:03:19.223536]<osimons>hehe
[02:05:12.497214]<osimons>good stuff - can do the same compat handling with javascript_escape. typically make a bitten.util.compat module where we keep these things that have limited timespans
[02:05:16.369700]<Hodgestar>I suggest "import json" > "import simplejson" > "import trac function" > "our independent copy of Tracs dumps + javascript_escape + dependencies" as the fallback path.
[02:06:44.367257]<osimons>in what module is the trac json code?
[02:07:02.006947]*osimons should really know this if he'd been paying attention to trac development...
[02:07:49.917279]<osimons>found it
[02:07:58.237959]<osimons>trac.util.presentation
[02:09:56.599892]<osimons>well, the trac code tries to import json first too - can't we just keep it simple and ditch simplejson? import from trac > our copy of json dumps (from trac trunk)?
[02:11:36.154224]<davidfraser>I've got a clean version I'm going to commit now that should handle all cases
[02:12:17.069822]<osimons>davidfraser: hang on
[02:12:22.966676]<davidfraser>Hodgestar/osimons: It's currently: import trac function > import json > our independent copy of Trac + javascript_escape etc - any problem with that?
[02:12:38.677347]<osimons>nah fine i guess.
[02:12:59.031645]<osimons>davidfraser: should we make a bitten.util.json instead?
[02:13:23.385242]<davidfraser>osimons: I've called it bitten.util.webio in case we need any related stuff - do you prefer bitten.util.json?
[02:13:43.387757]<davidfraser>I've copied in the tests from trac
[02:14:48.078359]<osimons>working with the XmlRpc plugin, i added JSON support there last year. one of the nice things is that we may want to extend the encoder to make it support other classes and so on - like pass it a build and it encodes it correctly (example)
[02:15:32.548253]<davidfraser>Right, I'll call it json instead of webio etc
[02:15:35.224432]<osimons>that way we may keep our main code simpler, and defer encoding/decoding to own modules. like we sort of do with xmlio today
[02:55:49.030019]<Hodgestar>davidfraser: I think Trac does "import json > import simplejson > custom stuff" itself? (If I'm remembering the code correctly).
[02:57:05.340929]<davidfraser>Hodgestar: Yes it does. I've committed the change - what happened to our CIA bot?
[02:57:32.735979]*Hodgestar pings CIA-37.
[04:07:06.188903]*osimons was unaware that he had been given the power to finally restore law & order in this channel...
[04:49:13.955820]<osimons>davidfraser: broken builds...
[04:52:33.993212]*davidfraser looks
[04:54:07.998063]<Hodgestar>Looks like a missing re import.
[04:57:45.463342]*davidfraser fixes
[04:58:12.866986]<Hodgestar>I really hate svn externals. :)
[05:00:15.654640]<davidfraser>I've actually run the unit tests now with Trac 0.11.1 and Python 2.5, so hopefully this time will pass...
[05:14:43.773715]<davidfraser>Do the builds still only run once a day?
[05:19:50.436570]<osimons>davidfraser: they build whenever walter has them online and operational
[05:22:55.902693]<davidfraser>osimons: Right-ho
[05:55:47.639795]<Hodgestar>Hooray! Builds green again.
[07:08:33.659252]<CIA-37>r855 by hodgestar in bitten/build/ (javatools.py tests/javatools.py): Handle py.test skipped statuses correctly. Fixes #564.
[07:08:33.701099]<CIA-37>r856 by hodgestar in branches/ (3 files in 3 dirs): Merge of [855] from trunk.
[07:09:33.588297]<CIA-37>r857 by dfraser in bitten/ (4 files in 3 dirs): Copy trac's `trac.util.presentation.to_json` method to prevent `json` module requirement, and incorporate fallback methods to support Trac 0.11 etc. ...
[07:09:33.601357]<CIA-37>r858 by dfraser in branches/ (5 files in 4 dirs): Backported [857] to 0.6.x
[07:09:33.614267]<CIA-37>r859 by dfraser in bitten/util/json.py: Added re import (required for worst-case fallback) - see #426
[07:09:34.270155]<CIA-37>r860 by dfraser in branches/ (0.6.x/bitten/util/json.py 0.6.x): Backported [859] from trunk to 0.6.x
[07:10:54.116932]<Hodgestar>wb CIA-37.
[08:11:21.743139]<CIA-37>r861 by dfraser in bitten/report/ (testing.py tests/testing.py): Fix incorrectly set template from [846] - see #426
[08:13:18.261555]<CIA-37>r862 by dfraser in branches/ (3 files in 3 dirs): Backported [861] from trunk: Fix incorrectly set template from [846] - see #426