IRC log of #zope3-dev for Thursday, 2006-08-17

*** philiKON has joined #zope3-dev00:09
*** RaFromLinuxWorld has quit IRC00:10
*** efrerich has quit IRC00:10
*** projekt01 has quit IRC00:11
*** schwendinger has joined #zope3-dev00:14
*** whit has quit IRC00:55
*** J1m has quit IRC01:05
*** whit has joined #zope3-dev01:12
*** projekt01 has joined #zope3-dev01:17
*** xyld has joined #zope3-dev01:32
*** schwendinger has quit IRC01:41
*** schwendinger has joined #zope3-dev01:44
*** benji has quit IRC01:50
*** schwendinger has quit IRC01:56
*** projekt01 has left #zope3-dev01:56
*** dobee has quit IRC01:57
*** srichter has quit IRC01:58
*** yota has quit IRC02:06
febbhi all02:21
*** flox is now known as flox|away02:42
*** RaFromBRC has joined #zope3-dev02:54
*** RaFromBRC has quit IRC02:55
*** febb has quit IRC02:58
*** hazmat has joined #zope3-dev03:11
*** ChanServ sets mode: +o hazmat03:11
*** srichter has joined #zope3-dev03:19
*** ChanServ sets mode: +o srichter03:19
*** philiKON has quit IRC03:28
*** rocky|away is now known as rocky|Zzz03:29
*** whit has quit IRC03:35
*** benji has joined #zope3-dev03:42
*** nathany has quit IRC03:42
*** roym has left #zope3-dev04:16
*** hazmat has quit IRC04:23
*** niemeyer has quit IRC04:29
xyldtrying to track down a problem with traversal and Five ... anyone know where requests for things starting with __ get filtered out?04:52
*** roym has joined #zope3-dev04:56
*** hazmat has joined #zope3-dev06:18
*** ChanServ sets mode: +o hazmat06:18
*** baijum has joined #zope3-dev06:21
*** bradb has quit IRC06:24
*** philiKON has joined #zope3-dev06:32
*** philiKON changes topic to "logs available at || paste code examples into or || bug day TODAY, live at"06:33
*** stub has joined #zope3-dev06:38
*** baijum has quit IRC06:41
*** whit has joined #zope3-dev06:42
*** baijum has joined #zope3-dev06:53
baijumhi all06:54
*** whit has joined #zope3-dev07:06
*** baijum has quit IRC07:15
*** baijum has joined #zope3-dev07:25
*** roym has quit IRC07:29
*** baijum has quit IRC07:37
*** alecm has quit IRC07:37
*** baijum has joined #zope3-dev07:41
baijumcan anyone help me in ?07:41
baijummake a test release as said here:
*** srichter has quit IRC07:56
*** alecm has joined #zope3-dev08:00
*** eins has joined #zope3-dev08:00
*** baijum has quit IRC08:03
*** xyld has quit IRC08:14
*** alecm has quit IRC08:31
*** jukart has joined #zope3-dev08:32
*** philiKON has quit IRC08:34
*** baijum has joined #zope3-dev08:35
*** srichter has joined #zope3-dev08:42
*** srichter has quit IRC08:42
*** srichter has joined #zope3-dev08:44
*** baijum has quit IRC08:56
*** zagy has joined #zope3-dev08:59
*** baijum has joined #zope3-dev09:01
*** hdima has joined #zope3-dev09:09
*** flox|away has quit IRC09:14
*** stub has quit IRC09:22
*** schwendinger has joined #zope3-dev09:25
*** baijum has quit IRC09:27
*** jinty has joined #zope3-dev09:31
*** baijum has joined #zope3-dev09:36
*** romanofski has joined #zope3-dev09:40
baijumthere is syntax error from
baijumcan anyone verify this is ok:
baijumI am not sure about it, because there is no test suite09:44
baijumerlier it was : base_dir = os.path.dirname(zope.__file__)[:-4]09:44
baijumbut now there is two 'dirname' ?09:45
*** batlogg has quit IRC09:47
* baijum understands syntax error will only get when installing unless there is not test cases?!09:47
*** dobee has joined #zope3-dev09:48
*** yota has joined #zope3-dev09:58
*** flox has joined #zope3-dev10:16
*** flox is now known as flox|away10:16
*** ChanServ sets mode: +o srichter10:38
*** tlotze has joined #zope3-dev10:39
tlotzeHow can I take the external_edit tab away from an object whose class is a subclass of
*** MJ has quit IRC10:40
*** kobold has joined #zope3-dev10:41
baijumsrichter, ayt?10:44
*** baijum has quit IRC11:07
*** baijum has joined #zope3-dev11:08
baijumsritcher, I was going to ask about this : (it's committed now)11:09
baijumthere was not test cases for that module, so at first I confused about the syntax error.11:10
*** baiju1 has joined #zope3-dev11:17
*** baiju1 is now known as baijum__11:20
*** MJ has joined #zope3-dev11:22
*** batlogg has joined #zope3-dev11:23
*** baijum has quit IRC11:29
*** baijum__ is now known as baijum11:29
baijumsritcher, can you help me to fix this,
*** projekt01 has joined #zope3-dev11:35
d2mwhat happened to apidoc ? http://localhost:8080/++apidoc++ throws an "NotFound: Object: < object at 0xb6ee2dec>, name: u'++apidoc++'" in the trunk11:36
baijumd2m, it's not working in 3.3 branch too here11:43
d2mbaijum: yes, i just did a svn up (last was 2 days ago) and its broken now11:43
*** batlogg has quit IRC11:53
baijumd2m, no change in `` other than in last few days (I reverted it in my copy, but still not working), so problem lies somewhere else11:53
d2mbaijum: seems some goofie changed default devmode to off (from on before)11:57
d2mbaijum: this is hell with a customized zope.conf11:57
d2mi wonder why i12:02
d2mneed to consume collector issues to get vital information on configuration changes :(12:03
*** ignas has joined #zope3-dev12:11
baijumd2m: :)12:13
*** alga has joined #zope3-dev12:16
*** baijum has quit IRC12:19
*** regebro has joined #zope3-dev12:31
*** efrerich has joined #zope3-dev12:36
*** batlogg has joined #zope3-dev12:40
*** baijum has joined #zope3-dev12:41
*** philiKON has joined #zope3-dev12:42
philiKONhi fellas. let's fix some bugs12:43
baijumhi philiKON12:43
baijumlet me commit this small doc fix (not critical :)
philiKONi think it's called developer-mode12:44
*** baijum has quit IRC12:44
*** baijum has joined #zope3-dev12:55
baijumphiliKON, can you verify the patch given here
philiKONlooks *very* cryptic to me13:00
philiKONbaijum, btw, devmode? i think it's called developer-mode in zope.conf13:00
philiKONhmm, no you're right13:00
*** projekt01 has quit IRC13:02
*** faassen has joined #zope3-dev13:03
regebroHi philiKON, Hi faassen.13:03
regebroMeh, just died. :(13:04
philiKONare you serious?13:04
philiKONbaijum, what's wrong with the original patch of issue 673?13:05
*** gintas has joined #zope3-dev13:05
*** gintas has quit IRC13:06
*** gintas has joined #zope3-dev13:06
philiKONlooks like it's back up again13:10
*** Aiste has joined #zope3-dev13:12
*** dunny has quit IRC13:14
*** baiju1 has joined #zope3-dev13:15
*** baiju1 is now known as baijum_13:15
*** baijum has quit IRC13:17
baijum_philiKON, original patch by thorsig may be enough for issue 673, I was just playing around...13:18
philiKONbaijum_, fix it :)13:19
baijum_philiKON, you mean the new one is ok? verified it?13:21
philiKONnot sure how much my opinion counts, but i dislike it :(13:22
philiKONthe arithmetic isn't that hard. your patch makes it seem pretty complex. i didn't even pretend to understand it13:23
*** baijum has joined #zope3-dev13:24
philiKONbaijum, something wrong with your inet connection?13:24
*** batlogg has quit IRC13:24
baijumphiliKON: yes13:24
philiKONbaijum, btw, your patch regarding #507. you were right and i was wrong13:24
philiKONit's devmode, not developer-mode13:25
philiKONi think the zcml feature is called developer-mode13:25
philiKONvery confusing13:25
baijumok, will commit a new one now13:25
baijumphiliKON, regarding 673, you mean the new patch is ok? verified it?13:25
philiKON[12:23] philiKONnot sure how much my opinion counts, but i dislike it :(13:26
philiKON[12:23] philiKONthe arithmetic isn't that hard. your patch makes it seem pretty complex. i didn't even pretend to understand it13:26
philiKONbtw, *you* are supposed to verify patches...13:26
philiKONmake sure they work13:26
faassenokay, Philipp, gimme an issue.13:28
philiKONfaassen, pick one :)13:29
faassenargh, okay, I will then.13:29
philiKONi mean, i could always give you one if you want one assigned13:29
philiKONbut i take it it's easier if you pick one yourself13:30
faassenis it true there's only a single zope 3 issue open?13:30
philiKONah, you mean zope 3.313:30
philiKONbut there are plenty of critical ones13:30
faassenbut they're critical for Zope 3.n where n > 3, right? :)13:31
philiKONn >= 313:31
faassenyou have an off by one error there.. :)13:31
philiKONok, let's go home then ;)13:31
philiKONi'll take 583 for warm up13:31
philiKONor perhaps i should worry about 673 first, to get it out of the way13:31
philiKONbaijum, still there?13:31
*** romanofski has quit IRC13:32
*** romanofski has joined #zope3-dev13:40
*** baijum_ has quit IRC13:41
*** wrobel has joined #zope3-dev13:45
*** baijum has quit IRC13:46
regebroOK, anything critical that looks doable I can take care of?13:48
*** kobold has left #zope3-dev13:48
philiKONregebro, check the tracker13:50
philiKONregebro, take any critical issue you like13:50
regebrophiliKON: Yeah, but there is so much! :) But I'm looking through it to see if I can find anything.13:50
regebro680 and 551 look doable.13:51
philiKONregebro, do it :)13:52
philiKONbtw, with 680 i think only .html shouldn't be a page template resource13:52
philiKONthings ending in .pt should still be13:52
philiKONregebro, about 54213:57
philiKONregebro, you just gotta follow the link :)13:57
philiKONsrichter's reply is there13:57
regebroOh. :)13:57
*** flox|away is now known as flox13:59
faassenI'm looking at issue 56214:04
faassenI'm going to try to verify that the object copier doesn't copy a version but makes a reference, even though that sounds like a stunning flaw.14:04
*** J1m has joined #zope3-dev14:05
faassenhey J1m14:05
philiKONfaassen, write that as a follow-up to the issue :)14:05
faassenokay, it doesn't happen for DTML Page.14:06
philiKONi dont' understand the issue14:06
philiKONhe's saying zope.copypastemove creates references, not copies?14:07
faassenyeah, that's what he's stating.14:07
*** J1m has quit IRC14:07
*** J1m has joined #zope3-dev14:08
philiKONi odn't understand the last paragraph in the issue14:08
faassenwell, he claims that the object copier only copies the reference.14:08
philiKONreference to what?14:08
faassenso that if he changes something in the original, it also changes in the copy, or vice versa.14:08
philiKONthat's weird14:08
faassento the object being copied. i.e. there's no copy.14:09
faassenyeah, it's definitely not what the implementation does.14:09
philiKONi wonder how it is with perrsistent references14:09
faassenit might be something special is up for names, though.14:09
philiKONsay, you have two *persistent* objects, a and b14:09
faassenanyway, there's definitely pickling going on and such.14:09
philiKONa.b = b14:09
philiKONnow you make a copy of a14:09
philiKONcall it c14:09
philiKONc = copy(a)14:09
philiKONc.b is b ?14:09
*** mkerrin has joined #zope3-dev14:10
faassenI don't think so, but we could verify it.14:10
philiKONperhaps he's epxeriencing something like this14:10
faassen    If object in the location refer to objects outside of the14:10
faassen    location, then the copies of the objects in the location refer to14:10
faassen    the same outside objects.14:10
faassenbut he's just talking about a normal attribute.14:10
philiKONwell, not sure if location means zope.location14:11
faassenanyway, lunchtime.14:11
philiKONperhaps it means persistent references14:11
faassenthis si in zope.location.pickling :)14:11
faassenI suspect his problem might have something to do with the name of the object.14:11
faassenwhich is managed rather specially compared to any other object.14:11
faassenanyway, his bug report is definitely not correct in that copies are being made, not just references.14:11
philiKONhis logic in the namechooser seems bogus too14:11
faassenidle for lunch14:11
philiKONi would just rejec tit14:12
faassenso I'm going to do a followup and decrease the importance of this one. :)14:12
philiKONyes please14:12
philiKONactually, ask for some more feedback too14:12
philiKONeet smakelik14:12
philiKONor however you spell it ;)14:12
d2mwould it make sense to add 'developer mode' information to IRuntimeInfo ?14:18
regebroAnybody keeping track of bugs fixed?14:18
MJHow much formality do people expect for cleaning up the zope.wfmc events?14:19
MJThere are events missing, and the events that are emitted mostly don't have interfaces.14:19
MJI'd love to go in and fix those shortcomings on the trunk, but is a proposal needed? A collector entry?14:20
*** stub has joined #zope3-dev14:20
* MJ dislikes process.14:20
regebroI seem to not be able to close bugs in the Zope3-tracker. If somebody want to fix that that would be welcom.14:20
*** nathany has joined #zope3-dev14:20
philiKONregebro, sure. did you port your fix to both trunk and 3.3?14:20
philiKONd2m, good idea!14:20
regebrophiliKON: Not yet.14:20
philiKONregebro, then i'm not closing the issue yet :)14:21
regebroOK. :)14:21
philiKONMJ, is there an issue?14:21
philiKON(in the collector?)14:21
philiKONJ1m, got a sec?14:23
philiKON(making zpkg understand which python versions are acceptable, etc.)14:24
philiKONi find it very frustrating. bash is a horrible scripting language14:24
philiKONthe original submitter did not submit a patch, just a "solution". and i can't make much sense of the other patch14:25
J1mI agree, which is why I stayed away from it. :/14:25
philiKONi wonder if we can just let go of this issue14:25
philiKONgiven that we might mvoe away from zpkg in 3.4 anyways14:25
J1mI think it is fairly serious though, isn't it?14:25
philiKONuntil then i'm tempted to fill the ACCEPTABLE list with "2.4.4 2.4.5 2.4.6" etc. :)14:25
philiKONwell, the script tries to do all sorts of magic to find the "right" python14:26
J1mOh. Is that a work around?14:26
philiKONi think so14:26
J1mworks for me.14:26
philiKONi'll test this first14:26
J1mI certainly don't want people to be unable to install Zope 3 w p2.4.4.14:26
philiKONi think does way too much magic. as a comment says, it could be much simpler. i wouldn't mind having tos pecify --with-python at all times14:27
philiKONthat *always* works, btw14:27
*** benji has quit IRC14:27
philiKONyou can always, even now, say --with-python14:27
philiKONthough i admit it's not obvious when configure yells at you for not having the right python version14:27
J1mI think that the goal is t avoid wasting time on responding to problems reported using the wrong version of Python.14:28
J1mI suppose though that that check could be done later in the process.14:28
philiKONin python :)14:29
J1mBut let's not worry about it util we have a re[lacement for the current mechanism.14:29
J1mHow many test failures were you seeing? Just that one?14:30
*** baijum has joined #zope3-dev14:31
philiKONJ1m, yes14:31
J1mI thnk I may have fixed that one.14:31
J1mI'm still getting another 1 or 2. (One is intermittent.)14:32
*** tlotze has quit IRC14:32
philiKONJ1m, adding future python 2.4 versions to acceptable makes it work, though you'll get a warning that you don't have the "optimum" installed. i guess that should be ok for now?14:39
philiKON(I just tested it with a fake future python 2.4 version)14:39
regebrophiliKON: OK, 680 is in 3.3 branch too.14:39
philiKONregebro, why don't i see your checkins?14:39
philiKONregebro, are you susbscribed to zope3-checkins?14:39
J1mI just approved that checkin14:40
J1mPeople should not have to subscribe14:40
regebroWell, I used to be... I'll make sure again.14:40
J1mI don't want to make people do that.14:40
J1mregebro: did you change your z.o email address?14:41
regebroJ1m: not for some time.14:41
philiKONTheuni, ping14:41
*** baijum has quit IRC14:42
regebroWell, I wasn't subscribed anyway, trying to subscribe now. :)14:43
faassenback from lunch. yay, an issue rejected, lots more to go. :)14:47
philiKONfaassen, :)14:49
philiKONnow it's me who's off to lunch14:49
regebroGah! I failed in running ALL the tests, and the fix broke something completely unrelated. :-/14:49
regebroAha. The testbrowser has ZPTs with HTML-extension and uses resource directories. Renaming them should solve it.14:52
MJphiliKON: No, there is no issue, at least none that I have put there15:00
MJphiliKON: I just implemented a workflow for a Zope3 project using zope.wfmc15:01
MJAnd I got rather intimate with the package, noting that the events are next to useless without interface declarations.15:02
MJI suspect that noone has been using the package in earnest yet.15:02
*** J1m has quit IRC15:06
*** mgedmin has joined #zope3-dev15:11
*** niemeyer has joined #zope3-dev15:13
TheuniphiliKON: pong15:18
*** batlogg has joined #zope3-dev15:18
faassendoes anyone know a lot about how exceptions are handled in zope 3?15:18
philiKONTheuni, hi there15:18
philiKONTheuni, why exactly do you prefer another beta?15:18
faassenI think I found a problem with the publisher hiding at least some exception tracebacks.15:19
philiKONwhat's wrong with a release candidate?15:19
philiKONMJ, i suggest a posting to zope3-dev15:19
TheuniphiliKON: in this case just a formality. an rc should be something that *we* expect to be what also could be release. i couldn't tell that after the quality of the first beta. :)15:20
Theuniwhat's the status of the broken trunk right now?15:20
* philiKON points to regebro ?15:20
philiKONduno, just got back from lunch15:20
philiKONTheuni, yeah quality of 1st beta was bad. i agree15:21
philiKONTheuni, i'm fine with a 2nd beta tonight15:21
philiKONrc in 1 week from now15:21
philiKONwe should probably have a bugday again next thursday15:21
regebroTheuni, philiKON: Running tests for the fix now.15:21
Theunioki doki15:21
philiKONTheuni, in any case it's a release :)15:21
philiKONTheuni, (re: failing ZEO tests)15:22
Theuniregebro: i hope you're not running them for the first time now :)15:22
regebroTheuni: Nope, but that was the first time they ran. :)15:23
regebroThat is, I ran the tests for the fix for 680 before the checkin, but i actualy didn't tun the tests for the testbrowser.15:24
regebroAnd they failed. Fixed soon.15:24
Theunibtw: i don't like the UI annotations in the style of "you have to turn this switch on to enable this option" ...15:25
philiKONalright! no more 3.3 release bugs15:30
philiKON16 critical bugs to go15:30
*** niemeyer has quit IRC15:34
*** baijum has joined #zope3-dev15:35
*** niemeyer has joined #zope3-dev15:36
* baijum will be disconnected is slow here :)15:37
philiKONwhere on this earth are you?15:37
philiKONbaijum, take a look at 50715:38
philiKONTheuni, made some good comments15:38
Theunii'm working on that already :)15:41
Theuni(notice that the issue is still assigned to me)15:41
philiKONtrue dat15:41
philiKONbaijum, pick something else then :)15:41
*** projekt01 has joined #zope3-dev15:43
philiKONhey projekt01! ready to fix some bugs?15:43
*** roym has joined #zope3-dev15:44
*** volvox has joined #zope3-dev15:47
philiKONalright, 15 bugs to go15:55
philiKONi'll be off to help my best friend on her wedding arrangement. couldn't resolve that time conflict today.. i'll be back later to make a beta2 release15:56
philiKONyou boys just keep on fixing those bugs :)15:56
*** baijum has quit IRC16:00
faassenokay, back to this window.16:01
faassenphiliKON: I have enountered what looks like a deeper underlying issue for exception handling.16:01
philiKONfaassen, i saw your comment16:01
* philiKON has to run16:02
faassenokay, see you. :)16:02
*** benji has joined #zope3-dev16:05
faassenhey benji16:07
benjihi, faassen; what's up?16:08
faassenbenji: bugday stuff16:08
faassenI'm trying to figure out why Zope 3.3 suppresses an exception16:08
faassenthen I'm trying to go look into that exception.16:08
benjifun :)16:08
faassenI was going to do the second only, but since the bug was reported something changed that eats exceptions.16:08
benjisomeone should go on an XXX hunt (I don't know if I'll have time today)16:09
benjiIf I remember correctly, there are about a half dozen that need looking into.16:09
*** kobold has joined #zope3-dev16:14
*** alecm has joined #zope3-dev16:18
faassensomehow in the case of a UnicodeError (in the webdav logic) the content-type doesn't get set in the response header.16:20
faassenand it'll trip over that in that case.16:20
faassenand raise another error masking the real error.16:20
*** rocky|Zzz is now known as rocky16:20
faassenI'm trying to understand where the content-type was set previously. I can't find relevant changes in zope.publisher.http so far.16:21
*** romanofs1i has joined #zope3-dev16:24
regebroOK starting on #55116:25
*** romanofski has quit IRC16:31
*** alecm has quit IRC16:34
*** J1m has joined #zope3-dev16:35
faassenJ1m: hey ..16:44
faassenJ1m: do you have some time to talk about some bug that's bugging me?16:44
faassenokay. I'm actually trying to solve this issue, but I ran into another issue.16:45
faassenwhat's going on is that the error reported to STDOUT is not the same as the one logged in the error reporting utility.16:45
faassenbecause there's an error that occurs during reporting.16:45
faassenthat is, during the reporting to HTTP, evidently.16:45
faassenI'm not sure why that results in the real error not being visible in the shell yet.16:46
faassenanyway, what is going on is that it tries to get a Content-Type from the response, it turns out to be None, and then it bails out as it expects something that starts with 'text/'16:46
faassensomehow this path worked back when the original bug was reported. hm, perhaps that ain't so..the original reporter might've looked in the error reporting utility instead.16:47
mgedminfaassen: I saw that bug recently, I think16:48
mgedminor something similar16:48
mgedminthe error view should set the content type header16:48
faassenyou get some unicode error that's unrelated to the real error (which in this case happens to be a unicode error too)16:48
faassenwell, it's not an error view, this is the no-error view default handing.16:48
faassenas far as I understand.16:48
faassenyes, that's the same.16:49
faassenI tried adding in the setHeader bit.16:49
faassenbut that suppressed any output for me.16:49
mgedminwhat about the error reporting utility in the ZMI?16:50
faassenthat one gets the right error.16:50
faassenwhat you describe in that issue looks like the same problem.16:50
J1mI don't have anything to add to this discussion.16:50
faassenheh. :)16:50
faassenI'm wondering whether this is somehow recently triggered.16:50
J1mBTW, it looks like someone needs to go through recent issues and mark them as needed for 3.3, if necessary.16:51
mgedminas far as I understand there are two problems here: one is that webdav fails with an error, and another is that error reporting fails with another error16:51
mgedminthere's a good rule: "fix error handling first"16:51
faassenmgedmin: yeah, I'm worrying about the latter right now.16:51
mgedminso that's issue 68616:51
MJQuick Q if someone knows the answer:16:51
benjimgedmin: +116:51
faassenyeah, I wasn't aware of this issue, but I ran into the same problem.16:51
MJIs there any access to the current request outside of browser views?16:52
MJI'd like to retrieve the URL in an event handler down the line.16:52
MJIt's okay if the request isn't available in certain circumstances.16:52
MJI don't think the request is ever registered in a thread-context..16:53
faassenMJ: there's a hack.16:54
MJfaassen: Hacks are fine.16:54
faassen        request = getInteraction().participations[0]16:55
* MJ is looking at the interaction for example..16:55
MJI was looking right then. Thanks!16:55
faassenfrom import getInteraction16:55
faassenMJ: busy with workflow, huh? I had the same issue in various workflow handling code.16:55
faassenhave a workflow event trigger the sending of an email which needs to display the URL for an object, say.16:56
MJfaassen: Yup. You can get at the process, for example, but the path there ain't in the interfaces..16:56
faassenI was using hurry.workflow though, not the core workflow. I'm curious to find out what your experience with core workflow are.16:57
faassenhow to mark an issue for Zope 3.3 release?17:00
*** Agis has joined #zope3-dev17:00
faassenAgis: hey17:00
*** russf has quit IRC17:00
Agisfaassen: :-)17:00
mgedminfaassen: you said that the fix for 686 "did not work for you"17:00
*** russf has joined #zope3-dev17:00
faassenmgedmin: yeah, I got no output, but I'll try to verify, perhaps I added it in another spot.17:01
faassenmgedmin: first I want to mark that issue as related to the Zope 3.3 release.17:01
mgedminwhat output did you expect?17:01
mgedminand where?17:01
faassenoutput on STDOUT, the error report.17:01
faassenanyway, maybe I was missing something, I'll review that.17:01
*** russf has left #zope3-dev17:01
mgedminnot all exceptions end up in stdout17:01
faassenyou'd expect the exception that shows up in the error log to appear on stdout too..17:02
*** gumpa has joined #zope3-dev17:02
faassenah, it's either 'critical' *or* Zope 3.3 release.17:02
faassenokay, well, it's critical so it's marked..17:02
faassenI upped it to critical.17:02
mgedminif there's a view registered for an exception that doesn't implement ISystemErrorView, or returns False from its isSystemError() method, then that exception will NOT appear in the log file (or in stdout)17:02
mgedminbut it will appear in the error reporting utility in the ZMI17:02
mgedminthis took me a while to figure out when I was writing my own custom exception views17:03
faassenbut there's no view registered for this exceptin in my understanding.17:03
faassenI mean, in the case of this unicode error, there doesn't appear to be a view as far as my debugging showed.17:03
*** henri_ has joined #zope3-dev17:04
faassen(I'm talking about the underlying unicode error)17:04
faassenI'll verify that too, though.17:05
faassenyeah, no view in this case.17:06
J1mphiliKON: were you planning to make the release?17:06
faassenmgedmin: do you know which code is responsible for feeding back the error to STDOUT?17:08
mgedminI do not remember, but I could find out quickly17:08
faassenmgedmin: let me take a look first, a moment.17:09
faassenthe exception side effects adapter?17:09
mgedminthat's the code that feeds te error to STDOUT17:11
faassenI wonder why it doesn't hook into the logging code instead.17:11
mgedmin(to the SiteError logger, actually, but the standard zope.conf tells it to use STDOUT, or maybe STDERR)17:11
mgedminI *think* the intent of that code is that not all exceptions are considered errors (or at least errors that are worth logging)17:12
faassenwell, but if errors are logged to the ZMI and *not* to the output17:12
faassento me that just sounds like you don't want this.17:12
mgedmingood point17:12
mgedminI just dont' know17:12
faassenI'm trying to figure out where this IExceptionSideEffect is registered.17:12
mgedminhey, response.handleException is called only if the view is not found17:12
mgedminand the error is logged to stdout only if the view was found17:13
faassenno, the view is found and set to None17:13
faassenisn't it?17:13
mgedminand bug 686 only happens in response.handleException17:13
faassenoh, wait, it isn't set to None again.17:13
d2mphiliKON: i've added a getDeveloperMode() to IRuntimeInfo, but the changes will affect 4 files (seems no way around)17:13
mgedminso you will never get the exception in stdout if you can trigger bug 68617:13
mgedminat least it appears to me from a cursory glance17:13
faassenyes, but I thought we don't trigger that anymore.17:14
faassenwith your fix.17:14
faassenbut there's still nothing on STDOUT17:14
faassen(whereas previously the spurious unicode error occurred)17:14
mgedminmy fix is inside handleException17:14
mgedminthe same code path is taken17:14
mgedminwith or without the fix17:14
faassenI'm trying to figure out where in the world it registers the IExceptionsSideEffects adapters.17:15
mgedminand this code path doesn't contain the logging to SiteError17:15
faassenbut I thought you said it was handled in the IExceptionSideEffects code?17:15
mgedminI'm pretty sure IExceptiosnSideEffects is irrelevant to both issues17:15
faassenoh, I thought you poitned me to that line.17:15
mgedminI mentioned ISystemErrorView17:15
faassenit might be I shifted a few lines in putting pdb statements.17:15
mgedminline 326 in my checkout is17:15
mgedmin                    if (ISystemErrorView.providedBy(view)17:15
mgedmin                        and view.isSystemError()):17:15
mgedminsorry about misleading you17:16
faassenoh, sorry, I think I read your line number wrong.17:16
faassenanyway, that code doesn't get executed in my case.17:16
faassenas view is None. I did verify that.17:16
faassenso that makes sense.17:16
faassenit never gets executed. I get your point now..17:17
faassenI really wonder what the reasoning is.17:18
faassenI'm going to write a post on that.17:18
*** alecm has joined #zope3-dev17:23
*** eins has quit IRC17:41
*** hdima has quit IRC17:44
*** MJ has quit IRC17:48
*** alecm has quit IRC17:56
*** agroszer has joined #zope3-dev18:00
agroszerin the Collector, is there a kind of "need-more-info" state?18:01
agroszerI'm looking at Issue 541 and unable to recreate the problem18:02
benjiagroszer: yes, you put it in that state by adding a comment that says "need more info"  :)18:02
*** agroszer has quit IRC18:11
*** gumpa has left #zope3-dev18:13
benjihmm, my joke was so bad it made him leave... I'm improving!18:16
d2mTheuni: thanks for creating the 'nodevmode' feature -- i've created a patch to add display of Developer Mode to ../++etc++process/index.html (Runtime Information), should i add it to ?18:17
regebroOK, fixed #551 too.18:17
Theunid2m: thanks. feel free, but reopen it18:17
srichtermmh, I do not like it ;-)18:18
srichterwhy not have another condition verb instead: not-have18:18
*** russf_ has joined #zope3-dev18:18
srichterI would hate to have: feature nofeature all the time, especially  since this is not feasible most of the time18:19
d2mTheuni: reopen is 'Accept' ?18:19
srichterTheuni: d2m: what do you think about having "not-have"18:19
d2msrichter: i only test for hasFeature('devmode'), but i don't know how this could be done inside ZCML18:20
Theunisrichter: fine with me, however something more module along the lines of tales' not: would be nice18:21
d2msrichter: not-have is ok with me18:21
benjisrichter: or perhaps <have not="devmode" />18:21
Theunid2m: resubmit18:21
srichterTheuni: we purposefully keeping zcml conditions very weak18:21
Theunibenji: that doesn't match the current syntax18:21
Theunisrichter: i know. so not-have is fine with me18:21
srichterwhat does it right now?18:21
benjitranslate to current syntax and re-apply :)18:21
*** romanofs1i has quit IRC18:22
srichtercurrently it does:18:22
srichter<configure zcml:condition="have devmode">18:22
srichterI suggest:18:22
srichter<configure zcml:condition="not-have devmode">18:22
TheuniJ1m told me about the intention of keeping the expressiveness weak18:23
Theuniand i agree swith that18:23
benjior <zcml:condition equals="don't-have feature"> :)18:23
srichterI think anything else would require major feature enhancements and increases expressiveness18:23
srichterbenji: I am not leaving, so your jokes get better ;-)18:23
benjiso if I get good enough, people will join just to hear the jokes... irc stand-up comedy18:24
srichtergo benji, he he ha ha lol ...18:25
*** srichter changes topic to "logs available at || paste code examples into or || bug day TODAY, live at || benji standup comedy"18:25
Theunii think "don't" is not propper written english! however, SteveA once proposed python extension for "if x ain't y"18:27
*** zagy has quit IRC18:28
Theuniregebro: could you please provide tests when fixing bugs?18:28
Theuniissue 551 does not include a test18:29
Theunioh. i just saw your comment on the collector18:29
Theuninever mind18:29
srichterTheuni: benji coming from TN will certainly support this syntax18:29
srichterit even feels more natural to him then18:30
benji+1 on ain't18:31
srichter>>> howdy(benji)18:32
benjithen we can have a better alias for hasattr: if x ain't got no "foo"18:32
srichteryo' bro'; what's u'?18:32
srichter>>> "na'ville" ain't states.NICE18:32
Theunifrom __future__ import slang18:33
srichter>>> benji ain't provides person.interfaces.IRedNeck18:36
benjiI appreciate the vote of confidence.18:38
*** zagy has joined #zope3-dev18:53
regebroTime for me to go back to vacation. Ciao!18:58
*** regebro has quit IRC18:58
*** henri_ has quit IRC19:01
*** jinty has quit IRC19:08
*** jinty_ has joined #zope3-dev19:08
*** kobold has quit IRC19:14
*** kamalgill has joined #zope3-dev19:18
* philiKON is back19:22
*** MJ has joined #zope3-dev19:26
*** zagy has quit IRC19:28
*** volvox has quit IRC19:28
philiKONTheuni, ayt?19:29
TheuniphiliKON: yupp19:30
philiKONi agree with stefan that "nodevmode" and devmode complicates things19:30
philiKONalso, what's up with the ftesting-*.zcml clutter now?19:30
philiKONhave you tested instances?19:30
faassenokay, I just deferred an issue. I hope we have a workflow to check out deferred issues again later. :)19:30
Theunisure. anyone who wants to spend time on infrastructure to do a "not-have" is welcome.19:31
philiKONi can do that19:31
philiKONquestion is whether that counts as a feature :)19:31
philiKONi would say it does19:31
TheuniphiliKON: i needed the ftesting environment but without the devmode. as i couldn't get to securitypolicy.zcml from my package i did it on the root.19:31
Theuniyup it does. so trunk only please. :)19:32
*** zagy has joined #zope3-dev19:32
Theunii haven't tested instances though ... i didn't anticipate a change there ...19:32
philiKONi also think that by introducing the nodevmode feature we have introduced another, well, feature19:32
Theunilet me see19:32
*** flox has quit IRC19:32
philiKONi'm just disturbed by the massive infrastructural change you did19:33
philiKONafter calling it "an easy one" in the issue19:33
philiKONi'm for the idea, but properly done, on the trunk19:33
Theunihmm. well. we need that information flag, and i very much don't like the tag on the link, because the devmode-default change needs to be communicated properly in 3.3 not in 3.4.19:34
Theunias zope 3.3 does not support no-have i consider nodevmode a reasonable workaround19:35
philiKONthis zig zag will be hell to document19:35
Theuniwhat zigzag?19:35
philiKONdo it this way in 3.3, that way in 3.4...19:35
Theuninah. there is no "do it this way" in 3.319:35
Theuniit's a workaround. it's not a "do it this way"19:36
philiKONi don't understand why we need nodevmode so badly19:37
philiKONi mean, sure, it's nice having a page at ++apidoc++ that tells users "hey, enable me"19:37
Theunino, it's necessary to not cause headache because of a policy change.19:37
philiKONwhat headache?19:37
Theunisee the issue19:37
philiKONah, because we made devmode off by default19:38
Theunione developer took half an hour to find out that in his instance dev-mode is now off by default19:38
Theuniyou can't discover that easily if you migrate existing instances19:38
philiKONwhen was devmode introduced? 3.2?19:38
Theunino idea19:38
philiKONseems so19:38
philiKONi was against having it in zope.conf anyways19:38
philiKONit's a zcml thing19:38
philiKONit doesn't configure a server or a database19:39
philiKONit configured software registrations19:39
Theunii would it expect to do something like debug mode19:39
Theuniin zope 219:39
philiKONi never understood why the devmode zcml feature needed to be special-cased in zope.conf19:39
philiKONTheuni, it isn't the same19:39
philiKONi think that's a false illusion19:39
philiKONstopping my rant now :019:39
Theunithen it maybe shouldn't be in there. but that's another totally different discussion.19:39
philiKONi'm still not happy about the ftesting clutter19:40
Theunii don't think it's a clutter19:41
Theuniwhy do you think it's a clutter?19:41
philiKONit's 3 files now instead of one. just because of 1 additional zcml feature19:41
Theunii see that we have to ftesting.zcml around19:41
philiKONand instances don' thave it19:41
Theunistupid difference to zopeskel19:41
philiKONthe whole ftesting.zcml thing needs to be rethought too19:41
philiKONi think you might want to take a different approach for your test19:42
philiKONi'm not sure19:42
philiKONshit, we still have 15 open bugs? when we started today it was 16 or something?!?19:42
philiKONwhat happened?19:42
Theunimaybe. i could be told to remove that "clutter" if you tell me how to replicate the ftesting.zcml in my module and how to include securitypolicy.zcml from there19:42
Theuni dunno, i saw people touching stuff19:42
Theuniwhat's the query you run for finding the 15 bugs?19:43
philiKONwhat's up with now? is anyone taking it?19:43
philiKONTheuni, select critical19:43
philiKONfrom the categories19:43
Theunithat's all issues19:44
Theunino bugs19:44
Theunialthough quite a few might be misclassified19:44
* mgedmin is for condition="not: have devmode" for symmetry with TALES not:19:45
Theunimgedmin: me too, partially, but i also agree on keeping zcml much more simple19:45
philiKONmgedmin, that would change the whole "verb object" philosophy19:45
mgedminif "lost devmode"?19:45
mgedmin"misplaced devmode"?19:45
philiKONdonthave devmode19:45
philiKONnotinstalled Products.Five19:45
philiKONi think adding those two verbs wouldn't be that big'a deal19:46
philiKONeven in 3.3...19:46
*** alecm has joined #zope3-dev19:47
Theunii think we could ask for it on the dev-list but i really don't feel like adding those verbs in 3.3.19:47
philiKONTheuni, any particular reason why you have like "enabled.zcml" and "disabled.zcml" files?19:47
Theuniwe need to get it out of the door. soon.19:47
TheuniphiliKON: actually i wondered about that initially if i could use a generic xml container element to do that in one file.19:48
Theunihmm. so i can nest configure within a file ... didn't try that.19:49
philiKONyou can nest as many <configure>s as you want19:49
Theunigood to know19:49
Theunithen that should be refactored. i agree.19:49
* Theuni refactors19:49
* philiKON looks at Theuni's ftest19:50
philiKONbtw, here's a reason why we should have a 'donthave' in zope 3.3:19:51
philiKONwhat if somebody enables both devmode and nodevmode featuers?19:51
Theuniit gives a conflict error when starting19:51
Theunii'm perfectly aware of that19:51
Theunithat's why it's called a workaround19:51
* philiKON grumbles19:52
philiKONJ1m, ayt?19:52
Theuninope, i think he'll be back in 10-15 minutes19:53
philiKONTheuni, btw, as for bug vs. issue in the tracker. everything that's there is an issue to me that needs to be resolved eventually (read: for 3.3 final)19:54
philiKONotherwise we can just as well a) defer it or b) give it a different status19:55
Theunithen it should be set to priority 3.319:55
philiKONwhat's critical then?19:55
philiKON"Oh, it's fairly serious, but we don't really wanna go into it"???19:55
philiKON"Yeah, I looked at it this last bugday and didn't really wanna fix it"???19:55
TheuniWell. There are issues and bugs. And there might be critical things that we decide to postpone, e.g. for 3.3.1 or whatever19:55
*** roym has left #zope3-dev19:56
Theuniimportance is not necessarily correlating with the urgency19:56
Theuniit might be important (critical) but we might have time to catch up with it19:56
philiKONhmm. i don't get this, but perhaps i don't have to19:57
philiKONTheuni, so, what about now?19:57
philiKONshall i take it?19:58
Theunii didn't touch it. i don't have time for any more bugs right now, take it if you want.19:58
philiKONactually, perhaps you're right about the urgency :)19:58
philiKONperhaps i should see what's really stopping a 3.3 release19:59
*** whit has quit IRC19:59
philiKONi think 507 should be addressed, some way or another. looks like srichter and I are not 100% satisfied yet19:59
philiKONlet's see what J1m has to say19:59
Theunii'm willing to take care for the stuff that needs to be done on 50719:59
Theuniit's also still assigned to me19:59
*** whit has joined #zope3-dev19:59
srichtercool, I am glad you are going back on 50720:00
Theunii've started the refactoring of the zcml but somethings biting me there20:00
philiKONid' be willing to help with the donthave and notinstalled verbs20:00
philiKONlike what?20:00
Theunisome registrations don't get in as before. looking into that later. need to do other stuff now.20:00
philiKONTheuni, when do you think you'll have time to look at this again?20:01
philiKONi want to make a beta2 tonight20:01
philiKONand i think 507 needs to be addressed somehow20:01
philiKON(well it sorta is already, but it'd be nice if done properly, whatever that means)20:02
Theunihmm. dunno. tonight for me, i guess.20:05
philiKONsame here20:06
philiKONi'll be online20:06
philiKONjust ping me20:06
*** Theuni has quit IRC20:27
*** gintas has quit IRC20:33
*** alecm has quit IRC20:34
*** alecm has joined #zope3-dev20:34
philiKONJ1m, so, now turns off devmode by default20:35
philiKONwell, actually, if nothing is specified in zope.conf, devmode is off20:36
philiKONit used to be on20:36
philiKONusers who upgrade to the new release we'll be making will find their apidoc etc. turned off20:36
*** Theuni has joined #zope3-dev20:36
philiKONthat's why theuni has created a page that will appear at ++apidoc++ when devmode is turned off20:37
philiKONtelling them to turn devmode back on, explicitly20:37
philiKONfor that, he needed to create a "nodevmode" feature20:37
J1mThat's nice20:37
philiKONbecaus ethere's currently no way to say:20:37
philiKONzcml:condition="not have devmode"20:37
philiKONor something20:37
philiKONi don't think it's nice having two features, devmode and nodevmode20:38
philiKONi think it should be one feature and the verb should be have or donthave20:38
philiKONverb being the verb in zcml:condition20:38
philiKONon a related note, i've had the use case of needing a negated "installed" verb20:38
philiKONzcml:condition="notinstalled Products.Five"20:39
philiKONquestion is, should we add those negated verbs to 3.3, or keep the workaround of two features?20:39
J1mare we talking about zope.conf? or site.zcml?20:40
philiKONadding them would be a feature, i guess20:40
J1moh never mind20:40
philiKONwe're talkign about adding them to zope.configuration20:40
*** flox has joined #zope3-dev20:40
J1mI guess I have a slight preference for notdevmode20:41
benjiI guess you guys missed my morning show; I though we'd decided to do zcml:condition="not-have foo".20:41
J1mI don't feel strongly about it.20:41
J1mI've for whatever has the least impact.20:41
philiKONbenji, oh, i must indeed have missed that20:41
philiKONimpact is hard to measure. i think it's just hard to document devmode vs. nodevmode, for example20:42
philiKONeveryone will think: whyd idn't they go for a negated verb?20:42
J1mI'll defer20:43
benjiwe also decided on a synonym: zcml:condition="ain't-got foo"20:43
philiKONsynonym for installed would be in-da-house?20:43
philiKONJ1m, ok :)20:44
benjiI was going more for hillbilly, not gangsta20:44
benjiI know they're hard to tell apart.20:44
philiKONit's all yanks to us20:44
* philiKON hides20:44
philiKONso, not-have and not-installed?20:44
*** batlogg has quit IRC20:45
philiKONTheuni, ?20:45
Theuniwhat about the "?"20:47
philiKONasking you for a vote20:47
faassenI'm out of here for the night.20:50
faassensorry I couldn't actually resolve more bugs. :(20:50
faassenit also was sort of like "Infrae bug day here", like too often. :)20:51
*** faassen has quit IRC20:51
*** stub has quit IRC21:00
TheuniphiliKON: +1 then21:01
philiKONTheuni, i do the verbs, you do the rest?21:01
philiKONis that TCP for acknowledged?21:03
Theuniack fin21:04
Theuni(actually that would be FIN and you say FINACK)21:04
philiKONsorry, my TCP is horrible... ;)21:04
benjiso from now on, instead of saying ping Theuni, I'll say SYN Theuni and wait for the SYN/ACK21:08
*** hazmat_ has joined #zope3-dev21:09
Theuniand i'll never say that and when you try more thann 4000 times your stack will overflow and you can't talk anymore to other people.21:10
philiKONsounds like a good protocol21:10
Theunithat's what the world relies on21:10
*** RaFromBRC has joined #zope3-dev21:17
*** gintas has joined #zope3-dev21:23
*** dunny has joined #zope3-dev21:32
*** russf has joined #zope3-dev21:37
*** romanofski has joined #zope3-dev21:38
philiKONTheuni, 3.3 branch now has not-have and not-installed21:39
philiKONand now trunk does too21:40
*** russf_ has quit IRC21:46
*** batlogg has joined #zope3-dev21:49
*** kamalgill is now known as kamal-lunch21:56
*** jinty_ has quit IRC21:57
*** mkerrin has quit IRC22:01
*** SiggyF has joined #zope3-dev22:07
*** nathany has quit IRC22:15
*** kobold has joined #zope3-dev22:21
*** WebMaven has joined #zope3-dev22:24
*** efrerich_ has joined #zope3-dev22:26
*** dunny has quit IRC22:26
*** efrerich has quit IRC22:34
*** alga has quit IRC22:44
*** zagy has quit IRC22:45
*** mgedmin has quit IRC22:46
*** kamal-lunch has quit IRC23:00
*** russf_ has joined #zope3-dev23:16
*** russf has quit IRC23:16
*** SiggyF has quit IRC23:21
*** kamalgill has joined #zope3-dev23:26
*** philiKON has quit IRC23:50
*** efrerich_ has quit IRC23:52
*** nebo has joined #zope3-dev23:56
*** nathany has joined #zope3-dev23:59

Generated by 2.15.1 by Marius Gedminas - find it at!