IRC log of #zope3-dev for Saturday, 2005-11-12

*** sashav has quit IRC00:01
*** sashav has joined #zope3-dev00:01
*** sashav has joined #zope3-dev00:02
*** sashav has quit IRC00:02
*** benji has left #zope3-dev00:19
*** Aiste has quit IRC00:39
*** JoaoJoao has joined #zope3-dev00:50
JoaoJoaogood evening00:50
*** M1 has joined #zope3-dev00:59
*** MJ has quit IRC01:00
*** M1 is now known as MJ01:00
*** Aiste has joined #zope3-dev01:00
*** zbir has quit IRC01:07
*** horizon5 has joined #zope3-dev01:16
*** horizon5 has joined #zope3-dev01:18
*** jinty has quit IRC01:22
*** efge has quit IRC01:31
*** efge has joined #zope3-dev01:32
*** fdrake has joined #zope3-dev01:33
fdrakeI've spent the week at the "packaging sprint" the Plone folks are holding at the CIGNEX offices, looking at what's needed to support Zope code that's importable from ZIP archives.01:35
fdrakeI've written up some notes about what I've done so far, and what still needs to be considered:01:35
fdrakeI plan on turning this into a Zope 3 proposal, but there are still things to be considered.  I'm pretty fried, but if anyone's interested in the branch, that should provide something to start with.01:36
*** SiggyF has joined #zope3-dev01:53
*** zbir has joined #zope3-dev02:00
*** d2m has quit IRC02:08
*** efge has quit IRC02:17
*** newpers has joined #zope3-dev02:21
*** deo has joined #zope3-dev02:45
*** yota has quit IRC02:58
*** sashav has joined #zope3-dev03:07
*** yutakashino has quit IRC03:09
*** horizon5 has quit IRC03:37
*** sashav_ has joined #zope3-dev04:19
*** sashav has quit IRC04:19
*** sashav_ is now known as sashav04:19
*** fdrake has left #zope3-dev04:23
JoaoJoaoWhat is the Zope3-ish way to catalog objects?04:39
*** newpers has quit IRC04:55
*** JoaoJoao has quit IRC05:58
*** newpers has joined #zope3-dev06:00
*** newpers has quit IRC07:45
*** deo has quit IRC08:02
*** dman13_ has joined #zope3-dev08:04
*** dman13 has quit IRC08:10
*** xanalogica has joined #zope3-dev08:40
*** xanalogica has left #zope3-dev09:02
*** sashav has quit IRC09:17
*** sm has quit IRC09:43
*** agroszer has joined #zope3-dev09:46
*** Volpe has joined #zope3-dev10:16
VolpeI'm having trouble with the pythonpath when using zope310:29
Volpewhen I try to run the example (messageboard) from the book.  the like from messageboard.interfaces import IMessage throws an import error 'No module names interfaces'  but there cleary is an interfaces module and just plain "from interfaces import IMessage" works fine. :\10:32
Volpeany reason for this?10:37
*** d2m has joined #zope3-dev10:41
Volpeif I open a python interactive session from lib/python (i.e. so lib/python is on the python path)  I can write 'from messageboard.interfaces import IMessageboard' but when I run the tests on the messageboard package it gives the ImportError (even though it lists lib/python in sys.path).  What could explain this?11:03
d2mVolpe: i think you export the PYTHONPATH and run the tests from the zope instance folder, like: python bin/test --dir=modulepath11:14
*** SiggyF has quit IRC11:18
Volpeyeah, I'm doing that (I think).11:19
*** Volpe has quit IRC11:28
*** horizon5 has joined #zope3-dev12:29
*** yota has joined #zope3-dev13:05
*** suse-joe has joined #zope3-dev13:14
*** strichter has joined #zope3-dev13:54
*** srichter has quit IRC13:55
*** strichter is now known as srichter13:55
*** ChanServ sets mode: +o srichter13:55
*** mp has quit IRC14:34
*** philiKON has joined #zope3-dev15:40
*** helmutm has joined #zope3-dev16:46
*** horizon5 has joined #zope3-dev16:54
*** horizon5 has joined #zope3-dev16:55
*** horizon5 has quit IRC16:57
*** tav has joined #zope3-dev16:59
*** MJ has quit IRC16:59
*** MJ has joined #zope3-dev17:02
*** zagy has quit IRC17:21
*** zagy has joined #zope3-dev17:22
*** J1m has joined #zope3-dev17:57
*** efge has joined #zope3-dev18:00
*** suse-joe has left #zope3-dev18:18
srichterJ1m: could you add a motivational paragraph for extrinsic references?18:20
J1mare you kidding?18:21
srichterno, I understand intrinisc references just fine, but then you dive into how you solve extrinsic references without giving an explanation what they are good for18:21
J1mI just checked in a motivational paragraph a few minutes ago.18:22
J1mI even used those words in the checkin message.18:22
srichterI know, and it only talks about intrinsic references18:22
srichterbut it does not explain extrinsic reference examples18:22
srichterfor example, you say the wheels on a car are intrinsic18:22
srichterso what is an example of an extrinsic reference?18:23
J1mah ok18:24
*** sashav has joined #zope3-dev18:28
J1mHow's that?18:32
* J1m is not feeling very literary this moring. :)18:33
* J1m is not feeling very literary this morning. :)18:33
srichterhe he18:33
J1mwe plan to release shortcuts too, fwiw18:33
J1mand a bunch of other stuff.18:33
J1mso much to do, so little time ...18:34
srichterok, much better18:34
srichterI understand it now :-)18:35
srichterJ1m: so every ExtrinsicReferences instance would be one type of reference18:39
srichter(like the verb in RDF language18:40
srichterok, I like it18:40
J1mAn interesting implementation note: this doen;t use intids. :)18:40
srichterthough Helmut would say now: another reference/relationship framework! :-)18:40
srichterJ1m: is it using weakrefs?18:41
srichterOk, I see18:41
J1mU put this out there in part in response to Helmut's plans to write a proposal.18:41
srichteroh, now I understand18:41
srichterthus you need the IKeyReference adapter18:41
srichterI like it18:42
J1mA relationship thing could be build on this.18:42
srichterit is very Pythonic and Zope 3 ish18:42
J1mIn particular, it is very small.18:42
srichterand since it is using btrees it is efficient I assume18:43
J1mThat's the plan. :)18:43
*** bradallen has left #zope3-dev18:43
srichterok, so +1 from me18:43
J1mI'll be interested to see if the people who want relationship systems find it useful.18:44
srichterI guess we should try to convert the st.relationship code to using it18:45
srichterthis way we could figure out if it meets our needs18:45
J1mI dunno18:45
J1mat least as a thought experiment.18:45
J1mI'm not proposing this as some kind of standard that people have to follow.18:45
srichterwell, one of the big criticisms of the ST relationship package is that it has a linear search18:46
efgeyeah :)18:46
srichterso this would solve this problem I guess18:46
*** natea has joined #zope3-dev19:02
srichternatea: so how did the meeting go?19:05
nateasrichter, pretty well19:06
nateaturbogears, django and aquarium were presented19:06
nateaand i gave an impromptu talk about plone/zope/zope319:06
nateathere were a lot of people interested in zope after i described where it's been and where it's going19:07
srichterhuh, interesting19:08
srichteris there anything planned for next time?19:08
nateabut i think a talk from you would be best19:08
nateano, i don't think so19:08
nateayou should contact chris curvey to schedule a talk19:08
srichterI wonder whether I should do another talk at Tufts19:08
srichterthis is much less overhead for me19:09
*** sashav has quit IRC19:22
*** natea has quit IRC19:23
helmutmhi jim, hi stephan - sorry, I was off my computer so i missed your discussion :-(19:30
srichterhelmutm: no problem19:31
srichterhelmutm: I was just given Jim some feedback about the extrinsic property code19:31
helmutmyes, i've read it, and also looked at the code19:32
helmutmso it might provide a great part of what i was working on19:32
helmutmbut i'd like to think a little more about it...19:32
srichteryeah, comments would be very welcome19:38
srichterI think we can profit from some experience of the Zope 2/Plone world19:38
helmutmone thing is just the plug-and-play aspect: in Plone you have the archetypes reference framework and you just use it...19:43
helmutmthat may mean that you have to tune it later or even change implementation19:44
helmutmbut you get a quick start19:44
srichterone goal of mine for 3.2 is to have a Zope 3 TTW development package19:45
srichterfirst step: make sure that J1m takes the time to sit down with me and fix the persistent interface bug. :-)19:45
*** BjornT has joined #zope3-dev19:45
srichterI will probably base a lot of the work on the new Boston skin19:46
srichterI want a click-and-done dev interface :-019:46
helmutmto be honest: i'd be happy to be able to do all the stuff i'd like to do just by writing python code...19:49
philiKONsrichter, you mean a goal for 3.3, right?19:52
srichteryes, of course19:52
philiKONok :)19:52
srichterhelmutm: J1m is working on archetypesxml for Zope 3, which would do that19:52
* philiKON is going to bed: g'night19:53
*** philiKON has quit IRC19:53
srichterI just want to be able to quickly write Web pages without any design19:53
srichterliek my personal one :-)19:53
*** sm has joined #zope3-dev19:53
d2msrichter: you answered my question regarding POST handling today -- i am really trying to figure out how (and if at all) a POST to the object is processed -- there seems to be no way to update an object except you add another view (if you do not want to use @@edit.html and all its troubles, different field names, different form-enctype,..)20:04
d2mif this is by design i did not find documentation on the topic20:05
srichterwell, I think PUT is defined in general20:05
srichterit uses the IWriteFile adapter of an object20:06
d2mPUT works ok for creating objects20:06
srichter(actually I am sure I jsut worked with that this week)20:06
srichterit should also work for updating20:06
d2mbut you get into troubles when using objectids without extensions20:06
srichterit just looks for an unnamed adapter then20:06
srichterwe use that all the time in SchoolTool20:07
d2mwhich defaults to a File20:07
d2mok, so it is on purpose that POSTs are not processed ?20:07
srichterthere is simply no view defined for it20:08
d2mbut you are forced to use the @@edit.html view then20:08
J1mwhat are y'all talking about.20:10
srichterabout not having a default POST view20:10
srichterlike we have a default PUT view20:11
J1mThat is different from the GET view?20:11
d2myes, the problem when you want to switch to another form component20:11
srichter(in SchoolTool's REAT API GET is often different from POST)20:11
J1mSeems like an odd use of post.20:13
d2mwith zope2 you are used to just post to the object itself and get the data processed20:13
srichtermaybe GET and POST is the same; I might have lied here20:14
d2mshould be the same with zope3 i think20:14
J1mIt is20:14
J1mThey are treated pretty much equivalently.20:14
d2mif you post to the object in Zope3 (and not to the @@edit.html view as it is now) the data is not processed, the object is not changed, its like a GET request20:15
srichterwell, because you post on the default view, which is @@index.html20:15
J1md2m, that's up to the thing you post to.20:16
d2mwell, ok, let that view handle post requests then20:16
J1mZope 3 has nothing to do with it.20:16
J1mViews handle all requests.20:16
J1mzope 3 doesn't make a distinction between GET and POST when deciding which view to call.20:17
d2mplease put that in big red letters on top of the ZMI20:18
d2mwell, the view should process the request20:18
* J1m is confused20:19
d2mok, could well be i am mixing up some things20:21
srichterd2m: how do you what request was sent?20:21
d2msrichter: i do not understand20:23
srichterhow will the /path/to/object view know whether the request came from edit.html, someotheredit.html or whatever?20:25
*** J1m has quit IRC20:27
d2msrichter, you are right, its completly my missunderstanging20:27
d2mi simply got confused by fact, that PUT and DELETE works well on the container20:28
*** helmutm has left #zope3-dev20:34
*** newpers has joined #zope3-dev20:35
agroszersrichter: I just checked schooltool.relations inspired by the above discussion20:42
agroszerin fact, that's cool20:42
srichterthank the POV guys20:43
agroszerjust one small thing that I miss, is to integrate it somehow with the schema/form framework20:43
agroszerwhat do you think it would be hard to implement?20:47
srichternot using a vocabulary I think20:51
srichterI am pretty sure this is commonly done in schooltool20:51
agroszerI think relations are used quite often in applications, so let's dream a bit:20:54
agroszerI could specify in the schema that the field is a relation20:54
srichternote that you can do a lot of what you about to dream with a List/Choice field combination and using vocabularies20:56
agroszeryes, what I don't tried yet is how it behaves with stool.relations20:56
agroszeras I see in stool it takes some lines to describe&handle the relations20:59
agroszerbut, relations could be described by let's say "parent" object, the "child", both interface are in the system,21:01
agroszerI jusst have to specify where do I store the children21:01
agroszerbut that's simple thinking21:01
agroszer(even coming from the relational world)21:02
agroszerNot thinking about the consequences, even the forms may be created using the above infos21:03
*** natea has joined #zope3-dev21:14
agroszereek, another thread21:16
agroszercould somebody run the zope/app/form/tests/
srichter[srichter@einstein Zope3-Tutorial]$ python -vpu1 -s test_utility21:18
srichterRunning tests at level 121:18
srichterRunning unit tests:21:18
srichter  Running:21:18
srichter  Ran 458 tests with 0 failures and 0 errors in 6.520 seconds.21:18
srichter[srichter@einstein Zope3-Tutorial]$21:18
agroszer*scrathing head21:19
agroszeron line 544 it expects21:19
agroszer"field: <class ''>"21:19
agroszerbut I get "field: <class '__main__.Foo'>"21:20
srichterthat's probably a doctest thing21:20
agroszeryes it is21:20
agroszerI started the "by hand"21:20
srichterwell, I never do that21:22
srichterit should work21:22
agroszerLet's say I'm a beginner21:23
srichterbut if it does not I am not going to try to figure out why21:23
agroszerand beginners sometimes do unexpected things :-)21:23
agroszerI think it's because when the "main" runs the tests, it imports each one, and that's why different module names come21:26
agroszerin the doctest, how can I write (.*) ?21:26
srichterthat's probably right21:26
agroszeror (ignore this)21:26
srichterI dunno21:26
agroszerI mean how to expect "field: <class '(ignore this).Foo'>"21:27
srichter<class '...Foo'>21:30
srichterof course you need to have the ELLIPSIS flag turned on21:30
agroszerI see, I remembered, that I saw it somewhere :-)21:30
agroszerhave you checked the diff's of the CustomSequenceWidgetFactory solution?21:35
srichterno, you should really send it to zope3-dev. sp more people have an opportunity to review it21:40
agroszerok, I'll do that21:41
agroszerjust one question21:41
agroszeris it OK to create an ICustomWidgetFactory interface, instead of IViewFactory?21:42
agroszersignature is the same21:42
srichterfor what?21:43
agroszeryes... first I thought it would be different, now it is the same21:45
agroszerand as checked IViewFactory is going to be deprecated21:46
agroszerthat's it I can't decide21:48
agroszeryou're the expert :-)21:48
srichterI see22:09
srichterI check it out22:09
agroszerInbetween I now, why it should be ICustomWidgetFactory22:09
agroszerthere should be an ICustomSequenceWidgetFactory too, because it's __call__ will have different signature22:10
agroszerI just noticed, that sequence widgets have again different __init__ signatures22:12
agroszeryay, object widget too22:14
agroszerthat means my current fix is more like a hack22:16
agroszerok, I'll post the situation to zope3-dev, let's have opinions22:19
*** sashav has joined #zope3-dev22:27
*** natea has quit IRC22:40
*** ignas has joined #zope3-dev22:41
agroszergood night22:45
*** agroszer has quit IRC22:45
*** ignas has quit IRC23:09
*** ignas has joined #zope3-dev23:27
*** ChanServ sets mode: +o tav23:29
*** sashav_ has joined #zope3-dev23:32
*** sashav has quit IRC23:32
*** sashav_ is now known as sashav23:32
*** drzoltron_ has joined #zope3-dev23:42
*** jhauser has quit IRC23:59

Generated by 2.15.1 by Marius Gedminas - find it at!