IRC log of #zope3-dev for Saturday, 2010-04-24

J1mYou might be able to get buildout to use yours, but it feels error prone.00:00
* J1m would be slightly (and pleasently) surprised if Fredrick was actually taking an interest.00:01
aclarkJ1m: Right, just wanted to confirm that. I'm not sure what I'll do but at least I know I can't rely on buildout to do what I want. (IOW, it'd be nice if there were an equivalent in buildout to do easy_install, i.e. install this egg from this location)00:01
*** danielblackburn is now known as danielb_afk00:02
J1maclark, yeah, although it's not a use case I'd want to put a lot of effort into.00:03
aclarkJ1m: If I didn't have enough to do already, I might consider forking PIL ;-)00:03
aclarkJ1m: Understood, np00:03
J1mIt sounds like already forked at least the release. The fact that it's a fork should be reflected in the project name.00:05
*** benji has quit IRC00:05
J1mThis PILwoTK :)00:05
J1mThat was such an evil release, I'm actually proud of it. ;)00:05
aclarkJ1m: Perhaps a PILwoTk 1.1.7 is in order then ;-)00:05
*** hathawsh has quit IRC00:06
* aclark ponders PILwNS for "PIL with name spaces" :-)00:06
*** allisterb has quit IRC00:10
*** andreypopp has joined #zope3-dev00:11
*** allisterb has joined #zope3-dev00:24
*** andreypopp has quit IRC00:35
*** afd___ has joined #zope3-dev00:46
*** afd__ has quit IRC00:49
*** andreypopp has joined #zope3-dev00:53
*** andreypopp has quit IRC00:55
*** JaRoel|4D has joined #zope3-dev01:02
*** JaRoel|4D has joined #zope3-dev01:03
*** aaronv has quit IRC01:05
*** mcdonc has quit IRC01:08
*** tarek has joined #zope3-dev01:09
*** gary_poster has quit IRC01:16
*** J1m has quit IRC01:33
*** aaronv has joined #zope3-dev01:42
*** aaronv has quit IRC01:49
*** aaronv has joined #zope3-dev01:57
*** dunny has joined #zope3-dev02:01
*** runyaga has quit IRC02:13
*** aaronv has quit IRC02:14
*** tarek has quit IRC02:33
*** fcorrea has quit IRC02:35
*** mcdonc has joined #zope3-dev02:38
*** aaronv has joined #zope3-dev02:54
*** matthal has joined #zope3-dev03:02
*** ccomb has quit IRC03:16
*** jan_s has joined #zope3-dev03:31
*** aclark has left #zope3-dev03:33
*** r0ver has left #zope3-dev04:13
*** aaronv has quit IRC05:06
*** cbcunc has quit IRC05:12
*** romanofski has joined #zope3-dev05:43
*** redir has quit IRC05:48
*** huajie has joined #zope3-dev06:01
*** TresEquis has joined #zope3-dev06:35
*** kursor has joined #zope3-dev06:47
*** jfkw has quit IRC06:55
*** davisagli has joined #zope3-dev07:30
*** romanofski has quit IRC07:46
*** redir has joined #zope3-dev07:46
*** redir has quit IRC07:46
*** redir has joined #zope3-dev07:47
*** davisagli has quit IRC07:49
*** kursor has quit IRC08:23
*** afd_ has joined #zope3-dev08:53
*** afd___ has quit IRC08:57
*** romanofski has joined #zope3-dev08:59
*** agroszer has joined #zope3-dev09:26
*** matthal has quit IRC09:38
*** matthal has joined #zope3-dev09:45
*** matthal has quit IRC09:59
*** Theuni1 has joined #zope3-dev10:15
*** matthal has joined #zope3-dev10:19
*** JaRoel|4D has quit IRC10:21
*** matthal has quit IRC10:34
*** agroszer has quit IRC10:34
*** agroszer has joined #zope3-dev10:38
*** zagy has quit IRC10:44
*** andreypopp has joined #zope3-dev10:44
*** tarek has joined #zope3-dev10:51
agroszerhmmmm.... I get failures with KGS3.4.0, even with virtualenv10:57
*** andreypopp has quit IRC11:03
*** andreypopp has joined #zope3-dev11:04
*** redir has quit IRC11:06
*** matthal has joined #zope3-dev11:14
romanofskimoin Theuni111:16
* Theuni1 waves romanofski11:20
Theuni1romanofski: bugfixing? :)11:20
*** ccomb has joined #zope3-dev11:20
romanofskiTheuni1: ow... bugfix day today?11:21
romanofskihm.. do you have the URL to the list by hand?11:21
Theuni1here's the wiki page11:22
Theuni1it has a link to the bug list11:22
romanofskihm... "can't set XUL content-type"... *G11:23
Theuni1memories :)11:23
* romanofski nods11:24
romanofskiTheuni1: I'm happy to help...11:24
* romanofski is browsing11:24
*** sweh has joined #zope3-dev11:24
romanofskiTheuni1: it looks like there are already patches for the most?11:25
romanofskihowdy sweh11:25
Theuni1still need to review them, though.11:26
swehhi roman11:27
swehhow're u down under?11:28
*** andreypopp has quit IRC11:29
romanofskigood good... just got hooked by Theuni1 for the bug day11:29
romanofskihow about yourself?11:30
Theuni1sweh: so what about you? you can fix one tiny bug, too! :)11:30
Theuni1(and maybe another one, and then a tiny other one, and ...)11:30
swehhm, sure I can, but did you look out if the window?11:31
Theuni1sun, fresh air and the awful noise of birds are the computer scientists arch enemies11:31
swehby the way, my daily bug days start on monday and end on friday :P11:31
Theuni1i'm talking about fixing bugs, not putting them in, thank you. ;)11:32
swehlol :)11:32
romanofskiTheuni1: send your employees to the east - they would now have no excuse because it'll be already dark outside ;)11:32
* Theuni1 got a checkin for revision 11133311:33
romanofskisweh: sounds like spring finally kicks in?11:34
Theuni1it does :)11:34
swehroman: indeed11:34
swehTheuni: I've got a date with my family today, so unfortunately no time for fixing bugs11:35
* sweh was just looking for some news while having breakfast11:36
* ccomb up11:40
romanofskiTheuni1: what should I do if .... let's say I'd like to pick #159467 (zope.testrunner should only profile actual tests, not setup) - checkout, check if the patch is working and commit?11:40
Theuni1this sounds like it might have been fixed already, actually11:43
Theuni1at least the patch won't work anymore because the test runner got heavily restructured after that patch11:43
Theuni1you might want to check out zope.testing.testrunner and look at the pipeline i built there11:43
romanofskiyeah - does that already count as a bugfix?11:43
Theuni1IMHO the profiling plugin should be stopped/started before and after the test runs11:43
Theuni1if it does, then I'd say this was fixed already11:43
*** sweh has quit IRC11:44
romanofskiI probably have too old information - my pointer to the zope root is still: svn+ssh://
ccomb /me looking at #22761711:46
*** andreypopp has joined #zope3-dev11:47
Theuni1sounds good to me11:51
romanofskiif it's zope.testrunner then I've found the module11:51
Theuni1romanofski: the command should be 'svn co svn+ssh:// zpoe.testing"11:51
Theuni1ccomb: cool11:52
Theuni1romanofski: your user account seems fine11:53
Theuni1agroszer: is zope 3.4.1 getting mechanize 0.1.8 in there?11:53
romanofskiTheuni1: yes - I checked... can still do a checkout *phew11:54
agroszerTheuni1, nope, 0.1.7b or?11:54
Theuni1agroszer: that seems to break under python 2.511:54
Theuni1  File "/home/ctheune/.buildout/eggs/mechanize-0.1.7b-py2.6.egg/mechanize/", line 3, in <module>11:54
Theuni1    from urllib2 import \11:54
Theuni1ImportError: cannot import name GopherError11:54
agroszerI'll take a look -- once I get there11:56
agroszerthere are a bunch of other breakages11:56
Theuni1i'm trying to fix an apidoc bug11:56
Theuni1and i can't reproduce with 3.4.0 KGS11:56
*** baijum has joined #zope3-dev11:57
ccombfirst I'm importing the two files maintaining-software and releasing-software from philikon's sandbox into the developer doc11:57
Theuni1ccomb: there's another place11:58
Theuni1ccomb: there's the ZTK documentation where I mirrored those docs already11:58
Theuni1if you want to move them, you might wanna extract them from there11:58
ccombah ok11:58
Theuni1so they can live in the developer doc and ZTK docs would only mention them11:58
Theuni1(somebody was talking about an inter-sphinx installation which might help with those kinds of reference, but i don't know about that and i think it's not relevent yet)11:59
ccombbb needs also that for the "what's new" section11:59
agroszerTheuni1, I'm wondering a bit about mechanize, cause I'm testing the KGS with py2.5 and that exception does not come up12:00
* Theuni1 scratches his head12:00
Theuni1happened to me using zopeproject12:00
agroszerTheuni1, you're using py2.612:00
Theuni1that's python2.612:00
*** afd_ has quit IRC12:01
Theuni1my bad12:01
* Theuni1 scraps his virtualenv12:01
*** afd_ has joined #zope3-dev12:01
*** allisterb has quit IRC12:04
romanofski*narf - the tests for zope.testing take quite some time12:10
Theuni1yes they do12:17
*** allisterb has joined #zope3-dev12:17
* Theuni1 wades through the remaining bugs with patches in Zope 312:17
romanofskiTheuni1: I just tried to run a few tests (limiting with -t) and it appears to me that the profiler is started and stopped before the layer setup and after the tearDown of the layers12:19
romanofskiseems that the bug has been already fixed by the refactoring12:19
Theuni1that's too coarse if you look closely12:20
Theuni1it shouldn't be done on layer setup/teardown but on a test level12:20
Theuni1otherwise the individual tests' setup and teardown will be profiled as well12:20
Theuni1OTOH i don't know whether the test runner even has a look for that12:20
Theuni1try checking the API of the test runner plugins whether they have something on a per-test level12:21
*** huajie has quit IRC12:21
* romanofski tries to find the API for the plug-ins12:21
Theuni1well, OTOH the bug is talking about avoiding layer setu12:22
Theuni1so that seems to have been fixed12:22
Theuni1i'd say if you find a way to also avoid individual tests' setup/teardown try to do that, otherwise report the bug as fixed12:22
romanofskiah ok... found the whole kit and caboodle12:25
*** projekt01 has joined #zope3-dev12:33
ccombI'm converting most namechooser doctests into unittests, leaving just a short informative docstring12:33
*** JaRoel|4D has joined #zope3-dev12:33
lisppaste6agroszer pasted "z3c.layer" at
agroszerccomb, can you take a look at the above paste?12:34
*** Theuni1 has quit IRC12:34
agroszerwhich package might output that <h2>Unauthorized</h2>12:34
agroszerhallo projekt0112:36
projekt01agroszer, Hallo12:36
projekt01agroszer, the paste bin shows z3c.layer, that's wvery old stuff12:37
agroszeryah I know12:37
agroszerI want to get a KGS3.4.112:37
ccombmaybe the string comes from a po file?12:38
projekt01are you using both z3c.layer and z3c.layer.* packages in KGS?12:38
projekt01ccomb, or from the layout template12:38
projekt01probably it comes from the layout template showing that some parts are comming from layout and others from pagelet content12:39
*** astoon has joined #zope3-dev12:43
*** Theuni1 has joined #zope3-dev12:49
*** JaRoel|4D has quit IRC12:52
romanofskiTheuni1: looks like I was wrong - the profiler profiles the layer setup12:53
Theuni1meep :)12:53
romanofskiTheuni1: looks like hotshot can't be enabled, disabled more than once and create a cumulative result :(13:01
Theuni1romanofski: can you give me the bug URL again?13:02
Theuni1There was a patch attached and I wonder how that worked.13:02
romanofskibefore_tests and after_tests13:03
romanofskiearly_teardown is already used13:04
romanofskiwhich seems to be correct13:04
romanofskibut late_setup is before the layers are setup13:04
Theuni1now i'm getting confused13:04
Theuni1that means that we're already enabling/disabling the profile anyway13:05
*** baijum has quit IRC13:05
romanofskithe profiler is enabled before the layers are setup and stopped after all tests13:06
Theuni1so we're only calling it once13:07
romanofskiwhat do you mean by that?13:07
Theuni1i'm trying to figure out two things:13:07
Theuni1are we calling enable/disable once or multiple times?13:08
romanofskionce IMHO13:08
romanofskiI tried to do that multiple times and IMHO doesn't work13:08
romanofskicould be, that I did it wrong tho13:08
Theuni1Then I'd say ask the reporter of the bug about that.13:08
romanofskibut I don't think so13:08
Theuni1I don't think it's all that important WRT profiling, because profiling the setup code doesn't actually distort timing of the code under test anyway.13:09
Theuni1It just makes everything run slower.13:09
romanofskiI'll attach a comment13:09
Theuni1I've set the importance to low. Can you write the comment and set the status to incomplete?13:10
Theuni1took a while to figure out, though13:11
romanofskiyeh.. I had to find the lines where the testrunner sets everything up13:13
*** regebro has joined #zope3-dev13:17
*** JaRoel|4D has joined #zope3-dev13:18
*** matthal has quit IRC13:19
* Theuni1 ponders some food13:20
romanofskiTheuni1: either I don't have the permission or I can't see the form to set the bug as incomplete13:21
romanofskifood... good call - there is already food on the table: late dinner today13:22
Theuni1romanofski: try again13:22
Theuni1there's still issues with the group management and the bug tracking on LP13:22
Theuni1i've added you to the group you need, but we'll have to clean that situation up later13:22
romanofskistill can't find the button/link whatever13:23
Theuni1did you try reloading the page?13:23
Theuni1i also set you as the assignee for now13:23
romanofskiyeh - two times13:23
Theuni1screw it13:24
Theuni1i've set it to incomplete for ya13:24
*** ignas has quit IRC13:26
romanofskiif I've time after dinner/evening movie I'll squish another bug13:27
Theuni1i'll be around for another 2.5 hours if you need me13:27
Theuni1TresEquis should come in later to take over :)13:27
* Theuni1 creates LP projects like crazy13:27
romanofskiis it only today or does it include tomorrow as well?13:29
Theuni1only today13:30
Theuni1but nobody will stop you if you want to fix bugs tomorrow, too ;)13:31
romanofskilet's see - maybe there is time... depending on the weather13:31
* Theuni1 keeps forgetting his lunch :/13:32
*** JaRoel|4D has quit IRC13:43
*** aaronv has joined #zope3-dev13:47
Theuni1all bugs that were registered with zope 3 and had patches are either triaged or fixed now13:49
*** yota has joined #zope3-dev13:53
Theuni1now really some food and then some actual fixing.13:56
Theuni1I'll be back in 30 minutes in case someone needs me as a guide.13:56
ccombis it ok for the namechooser to return something even if we give it an int or None or anything?14:00
ccombas a name14:00
ccombI'm replacing14:01
ccombname = unicode(name.replace('/', '-').lstrip('+@'))14:01
ccombname = unicode(name).replace('/', '-').lstrip('+@')14:01
ccombso that it works with None as a name14:01
agroszergaaah, breaks z3c.layer tests14:12
Theuni1ccomb: what does the interface say?14:14
*** astoon has quit IRC14:14
ccombTheuni1: that chooseName should never raise an error14:14
ccombI'm adding many unittests14:15
ccombalso replacing '.' with os.path.extsep14:15
ccomband checking the type before the value in checkName14:15
*** JaRoel|4D has joined #zope3-dev14:17
Theuni1ccomb: i wouldn't do the extsep14:17
Theuni1we're talking about URIs which should be generated independently on where the server is running14:18
Theuni1otherwise semantics of your application may change when you move the server between OSes14:18
Theuni1which is exactly the opposite of cross platform ;)14:18
Theuni1i would be careful about the unicode thing14:19
Theuni1converting anything to unicode may raise an error14:19
Theuni1you might want to protect it like this:14:19
Theuni1    name = unicode(name)14:19
Theuni1except UnicodeDecodeError:14:19
Theuni1    name = ''14:19
Theuni1name = ...replacements here14:19
ccombok for unicode14:20
Theuni1the most important thing IMHO is that the assumption about chooseName never breaking holds14:20
Theuni1as the interface tells the name only *may* be taken into account14:20
Theuni1and we can try hard by converting ints etc. into strings, but we need to also be failsafe14:20
agroszera string can be converted by omitting bad characters14:21
agroszerbetter try that too first before setting name = ''14:21
Theuni1agroszer: fine with that14:21
agroszername = '' seems pretty rude, huh?14:22
ccombfor the extsep thing, it is only used to keep the extension in case the name already exists14:22
ccomb'foobar.rst' -> foobar-2.rst'14:22
ccombinstead of 'foobar.rst-2' which would break mimetype detection14:22
ccombso I thing detecting the separator is better with os.path.extsep14:23
Theuni1for that to work perfectly you would need to know the platform of the person who uploaded the file14:23
Theuni1not the platform of the server14:23
ccombmmmh right14:23
Theuni1and i think sticking to '.' will be fine :)14:23
agroszerwhat does the mac use?14:23
Theuni1i bet its '.' on almost all platforms, but the API doesn't guarantee that to you14:24
Theuni1the formally correct thing would be to use the clients notion of the separation but we can't have that14:24
Theuni1using the server's is definitely formally wrong. so i reaaaaaaaaally think '.' is the way to go. ;)14:25
ccombok I'm sticking with '.'14:25
Theuni1hooray :)14:25
* Theuni1 picks another bug14:26
*** TresEquis has quit IRC14:28
*** JaRoel|4D has quit IRC14:29
*** TresEquis has joined #zope3-dev14:32
*** redir has joined #zope3-dev14:38
Theuni1looks like i found myself a rat hole14:43
agroszerwhich one?14:45
agroszerouh that one14:46
agroszerI think I've seen even a patch for that14:47
agroszerbut the patch was for Clientform14:47
agroszerTheuni1, I think we need to enforce dumping egg versions for tags14:52
agroszermy problem is that I try to test a package on a tag14:53
agroszerand I have no idea whether to take the KGS3.4.0 or which pacakge versions14:53
agroszercause it just fails14:53
agroszerwith various problems14:53
Theuni1yeah, i think we should discuss that on the list or in the meeting14:54
Theuni1i've had a similar point earlier14:54
agroszerand if you want to release a next minor, you'll want your tests to pass14:54
*** alga has joined #zope3-dev14:55
agroszerand on top of that I need zc.buildout 1.3.1, gaaah14:56
*** JaRoel|4D has joined #zope3-dev14:59
*** JaRoel|4D has quit IRC15:04
Theuni1what are you up to? :)15:04
agroszerstill fighting with the KGS15:05
agroszerz3c.layer 0.2.4 breaks z3c.formdemo15:05
ccombagroszer: this is security fix that should break most applications. there is a warning about this already15:06
agroszeryes, now I'm trying to figure how to solve it in z3c.formdemo15:07
ccombprobably by adding more security registrations15:07
agroszerccomb, weird is:15:08
agroszer<grantAll role="zope.Anonymous" />15:08
agroszerso everyone ought to be a superuser15:09
Theuni1no everyone15:09
Theuni1only anonymous15:09
Theuni1doesn't apply to authenticated15:09
Theuni1there's a separate "everyone"15:09
agroszerand z3c.formdemo runs without auth15:09
agroszerviews are registered for IContainer15:10
agroszerand I get15:11
*** Charlie_X has joined #zope3-dev15:11
agroszerForbiddenAttribute: ('__getitem__', < object at 0xa88492c>)15:11
*** astoon has joined #zope3-dev15:11
* Charlie_X waves15:11
* Theuni1 stares at the list of packages zope.testbrowser requires15:11
* Charlie_X offers Theuni1 some eyedrops and aspirin15:11
* Charlie_X has just correctly built Python 2.7 from MacPorts.15:12
agroszerThe security policy defines a special role named "zope.Anonymous".  All15:13
agroszerprincipals have this role and the role cannot be taken away.15:13
agroszerstill does not apply to everyone?15:13
Theuni1i think there's separate definitions of roles and groups15:15
Theuni1I guess what I said applies to the groups not the groups.15:16
Theuni1not the roles15:16
Theuni1something is broken for me grrr15:16 doesn't sucessfully complete tests when run on python 2.415:16
*** JaRoel|4D has joined #zope3-dev15:17
Charlie_XI removed Python 2.4 from my system this week. It's broken on OS X 10.6 and since Zope 2.12 I haven't needed it for anything.15:18
*** alga has quit IRC15:18
* ccomb wonders whether a bugfix in 3.11 should be backported in 3.10, 3.9, 3.8, 3.7, 3.6, and 3.515:19
* ccomb feels the pain15:20
ccombdo we really need releasing so many major versions?15:20
Theuni1ccomb: i don't care too much about fine-grained backporting15:20
agroszerI guess if someone needs it he'll backport15:21
ccombI'll probably backport at lest in the KGS 3.4 branch15:21
agroszerthat would be great15:21
*** redir has quit IRC15:24
*** huajie has joined #zope3-dev15:32
*** TresEquis has quit IRC15:38
*** TresEquis has joined #zope3-dev15:39
ccombthe KGS 3.4.1 really needs buildout 1.3.115:40
ccombI cannot run tests on without it15:40
agroszerhmmmm, if a user does not have perm to __getitem__ ona folder, then he also cannot get a view on the folder??15:41
agroszerccomb, you're right15:41
Charlie_XI'm wondering how to go about applying, testing and publishing a patch for zope.structured I made a few years ago. Tres kindly reassigned the bug to me.15:41
TresEquisCharlie_X: check it out, boostrap the buildout, run bin/test --all15:42
TresEquisthen try applying the patch15:42
TresEquisand re-run tests15:42
agroszerI'll file a bug15:42
Charlie_XMorning TresEquis15:43
TresEquisI'm gonna be gone to a jazz rehearsal for a couple of hours shortly15:44
*** JaRoel|4D has quit IRC15:46
*** Charlie_X has quit IRC15:46
Theuni1TresEquis: if you have time later, i'd love if you could review my work on #98437 then15:47
*** Charlie_X has joined #zope3-dev15:48
TresEquisI'm going to push the mechanize 2.0 changes to a branch15:48
TresEquisin case they step on your stuff15:48
TresEquisCharlie_X: have you looked at the updated developer docs?15:56
Charlie_XVery briefly15:56
Charlie_XAll tests pass after my patch.15:57
Charlie_XMaking it conform to PEP 8 now16:00
TresEquisdid your patch add any new tests for the issue?16:00
agroszerccomb, you mind taking a look?16:03
agroszerI don't get this16:03'zope.View',self.context)16:03
agroszersais I have permission tothe root folder16:03
agroszerbut self.context[name] still fails16:03
agroszerwhere self.context is the root folder16:04
Charlie_XTresEquis: my patch effectively doubles the number of tests.16:05
ccombagroszer: is there a 3.4.1 versions.cfg somewhere?16:05
agroszernot yet16:06
TresEquisromanofski: I just made an LP project for zope.testrunner (Lennart's split out version)16:06
*** jfkw has joined #zope3-dev16:06
TresEquisand tagged you with lp:159467 on that project as well16:06
agroszerccomb, as soon as I got all tests pass16:07
romanofskiTresEquis: alright - I guess that's were any further patches should go in?16:08
ccombis the zope.release trunk updated?16:08
Theuni1TresEquis: i didn't finish the testbrowser part of #98437  but i changed not setting any referer at all, just looking for an OK from you.16:08
agroszer3.4.1 goes into branches/3.4, or?16:08
Theuni1for testbrowser i need to do some more16:08
Theuni1gotta run to catch up with my GF now16:09
ccombyes 3.4 branch, I think the trunk is unuseful now16:09
agroszer2 more failures and z3c.formdemo to go16:09
TresEquisTheuni1: I think it should be fine16:12
TresEquisCharlie_X: I would go ahead and commit when you think its ready16:13
*** andreypopp has quit IRC16:18
*** andreypopp has joined #zope3-dev16:19
*** andreypopp has quit IRC16:21
ccombagroszer: I have a 404 on mechanize with your branch16:35
agroszerlook into the addresbook sample16:35
agroszerthat should fail with ForbiddenAttribute16:36
agroszer, /z3c/formdemo/addressbook/README.txt", line 22, in README.txt16:36
ccombI mean during the buildout16:38
agroszerI guess I have it in the buildout cache16:39
agroszerseems like 0.1.7b was removed16:42
agroszerthat sucks16:42
agroszerccomb, I'll drop you the targz by mail16:43
ccombI have 0.1.8 in my cache16:44
*** allisterb has quit IRC16:44
ccomb8 failures with 0.1.816:45
agroszersame here with 0.1.7b16:46
agroszereach README.txt fails16:46
agroszermaybe I'll bump it in KGS 3.4.116:47
ccombsame with 0.1.7b, 8 failures16:48
ccombsince 0.1.8 is there, probably it should be used instead, yes16:51
agroszerit just fails16:52
agroszerTraceback (most recent call last):16:52
agroszer  File "/home/adi/.buildout/eggs/zope.testbrowser-3.4.2-py2.5.egg/zope/testbrowser/", line 22, in <module>16:52
agroszer    from zope.testbrowser import browser16:52
agroszer  File "/home/adi/.buildout/eggs/zope.testbrowser-3.4.2-py2.5.egg/zope/testbrowser/", line 22, in <module>16:52
agroszer    import mechanize16:52
agroszer  File "/home/adi/.buildout/eggs/mechanize-0.1.8-py2.5.egg/mechanize/", line 122, in <module>16:52
agroszerImportError: cannot import name Firefox3CookieJar16:52
agroszer(for me)16:52
agroszerI guess I'm missing sqlite316:54
ccombok I see the ForbiddenAttribute during traversal16:55
*** ignas has joined #zope3-dev16:56
*** allisterb has joined #zope3-dev17:01
*** fcorrea has joined #zope3-dev17:06
Charlie_XWhere are the translation files? I've made a tiny change for choice widgets and want to update the German translation.17:07
agroszerwhich package?17:07
agroszerI guess
*** JaRoel|4D has joined #zope3-dev17:11
agroszerccomb, mechanize 0.1.9 is the solution17:13
ccombsomething special ?17:14
agroszerjust a workaround for Firefox3CookieJar17:14
Charlie_XHow do I get the changed default ("no value" -> "nothing selected") into the locales?17:20
ccombfinished with zope[.app.].container17:23
ccombno new release yet17:23
agroszerhahh, that was to blame?17:24
agroszerCharlie_X, honsetly: no idea17:25
agroszerI guess i18n was not maintained since 3.4.0 or earlier17:25
agroszerprojekt01, you around?17:26
agroszercould you do a py2.5 win32
agroszercould you do a py2.5 win32 egg?17:27
agroszercould you do a py2.5 win32 3.4.3 egg?17:27
agroszerI get it right slowly...17:27
projekt01which version?17:27
projekt01agroszer, but that's a very old one?17:28
agroszerstill working on KGS3.4.117:29
agroszerit almost passes now on linux17:29
agroszernow I'm getting to windowze17:29
projekt01agroszer, Ok done17:30
projekt01let me know if you need more17:30
*** alga has joined #zope3-dev17:34
*** lisppaste6 has quit IRC17:34
*** alga has quit IRC17:37
*** dunny has quit IRC17:39
*** hazmat has quit IRC17:42
*** huajie has quit IRC17:42
*** lisppaste6 has joined #zope3-dev17:51
*** baijum has joined #zope3-dev17:53
*** baijum has quit IRC17:59
*** JaRoel|4D has quit IRC18:03
ccombagroszer: if I just add a security declaration for, I fall down to 3 failures18:06
ccombdoes the trunk have the same problem?18:07
agroszergood q18:07
agroszerlet me check18:07
ccombI'm commiting to your bhancr18:09
agroszergaaah trunk is even worse18:10
agroszerIOError: [Errno 2] No such file or directory: '/home/adi/.buildout/eggs/'18:11
ccombok now:18:11
ccombForbiddenAttribute: ('__annotations__', < object at 0x4d46c08>)18:11
ccombalsmost the same18:12
agroszerI don't get this, if the user has all perms, why is this coming?18:12
*** JaRoel|4D has joined #zope3-dev18:13
Charlie_XI don't know anything about ForbiddenAttribute but it sounds like the sort of thing that "not even god should change".18:13
*** bitmonk1 has joined #zope3-dev18:13
agroszerit's a permission problem, but the user ought to have all permissions :-O18:14
ccombEven if you have all the permission, a missing security declaration is *always* preventing from accessing the content18:14
ccombIf I remember18:14
* bitmonk1 couldn't add himself to the bugday in wiki, but is interested in helping with bb/ztk. anyone looking at 98395?18:15
*** JaRoel|4D has quit IRC18:25
agroszerbitmonk1, assign to yourself18:29
*** alga has joined #zope3-dev18:29
bitmonk1i don't see the option, checking again ..18:30
bitmonk1got it.18:30
Charlie_XLaunchpad's icons are not very intuitive18:30
agroszer:-D you need lots of karma18:30
bitmonk1Charlie_X: an old coworker of mine is listed as being on the UI team, something tells me this particular bit of UI is him. :)18:30
* bitmonk1 hides from the edwin18:31
Charlie_XIt's miles better than bugzilla but I had the same problem earlier.18:31
bitmonk1oh yeah bugzilla pffth.. once you know it, this sort of ui is very concise, but it just doesn't reach out to you the first time around.  now i know to click the diagonal exclamation point balloon next to 'unassigned' for an assignee popup.18:32
agroszerthat's a pencil... ;-)18:33
agroszerisn't it?18:33
*** alga has quit IRC18:33
*** astoon has quit IRC18:35
Charlie_XI suppose it could be a pencil. But the yellow circle certainly makes an exclamation mark the most likely interpretation.18:35
Charlie_XI've always assumed it was a warning symbol.18:35
bitmonk1pencil would explain the lean, as if being held by an obviously superior right-handed person [ *ducks* ]18:37
ccombannotation adapters are not trusted?18:38
*** Charlie_X has quit IRC18:39
bitmonk1i would expect any adapter to be trusted .. are you getting an error that implies your ask?18:39
ccombI'm still trying to make z3c.formdemo work18:40
*** Charlie_X has joined #zope3-dev18:40
Charlie_XOpera just went boom18:40
* bitmonk1 just joined in the past hour18:41
ccombI don't like being forced to add a <require attributes="__annotations__"18:41
agroszerthat's ugly18:42
agroszersomething is wrong there18:42
Theuni1TresEquis: there's a project group now18:45
Theuni1in launchpad. i've put all projects that i found into it18:45
*** Theuni1 has quit IRC18:51
*** ignas has quit IRC18:59
bitmonk1ccomb: talk it through, why do you have to do this?18:59
Charlie_XTresEquis: I committed both changes for the two bugs you tagged for me. I don't understand much of the other CMF bugs so I am continuing to work on Yuppie's CookieCrumbler branch.19:01
ccombno I think it's just another missing declaration on an annotation adapter19:03
bitmonk1ok :)19:03
ccombtell me if I'm wrong, when using an annotation adapter, either you declare it as trusted, or you declare security for its interface19:03
bitmonk1that is not a question i know the answer to, or obvious to me from a glance at annotation docs19:10
*** redir has joined #zope3-dev19:10
*** astoon has joined #zope3-dev19:13
*** projekt01 has quit IRC19:13
TresEquisok, back from rehearsal19:22
TresEquisanybody still squishing crustaceans here?19:22
TresEquisarthropods, anyway19:22
* bitmonk1 stomp19:24
bitmonk1in fact .. i think this patch is ready.19:24
bitmonk1well, i'm unclear if the test needs to be further improved.. but still squishing away. :)19:26
ccombjust 1 more failure on formdemo19:30
bitmonk1yah i need to test if this output is actually unicode, rather than just checking the content-type header's charset.  hm.19:31
TresEquisccomb: re
TresEquisdo you actually see the forbidden error propagating?19:35
TresEquisoops, that was agroszer, not ccomb19:35
ccombTresEquis agroszer the view is asked without @@19:37
ccombI think if we directly ask for the view using @@, there is no permission denied19:38
ccombthe traverser first tries to getitem and then falls back to the view19:38
TresEquisah, ok19:39
TresEquisand this is with spacesuits, right?19:39
TresEquisaka security proxies19:39
TresEquiswhich make __getitem__ raise the forbidden exception?19:39
ccombI've tried replacing all the views with @@ in the template and it worked19:40
TresEquisI think a new unit test, with a dummy context whose __getitem__ raises the error, would be enough to add the fix19:40
TresEquisbut this is a case where there *is* an item with the same name19:40
TresEquisas the view19:40
TresEquisor is __getitem__ raising the error unconditionally?19:41
TresEquis(which it probably shouldn't)19:41
ccombyou want the traverser to try the view even if it failed on getitem due to security?19:41
TresEquisthat is what the bug report suggests19:41
bitmonk1nope, the result itself is not unicode because it is the whole http response including the headers.  dur.19:41
TresEquisI think it is generally true that templates should use 'context/@@view' to form URLs19:42
TresEquiswhen they know they want a view19:43
TresEquisI don't think that falling back to a same-name view is the right thing to do if there is an item with the same name, but the user isn't allowed to see it19:43
ccombthe last error on formdemo is an UnpickleableError: Cannot pickle <type ''> objects19:44
TresEquisthe odd case is if __getitem__ is raising an exception without checking for the item19:44
bitmonk1what is the $Id .. etc .. syntax in the docstrings?  I19:47
bitmonk1I'm familiar with it, but there's no googling $Id19:47
* bitmonk1 ends up reading about freud19:48
*** fcorrea has quit IRC19:49
bitmonk1anyway 98395 has a patch with tests that first failed, and now pass.  going to throw together some breakfast and pick up another / field feedback. :)19:49
*** agroszer has quit IRC19:57
*** agroszer_ has joined #zope3-dev19:57
*** aaronv has quit IRC20:13
*** bitmonk has joined #zope3-dev20:14
* Charlie_X is eating liquorice - does that count?20:15
TresEquisbitmonk1 it identifies the VCS version of the file20:19
TresEquisgenerally cruft20:19
Charlie_Xeek! disposing of auto_login_page causes an error in GenericSetup's utils! :-(20:21
Charlie_XBut only when CMFDefault's tests run...20:22
Charlie_Xah, it's a profile issue20:23
bitmonkTresEquis: sure i just, didn't know if i need to muck with it20:24
bitmonklooks like my patch needs a couple touches on review..20:24
TresEquisRemoving it would be "janitorial"20:25
TresEquisshould normally not be done in same commit as "real" fixes20:25
TresEquisand might need to be deferred until all "old" patches have been reviewed / applied, like other janitorial cleanups20:25
TresEquisbecause they can cause spurious patch failures20:25
*** fcorrea has joined #zope3-dev20:27
bitmonkTresEquis: absolutely.20:28
bitmonkmore like, i copied a file to a new one, to create a new test.  apparently it's a non-issue.20:29
TresEquisin a new file, just zap it20:29
TresEquisfossil from CVS days, actually20:29
ccombagroszer_: formdemo is ok20:34
ccombagroszer_: I'm renaming adamg-1.5.4 to 1.520:39
bitmonkhokay i feel pretty comfortable with the patch for 98395, what status should i set this issue to.  fix committed wouldn't be appropriate since i've only attached a patch and am not a committer.20:44
Charlie_XNo, In progress makes more sense.20:46
TresEquisbitmonk you can either upload the patch as an attachment, or push a bzr branch to launchpad20:54
TresEquisand ask me to review20:54
bitmonksure the patch is attached, but .. bzr is fine.  my patch is against svn trunk atm.20:55
TresEquisbzr is just an option20:55
TresEquisfor getting your branch uploaded with maybe multiple discrete changes20:55
TresEquisbut a patch is fine, too20:56
bitmonksure.. i would actually prefer it, i didn't know there was vcs-import of zope stuff.  sweet.20:56
TresEquiswhich package?20:56
TresEquis $ bzr co lp:zope.publisher gets you the trunk20:56
bitmonki am looking at the branches page now20:56
TresEquis$ bzr branch lp:zope.publisher20:57
TresEquisI mean20:57
bitmonkgot it.20:57
TresEquisWe can't push back to the branch, so checkout doesn't make sense20:57
TresEquisthen you can do:20:57
TresEquis$ patch < your.patch20:58
TresEquis$ bzr commit --fixes lp:9839520:58
TresEquis$ bzr push lp:~bitmonk/zope.publisher/lp_9839520:58
TresEquisoutlined on the shiny new "contributing as a non-committer using Bzr page"20:59
*** astoon has quit IRC21:00
* bitmonk is actually justizin on launchpad because .. bitmonk is linked to email or rackspace or who knows what of the past ;d21:01
bitmonkbut i get the point.21:01
bitmonkdone and ready for review TresEquis21:04
*** jfkw has quit IRC21:08
TresEquisOK, looks pretty good21:12
TresEquisone nit is our 80 column fetish21:12
TresEquisbut the extra tests look to exercise the new functionality fully21:12
* bitmonk never knows what to do about 80 columns when others are involved ;d21:13
TresEquisThere's docs for that ;)
TresEquisI tried to get most of our current practice into those docs21:14
TresEquisthe bzr stuff is pretty much speculative:  you're the first actual non-committer to try them out21:15
TresEquisfor which I'm grateful ;)21:15
TresEquisI can fix those before pushing to SVN21:15
TresEquisah, another fix:  needs to update CHANGES.txt21:15
bitmonkThe docstring should contain a reference about version control status. The example given is valid for at least CVS and Subversion. <- the example is "$Id$", this is the thing i ripped out.  just noting.21:16
bitmonki'm happy to do both.21:16
bitmonkon that is21:16
TresEquishmm, I don't see that in the patch21:16
bitmonksorry, segway21:16
TresEquisah, ok21:16
TresEquischanges to those docs need to be run by the ZTK steering group (of which I'm a member)21:17
TresEquisbut I would be in favor of dropping the $Id$ fossil21:17
bitmonkfair enough, just pointing out since we discussed it so you'd be aware.21:21
*** astoon has joined #zope3-dev21:21
* bitmonk observe what appears to be an 86 column fetish :-P21:25
bitmonkpushed TresEquis with CHANGES.txt entries and multilined the error message to fit in 80 columns.21:29
bitmonker, entry.21:29
bitmonkalso happy to help try out the bzr non-committer access.  super duper handy.21:30
*** shen-long has joined #zope3-dev21:30
TresEquisah, I jumped the gone and tidied the long lines -- looks like there were some pre-existing fouls, too21:32
TresEquisyour branch is now merged to the SVN trunk21:32
bitmonksweet.  picking through what remain in lp.21:32
TresEquis209550 is another zope.publisher one with a patch21:35
* bitmonk takes a look21:37
bitmonkah i was in fact just looking at this..21:38
*** fcorrea has quit IRC21:41
*** jfkw has joined #zope3-dev21:42
bitmonkalright well that patch doesn't apply cleanly so.. i'll work on that. :)21:46
TresEquissounds good21:51
TresEquisBIAB -- daddy taxi duties call22:02
bitmonkgoing to assign myself, suck down to the netbook, and travel with it for a couple hours. :)22:02
bitmonkpretty straightforward just, picking through stuff by hand.  the tests are what don't quite cleanly apply.22:03
*** matthal has joined #zope3-dev22:06
*** bitmonk1 has quit IRC22:21
*** dunny has joined #zope3-dev22:30
bitmonkalright i've actually got 209550 ready TresEquis, but the push to lp is hanging ..22:32
bitmonkthar she blows.. updating the ticket with a note it is ready for your review after taxi duties, TresEquis.22:37
TresEquisthanks, taxi back in the statnd22:42
*** ignas has joined #zope3-dev22:43
*** agroszer_ has quit IRC22:45
Charlie_XRight, I've added a logout view to Yuppie's branch. Still need to sort out the tests but things are looking good.23:01
*** andreypopp has joined #zope3-dev23:04
Charlie_Xbitmonk: 80 characters is there for several reasons but the most cogent is your ability to read a lot of characters per line is limited. And line breaks don't hurt23:04
*** matthal has quit IRC23:09
*** allisterb has quit IRC23:09
TresEquisCharlie_X: more important is *other* people's ability to read the lines, particularly in the popular editors23:11
*** matthal has joined #zope3-dev23:11
andreypoppHello, what is the status of ?23:14
*** alga has joined #zope3-dev23:17
TresEquisandreypopp: nobody has tackled it yet23:18
andreypoppTresEquis: ok, it has patch with tests23:18
TresEquisandreypopp: are you a Zope SVN committer?23:19
*** JaRoel|4D has joined #zope3-dev23:21
andreypoppthe possible problem with this issue is it can break b/w23:22
andreypoppi don't really know23:22
TresEquisyou have two attachments to the issue:  does the second (the patch) include the first?23:22
andreypoppsthe first one is just test to show the bug, the second is fox with unit test23:23
andreypoppignore the first one23:23
andreypoppfox = fix23:23
*** afd_ has quit IRC23:24
*** allisterb has joined #zope3-dev23:25
TresEquisandreypopp: why does the fix not just propagate the Invalid exception?23:30
TresEquisinstead of wrapping it as a WrongContainedType exception?23:30
andreypoppis it not the policy for Object validation error to be wrapped in WrongContainedType?23:31
andreypoppThis is how code that collects errors at top level can differentiate errors from fields23:32
TresEquisI may misunderstand23:33
TresEquisthere is test code which wants WrongContainedType to be raised23:33
TresEquisthe 'getErrors' bit23:33
*** ignas has quit IRC23:34
TresEquisbut other fields have _validate methods which raise different errors23:35
andreypoppthis is because I do not propagate Invalie exc and wrap it in WrongContainedType23:35
andreypoppOk, for example we use zope.schema.getValidationErrors, and have failed invariant check something deeper inside Object field… we cannot differentiate it by field23:36
andreypoppbut if we wrap it inside WrongContainedType — we can23:37
andreypoppI think it is reasonable23:37
andreypoppTresEquis: sorry, I have to go, please update bug ticket if you reject or commit fix, ok?23:40
*** andreypopp has quit IRC23:41
TresEquisHere is the set of tagged bugs resolved today in the ZTK:

Generated by 2.15.1 by Marius Gedminas - find it at!