IRC log of #zope3-dev for Thursday, 2006-01-05

*** pcardune has joined #zope3-dev00:18
*** andres has joined #zope3-dev00:33
*** zbir has quit IRC00:47
*** _pcardune has joined #zope3-dev00:49
*** GaryPoster has quit IRC01:01
*** natea has quit IRC01:06
*** retsu has quit IRC01:06
*** _anguenot has quit IRC01:14
*** _anguenot has joined #zope3-dev01:14
*** andres_ has joined #zope3-dev01:33
*** andres has quit IRC01:33
*** elbixio has joined #zope3-dev01:36
*** sashav has quit IRC01:49
*** natea has joined #zope3-dev01:54
*** sashav has joined #zope3-dev01:55
*** zbir has joined #zope3-dev02:11
*** sashav has quit IRC02:13
*** tarek has quit IRC02:19
*** elbixio has quit IRC02:21
*** J1m has quit IRC02:25
*** retsu has joined #zope3-dev02:27
*** natea has quit IRC02:29
*** philiKON has quit IRC02:38
*** dman13 has joined #zope3-dev03:18
*** retsu has quit IRC03:21
*** yota has quit IRC03:35
*** benji has quit IRC03:39
*** MiUlEr has joined #zope3-dev04:02
*** tahara has joined #zope3-dev04:18
*** WebMaven has joined #zope3-dev04:32
*** retsu has joined #zope3-dev04:34
WebMavensrichter: yo.04:46
WebMavensrichter: are you working on a new edition of the 'Zope 3 Developers Handbook'?04:47
*** hazmat has joined #zope3-dev05:23
*** retsu has quit IRC05:38
*** hazmat is now known as haz_out05:48
*** retsu has joined #zope3-dev05:48
*** dman13 has quit IRC07:19
*** natea has joined #zope3-dev07:28
*** eins has joined #zope3-dev08:26
*** dobee has joined #zope3-dev08:45
*** dobee has quit IRC08:57
*** sashav has joined #zope3-dev09:21
*** MJ has quit IRC09:22
*** dobee has joined #zope3-dev09:25
*** sashav has quit IRC09:26
*** rom|aw has joined #zope3-dev09:46
*** zagy has joined #zope3-dev09:49
*** rom|aw is now known as romanofski10:03
*** pcardune_ has joined #zope3-dev10:04
*** jhauser has joined #zope3-dev10:05
*** _pcardune_ has joined #zope3-dev10:06
*** Theuni has joined #zope3-dev10:06
*** agroszer has joined #zope3-dev10:15
*** retsu has quit IRC10:21
*** pcardune has quit IRC10:21
*** _pcardune has quit IRC10:21
*** sashav has joined #zope3-dev10:31
*** d2m has quit IRC10:34
*** vlado has joined #zope3-dev10:42
*** MJ has joined #zope3-dev10:42
*** d2m has joined #zope3-dev10:49
*** retsu has joined #zope3-dev10:56
*** yota has joined #zope3-dev11:12
einshow do I set my own permissions for adding objects through "/+/"? Zope requires zope.ManageContent permission to access "+" view11:13
einsromanofski hi11:13
romanofskieins: hi!11:22
*** tarek has joined #zope3-dev11:25
*** retsu has quit IRC11:27
taharaeins, you can set a permission using containerviews directive.11:41
*** tarek has quit IRC11:45
*** tarek has joined #zope3-dev11:46
VladDrachmm I didn't know metal fill-slot can't be limited through tal:condition11:48
VladDraclame :(11:48
einsthanks tahara, it worked :)11:48
VladDrac<metal:block fill-slot="someslot" tal:condition="condition"> .. </metal:block>11:49
SteveAVladDrac: well...11:57
SteveAyou're relating two different things11:57
SteveAmetal is one language11:57
SteveAit runs first11:57
SteveAtal is another language11:57
SteveAit is processed after the metal11:57
VladDracso I noticed12:03
VladDracI have a workaround, move the condition inside the fill-slot and reinclude the macro if false12:04
VladDracbut that's less clean imho12:04
VladDracconceptually, <metal:macro is just another tag that can conditionally be included/excluded using tal12:05
VladDracjust like html, xml, whatever12:05
SteveAnot really12:05
VladDracbut okay, it works, and it still alot better than kid :)12:05
SteveAbecause it is run first12:05
SteveAlike macros in C12:05
VladDracconceptually == in my brain :)12:05
SteveA tal:replace="structure template/macros/macroname" tal:condition="whatever"12:06
SteveAis maybe what you'd like to do12:06
SteveAi don't think that will work12:07
VladDracI want to leave a slot untouched / replaced depending on a condition12:07
*** ChrisW has joined #zope3-dev12:11
ChrisWanyone here?12:11
TheuniSteveA:  metal and tal are processed alternating iirc ...12:23
ChrisWah, life :-)12:23
ChrisWis there any way to set a default skin at the site level?12:24
Theunishouldn't it work by making a <tal:block> around the use-macro that does the condition?12:24
TheuniHmm. Wait. this would just work even if the metal is executed.12:25
Theunioh wait again ... you're talking about fill-slot12:25
* ChrisW wonders if he's missing something of is Christian is really having a monologue ;-)12:26
TheuniChrisW:  there was discussion before you entered :P~12:26
ChrisWaha :-)12:26
* ChrisW is all excited having finally gone through all of Philipp's book :-)12:27
ChrisWapparently I'm a Zope 3 expert now ;-)12:27
ChrisWand bumping into the rough edges now ;-)12:28
ChrisWI am finally working on Swishdot ;-)12:28
ChrisWI have real code, for Zope 3 ;-)12:28
ChrisW(not much, mind)12:29
ChrisWokay, I'm only online for a short-ish time but I have looots of questions12:29
ChrisWzope3-users should get interesting when I'm back online properly ;-)12:29
taharaI refered schooltool source.12:34
taharasee schoolToolTraverseSubscriber in
ChrisWsorry, not sure what you're referring to?12:36
ChrisWthat lets you set a skin per site?12:36
taharaah, I'm sorry.12:36
ChrisWoh, right12:36
ChrisWI'm suprised that isn't a core part of Zope 3's Site interface...12:36
*** mgedmin has joined #zope3-dev12:38
*** alga has joined #zope3-dev12:38
*** philiKON has joined #zope3-dev12:47
ChrisWhey Phil :-)12:57
ChrisWyou here?12:58
ChrisWanyone know how big a checkout of is going to be?12:59
ChrisWseems like it might be a good example app...12:59
*** MJ has quit IRC13:00
*** ignas has joined #zope3-dev13:00
philiKONhey ChrisW13:01
ChrisWhey Phil :-)13:01
ChrisWjust wanted to thank you for the grat book13:01
ChrisWfinally had a chance to read it all13:01
ChrisWjust finishing up the security chapter now :-)13:01
philiKONcool :)13:01
philiKONi'm glad to hear that13:02
*** mgedmin has quit IRC13:02
ChrisWheh, I have so many questions now though13:02
ChrisWnothing like working offline for a couple of weeks with a book and a trunk checkout to leave me with lots to ask13:03
*** mgedmin has joined #zope3-dev13:03
ChrisWmost of it seems to be around how to make things happen at the site level13:03
ChrisWlike, for example, setting a default skin for the site13:03
philiKONoh, hehe, yeah13:03
ChrisWsomeone suggested schooltool does this, but I saw no evidence of it13:04
philiKONtechnically possible13:04
ChrisWin zope 3 core13:04
philiKONjust the UI is missing13:04
ChrisWwhich is very suprising13:04
ChrisW(bear in mind I'm used to running many "sites" of one Zope instance)13:04
philiKONthough I thought schooltool didn't use sites13:04
ChrisWwhere should I look?13:04
philiKONyes, running several sites in one zope instance is possible13:04
philiKONthat's whole idea of sites13:05
philiKON(sites the concept)13:05
philiKONthe functionality is there, as i said13:05
ChrisWyes, but it seems everything is geared up to provide stuff globally: skins, etc13:05
philiKONjust a nice UI isn't13:05
ChrisWwhere would I look?13:05
philiKONfeel free to write one :)13:05
ChrisWhow should I wire it up?13:05
ChrisW(btw: I have finally started writing Swishdot now... all sorts of fun to be had :-D)13:05
mgedminChrisW, schooltool *used* to set the skin for a site in a traversal subscriber13:06
ChrisWis there a style guide for what macros should be provided/slots filled/etc in order to be compatible with the Zope 3 style?13:06
philiKONi still remember your talk from EuroZope barbecue in 2001 or so where you "introduced" swishdot13:06
mgedminthen we decided to scrap the 'SchoolTool application as a content object inside a Zope 3 folder' use case13:06
ChrisWalso, how far along is the "viewlet" concept?13:06
mgedminand specified the default skin in zcml13:06
ChrisWah, okay13:06
mgedminif you're interested, you can find the old subscriber in subversion's history13:06
ChrisWso schooltool can no longer easily co-exist with other Zope 3 applications... that's a shame :-S13:07
philiKONwell, that's not exactly true13:07
philiKONit just assumes certain defaults13:07
philiKONyou are free to change them :)13:07
ChrisW(and, fwiw, the original Swishdot code has been in CVS for 5 years or so now ;-) )13:07
philiKONanyways, there's a difference between setting the default skin and setting the skin dynamically upon traversal or some other funky way13:08
ChrisWphiliKON: what are you talking about there? (too many cross conversations ;-) )13:08
philiKONi'm trying to explain to you what "default skin" means13:08
philiKONwhen a request is made, the default skin is applied to it13:08
ChrisWwell, I think it'd be pretty important to be able to set the default skin at a site-level13:08
philiKONso, unless something else in the request's life changes that, the default skin applies13:09
ChrisWthink of one Zope instance serving out 3 or 4 Squishdot sites...13:09
philiKONyes, yes13:09
ChrisWso what's the "right" way to do that?13:09
*** andrew_m has joined #zope3-dev13:09
philiKONlemme finish, goddam13:09
philiKONnow, a skin is an interface that's registered as ISkin and with the name we know it as  (e.g. "Rotterdam")13:10
philiKONthe default skin is determined by looking up the default skin *name*13:10
philiKONwhich works somethign like this:13:10
philiKONah, mmh, i take that back13:11
philiKONwe actually *do* look up the skin itself13:11
philiKONso, anyways, a UI that would do this at site level would simply register the skin interface as an adapter13:12
philiKONfrom IBrowserRequest to IDefaultSkin13:12
philiKONof course, since interfaces aren't picklable right now, this can only apply to skin interfaces defined in python13:13
ChrisWokay, but how do registrations like that take place at the site level?13:13
philiKONthere's a registration API13:13
ChrisW"skin interfaces defined in python"?13:13
ChrisWand why aren't interfaces pickleable?13:13
philiKONeven if they were it wouldn't do good13:13
ChrisW(I saw persistent schemas mentioned in the ZODB, do they not work now?)13:13
philiKONright, i don't think they work13:13
philiKONeven if interfaces were picklable, you would persist an interface created in one run of the zope app and after a restart, ZCML would generate a new interface object which would then not be the same as the one sitting in the ZODB13:14
philiKONauto-generation of skins and layers just needs to end, i actually have a proposal for this that is just waiting to be implemented13:15
philiKONre: skin interfaces defined in python:13:15
*** nixnutz_ has joined #zope3-dev13:15
philiKONyes, ZCML is evil that way13:15
ChrisWbear in mind I've read the book, but apart from that, I'm still a big fat n00b ;-)13:15
ChrisWZCML is evil full stop13:15
ChrisW<rant coming>13:15
philiKONwell, if it were only about on/off switches, it wouldn't be so evil13:16
ChrisWso we've come up with yet another fucking language, great, complexity for complexity's saske, woohoo!13:16
ChrisWyes, but it isn't13:16
philiKONright, it isn't13:16
ChrisWthink how much ZCML you wrie in tha tbook13:16
philiKONit's not for complexity's sake13:16
ChrisWand most of that is what I'd call programming13:16
ChrisWwell, having it as another language is13:16
ChrisWespecially as it likely maps onto python methods and objects underneath13:16
philiKONif it'd be python, people start doing evil hacks13:16
philiKONthey start introducing 'if' clauses etc.13:17
philiKONwe don't want that13:17
philiKONjust like we don't want to use python for generating HTML13:17
ChrisWhave them behave like security proxies, where if people try and do evil hacks, it tells them not to13:17
* mgedmin sometimes fears zope 3 is doomed to be huge and complex forever13:17
* ChrisW is with mgedmin13:17
philiKONmgedmin, there are ways to make sure it won't be13:17
ChrisWthat said, I still have dreams of my cms which will hide the complexity unless you really need it...13:17
philiKONChrisW, anyways, to read about the status quo of skin interfaces etc., read the beginning of
ChrisWthe constant use of adapters does worry me from a performance point of view13:18
* andrew_m dreams of a layer between zope3 and a CMS that simplyfies stuff..13:18
ChrisWevery single thing seems to involve insantiating a class and then calling it13:18
philiKONandrew_m, well, why not simplify zope3 itself13:18
philiKONChrisW, so?13:18
ChrisWwhile that's not hugley expensive, it's also not trivially cheap ;-)13:18
*** MJ has joined #zope3-dev13:19
ChrisWzope3 needs it's complexity to meet everyone usecases and be ultimately felxible13:19
philiKONbut i think there are ways it can "scale down"13:19
ChrisWhowever, _most_ people don't need to be exposed to that13:19
philiKONe.g. with the revival of bobo13:19
ChrisWI am very happy to see how things have come together13:19
ChrisWand yes, I can see how only little bits of it now _need_ to be used, which is great13:19
ChrisWI can see several mini-frameworks spinning off that do simpler things for what people need13:20
ChrisWand yet still allow them to side-grade to "full zope 3" if they really want it all13:20
philiKONwell, yes, the idea of z3 is to be modular13:20
philiKONi think with our making smaller, it'll be even more modular13:21
ChrisWI can also see frameworks like Plohn and the one I'm brewing havign a full Zope 3 source tree but also providing much more targetted and flexible functionality for, say, a CMS13:21
ChrisWstill Swishdot's going to make an interesting little experiement for me13:21
philiKONyeah, would also be a nice demo app13:21
ChrisWoh, tht reminds me, two big topics I need to dig on:13:22
philiKONi'm sure people would be willing to contribute, too, just to be able to play with z313:22
ChrisWwhere should I look? what should I read?13:22
philiKONerr, sorry13:22
philiKONfwrong window13:22
andrew_mjust a quicky: does someone know how i can get this very trivial doctest to work (i.e. "..." at the beginning of output):
philiKONChrisW, the doctests13:22
philiKONChrisW, btw, are you not getting my /msgs?13:22
ChrisWI am, are you not getting mine?13:23
ChrisWoh, maybe not13:23
ChrisWfucking freenode bullshit13:23
philiKONnope, you're not registered then13:23
andrew_mChrisW: you have to register / identify13:23
philiKONandrew_m, #doctest: +ELLIPSIS13:24
mgedminChrisW, the README and the source code of zope.wfmc are pretty good for understanding workflow13:24
ChrisWokay, and for cataloging?13:25
mgedminandrew_m, I don't think you can -- the '...' is parsed as a continuation line; you must start the expected output section with something else13:26
andrew_mphiliKON: ok, thanks. but it still refuses ellipses at the very beginning (first line of output)13:26
ChrisWphiliKON: docs?13:26
andrew_mmgedmin: yeah :/13:26
mgedminandrew_m, in the worst case you can always do '>>> print "-"; do other stuff'13:26
ChrisWand, btw, I'm wary of doctests as documentation, they tend to end up only making sense if you're the author of them ;-)13:26
andrew_mmgedmin: heh, ok, that is an idea13:26
mgedminChrisW, documentation with doctests inside is great (working examples! an example is worth a thousand descriptions)13:27
andrew_mChrisW: hmm.. i have leaned a lot from doctests though13:27
mgedmindoctests that test the internals/corner cases/etc are useless as documentation13:27
mgedmindoctest is a tool with many uses13:27
mgedminI count at least 413:28
mgedmin1) short function usage examples in function docstrings13:28
mgedmin2) long narrative-style developer documentation about the use of a package/module13:28
mgedmin3) unit tests13:28
mgedmin4) API design tool -- write some science fiction, try to make it into a doctest, then implement the code so that the test passes13:29
ChrisWand only 2 is useful if you're in my position13:29
ChrisWI have to admit to being annoyed by 1, since most people don't really get the meaning of "short"13:30
ChrisWand doctest setup usually makes the docstring anything but short ;-)13:30
mgedminit takes a while to get13:30
andrew_mChrisW: yeah.. 1) makes me nervous (too much blah inside the code)13:31
mgedminJim's PyCon doctest presentation, and Phillip J. Eby's blog helped me13:31
ChrisWI _like_ doctests in files13:31
ChrisWbut I don't think they're often documentation13:31
ChrisWzope.testbrowser was a good example for me ;-)13:31
philiKONsandwich time!13:33
*** J1m has joined #zope3-dev13:44
*** AJC has joined #zope3-dev13:49
*** mkerrin has joined #zope3-dev13:49
*** ChrisW has left #zope3-dev13:49
*** J1m has quit IRC13:53
*** stub has joined #zope3-dev13:54
*** marwin has joined #zope3-dev14:30
marwinhi guys. is this list for technicans or users ?14:32
philiKONyou can ask questions about zope 3 here14:32
marwincool :)14:33
marwini used zope 2 before. now i have a look to zope 3... a very basic thing I nowhere found explained is: What is the difference between default and tools in ++etc++site ?14:34
philiKONnothing really14:35
philiKONjust a different site management folder14:35
marwinso if I add something to default, it is exactly the same as I would add it do tools? there nothing different like visibility, local/global or something?14:36
marwinok :)14:37
marwinoups, just realised that I created 'tools' by myself... I thougt it is also a default folder.... sorry about that.14:52
*** roym has joined #zope3-dev14:57
roymin z3, is it possible to pack the database from the GUI?14:58
roymI see a test: >>> ZODB.tests.util.pack(db)14:58
Theunicheck the process management14:59
Theuniit might be there14:59
Theunii'm not sure though14:59
philiKONroym, localhost:8080/++etc++process15:02
roymI love how compact it is (vs Zope2) after packing.15:02
*** tonico has quit IRC15:16
*** _anguenot has quit IRC15:17
*** stu1 has joined #zope3-dev15:24
*** stub has quit IRC15:24
*** tonico has joined #zope3-dev15:28
*** elbixio has joined #zope3-dev15:28
AJCdoes it make sense to use rotterdam as a basis for a new skin, even if this new skin is extremely simple?15:29
philiKONperhaps not15:30
philiKONusing rotterdam as a base skin is a pain15:30
AJChow so?15:30
*** eins has quit IRC15:30
philiKONit's inflexible15:31
*** _pcardune_ has quit IRC15:33
*** pcardune_ has quit IRC15:33
AJCok, good enough for me :-)  thanks.15:33
*** stu1 is now known as stub15:33
AJCis the difference between CPS and 3lab that the first uses Zope 2 and the latter Zope 3?15:35
*** retsu has joined #zope3-dev15:36 i meant15:36
*** tonico has quit IRC15:39
philiKONAJC, runs CPS is15:40
philiKONAJC, runs CPS15:40
AJCok, i was a bit confused by the relationship.  so the next-gen open source ECM that z3lab wants to make is CPS?15:41
*** benji has joined #zope3-dev15:41
jhauserz3lab is not about a product or a specific framework15:42
philiKONAJC, z3lab is just a site15:42
philiKONAJC, it's not really a project15:42
jhauserz3lab at the moment collects components15:42
jhauserit represents somewhat the z3ecm project15:42
*** tonico has joined #zope3-dev15:42
jhauserthe status of which I also do not know15:43
philiKONthere's no status because there's no real project15:44
AJChmm, i think i see :-)  maybe you should have a small paragraph explaining this + linking to CPS, i had no idea about CPS until yesterday...15:45
*** jinty has joined #zope3-dev15:46
*** J1m has joined #zope3-dev15:47
philiKONAJC, so we should also have a link about Plone and Silva? seems a bit unnecessary15:47
*** zbir has quit IRC15:50
AJCi see your point.15:52
*** retsu has quit IRC15:57
*** zbir has joined #zope3-dev16:01
*** tav has joined #zope3-dev16:02
*** j-w has joined #zope3-dev16:03
marwinwhat is the simplest way to create a user login which can do the same as the default admin user?16:08
marwinis there something like a user_folder ?16:08
philiKONcreate a Pluggable Auth Utility16:10
philiKONcreate a principal folder inside16:10
philiKONcreate a principal16:11
philiKONthen, grant that principal the zope.Manager role16:11
marwinhm. what is the prefix? or where is something like that documented?16:15
*** stub has quit IRC16:16
philiKONmarwin, APIDoc, documentation, books...16:17
*** jukart has joined #zope3-dev16:17
marwinphiliKON, did you write "Web Component Development with Zope 3" ?16:18
marwinwe ordered it yesterday :)16:19
andrew_mmarwin: good idea.. helps a lot especially in the beginning16:19
J1mSome things I want in the next release cycle:16:30
J1m- Move as much as possible out of zope.app16:30
J1m- Much simpler local component management.16:31
J1m- Better publishing APIs.16:31
J1mRethink/do repo/packaging w eggs16:31
*** retsu has joined #zope3-dev16:38
philiKONJ1m, two questions16:47
benjiI'd suggest we make Z3 compatible with Paste, and thus get egg support16:47
philiKONa) will the moving out of be only mechanical (meaning, the 'app' in the import path will disappear), or will it basically be large refactorings as well?16:48
philiKONb) how simple will the local component management be?16:48
philiKONdo we know yet?16:48
J1ma) It will be mechanical, except that it is motivated by making things more modular and is likely to inspire further refactoring.16:50
J1mb) :16:50
J1m- Get rid of registration managers16:51
J1m- components can be registered either from anywhere. This means site-management folders will be less magic. They will simplty provide a convenient way to keep things out of content space.16:52
J1m- From python, components will be registered more or less as they are now with the global component registry.16:52
philiKONthat sounds great16:53
philiKONquestion about the first bullet point:16:53
J1mSo, for example, creating a local utility will be as simple as calling provideUtility on some local thing.16:53
philiKONhow will registrations be managed?16:53
srichterwhich brings me to the point that we should really rid ourselves of the old local CA BBB code16:53
J1mInternally by the component manager (currently called site manager)16:54
srichterit will be hell to write new BBB code otherwise16:54
philiKONJ1m, i see. so basically, registration managers are folded into site managers16:55
philiKONother than that, this is exactly what i had in mind for CMF 2.0 local components16:55
philiKONbeen meaning to put that into my draft proposal16:55
philiKONthe one i'll work on at pycon16:56
philiKONperhaps we can take some time to brain storm on this16:56
philiKONso the architectures will be similar16:56
philiKONmeaning, brainstorm on it at pycon :)16:56
*** retsu has quit IRC16:57
* jinty would really like to be involved in any discussion about zope+eggs16:58
jintyas it has the potential to make debian packaging much easier, or harder16:59
philiKONbenji, interesting suggestion. i need to read up on paste16:59
benjime too :)16:59
philiKONjinty, which reminds me, i've always wanted to write this blog entry why eggs and tradition packaging formats like debs and rpm don't contradict...17:00
philiKONjinty, i think it'll make debian packaging much easier17:00
mgedminphiliKON, +1 for writing it17:00
jintyphiliKON: eggs and packages were bad for each other until we got the --single-version-externally-managed shoot myself in the foot option17:01
philiKONwhat's --single-version-externally -managed?17:02
jintyit's a little option for setuptools that does not build eggs17:02
philiKONso, what does it build?17:03
jintybut exploded directories maintaining the egg metadata17:03
philiKONbut why can't a deb just install an egg?17:03
philiKONi mean, debs install JAR files too17:03
jintyI don't think it's necessary17:04
jintybut I am not sure of the exact reasons17:04
jintybut here goes anyway17:04
jintyan egg is a single file, so un-installing is easy, i.e. you dont have to keep track of the files17:05
jintybut a debian package keeps track of the files, and make sure they are un-installed correctly17:06
philiKONyeah, but can a debian package with single files actually manage separate versions of something17:06
jintyi don't understand the question17:06
philiKONwith eggs, you can install more than one version of the same package17:07
philiKONe.g. you can install zope 3.1 and 3.2 in sites-packages17:07
jintybut only one is active at a time?17:07
jintyi.e. import zope will only import one of those, right?17:08
philiKONwell, if my personal package says "require("Zope <= 3.10")", then eggs will figure out and put zope 3.1 into the pythonpath17:08
benjijinty, you can use the -m (IIRC) option to have multiple eggs "active" (but I'm not sure about the details, I'm an egg newbe)17:08
philiKONright, multiple ones can be active17:08
philiKONand it decides from the metadata whidch one to use17:09
jintyI think what will happen in this case is that the eggs in debian packages will contribute one of thse versions17:10
jintybecause they have the metadata17:11
jintybut I am rapidly getting out of my depth here17:11
jintythere is a huge thread on debian-python that goes through most issues17:13
jintyif you can dee through the flames (me is embarrassed at the behaviour of some debian developers)17:13
philiKONheh, yeah17:14
philiKONexactly the point of "eggs vs. tradition packaging"17:14
jintyeggs try to solve a problem that was already solved in debian for many years17:15
jintythat gets people irritated17:16
philiKONbut it solves more even17:16
jintystarts here:
benjiunfortunately the "all the world is debian" attitude doesn't help much here  :)17:17
mgedminyeah, some of us use ubuntu17:18
* mgedmin ducks and runs17:18
srichterbenji: I agree17:19
srichterI think it is very important for Python to have a solid dependency/packaging system independent of any OS17:20
philiKONjinty, holy jeebus, what a thread17:20
dobeehi all, is ther some kind of timezone change to zope 3.2 b3? on a fresh instance when i create a folder the dates i see in the browser does not match the ones i get in the log. the log dates match my machine17:21
dobeemy machine: Darwin bd-mb 8.3.0 Darwin Kernel Version 8.3.0: Mon Oct  3 20:04:04 PDT 2005; root:xnu-792.6.22.obj~2/RELEASE_PPC Power Macintosh powerpc17:21
jintybenji, srichter: I agree, but not actually breaking debian (or ubuntu) would be nice as well17:22
srichterof course17:24
srichterbut if there are conflicting recommendations/policies you have to choose one17:24
benjiI was just noting that the scope of our considerations is much wider than that of the debian team (non-debian linux, Windows, Mac, etc.)17:25
J1mphiliKON, I want Z2 and Z3 to manage local components the same way, ideally with the same code.17:29
philiKONme too17:29
J1mThere are all sorts of other things I want for Z2. :)17:29
philiKONbut we want something like this in cmf 2.x17:30
philiKONwhich might only require zope 2.9/3.217:30
philiKONor perhaps not17:30
philiKONlots of tings to brainstorm at in dallas17:30
*** retsu has joined #zope3-dev17:31
*** _anguenot has joined #zope3-dev17:33
philiKONg'night y'all17:33
romanofskinighty philiKON17:33
*** jinty has quit IRC17:40
mgedminstylistic question: should doctest sections in a ReStructuredText file (*.txt) be indented with 2 spaces or 4?17:42
* mgedmin prefers 417:42
* mgedmin 's editor prefers 4 would be a bit more accurate17:42
benjimgedmin, 417:43
* philiKON prefers 2 space indention in all reST files17:43
*** marwin has quit IRC17:43
philiKONmeaning, the >>> is on the 3rd column17:43
mgedminI know17:44
mgedminin a preformatted section::17:44
mgedmin    if there is python code:17:45
mgedmin        it is better to keep all indents a multiple of 4 spaces17:45
mgedminotherwise it looks strange17:45
mgedmin    >>> and I want to align all indented blocks the same way17:45
benjithat's one reason why I prefer 4, that's how the code will be indented17:46
philiKONso, you indent quotes 4 spaces as well?17:46
benjiI do17:46
J1mI indent >>>s 4 spaces.17:46
philiKON*sigh* ok :)17:46
*** jhauser has quit IRC18:01
*** sashav has quit IRC18:05
*** jhauser has joined #zope3-dev18:05
*** dobee has quit IRC18:11
*** romanofski has quit IRC18:16
*** mgedmin has quit IRC18:21
*** MacYET has joined #zope3-dev18:22
*** andrew_m has quit IRC18:23
srichterphiliKON: I prefer 2 :-)18:25
*** jukart has left #zope3-dev18:26
*** zagy has quit IRC18:27
benjiCardinal Biggles, prepair the comfy chair for the heratic, srichter!18:30
*** dobee has joined #zope3-dev18:31
* MacYET hands some soft cushions18:31
* benji searches franticly for... THE RACK!18:32
*** natea is now known as natea|away18:32
*** andrew_m has joined #zope3-dev18:34
*** elbixio has quit IRC18:45
*** mgedmin has joined #zope3-dev18:49
*** efge has joined #zope3-dev18:49
efgeJ1m: I saw your email. Actually I just committed my temporary fix in I'll remove it if needed.18:50
J1mI'm just about to send a second message.18:51
J1mDo you know why the temporary storage isn't part of the multidb to start with?18:51
efgeJ1m: yes, it's opened by the mountpoint traversal, and it doesn't add it to the multidb connection set18:52
J1mI'm not asking about the connection.18:52
J1mJust ream my mail ...18:52
J1mJust read my mail ...18:53
efgethat's a problem only the first time it's done in the zope life, when the DB for the temporarystorage hasn't been created yet18:53
J1mWhy isn't the temporary database used by the session machinery part of the multiedb to start with?18:55
efgeJ1m: ok I understand the question18:55
J1manyway, aside from not having enough tests, I don't think there is a ZODB bug here.18:56
efgeI don't know why, something in the DBTab layers fails to register it18:56
efgeJ1m: agreed18:56
J1mSo I'll tell Tim to proceed with 3.6 final so I can proceed with 3.2 final.18:57
MacYETwhen you expect 3.2 final?18:58
MacYETwhen do you expect 3.2 final?18:58
efgeJ1m: hm the database isnt't added to the multidb until it's opened18:58
efgeand it's not opened until the mountpoint traverses it18:58
efgedatatypes.DBTab.getDatabase() does that18:59
J1mSounds like a DBTab bug.19:00
*** mgedmin_ has joined #zope3-dev19:00
*** MJ has quit IRC19:00
J1mIs it possible for that code to add the database to the multidb and then retry get_connection?19:01
efgeJ1m: should dbtab open all registered databases at startup?19:01
efgeJ1m: that could work yes19:01
J1mYes, it should19:02
J1mBut if that's too hard to change, then a sane runtime approach would be fine.19:03
J1mWhere sane is as described above.19:03
efgeJ1m: ok I'll try that and check it in if it's working19:04
J1mCool. Thanks!19:04
efgeprobably not today though19:04
*** xenru has joined #zope3-dev19:05
efgeJ1m: DBTab.getDatabase has an is_root arg which passed by zope's startup code but never use19:11
efgewe could open all databases at that point if is_root is true19:11
J1mDoes it know about all of the databases at that point?19:11
J1mI've always thought that DBTab should *really* be like fstab.19:12
efgeyes that's after configuration has been loaded, the mount_paths dict has been filled19:12
J1mfstab doesn't define devices. It just configures how they're used.19:12
MacYETjim: where is the generations package?19:13
J1mIN the long run, I'd prefer that the multi-databases be defined outside of DBTab and have DBTab just specify where databases (by name) get mounted.19:13
J1mOf course, that should move out of :)19:14
efgeJ1m: quick ZODB question: I don't have a clear understanding of the difference between a tid and an object's serial, are they the same?19:16
*** mgedmin has quit IRC19:16
J1mThey can be different.19:16
*** zagy has joined #zope3-dev19:16
J1mThey are usually the same, but they aren't guarenteed to be the same.19:17
*** mgedmin_ has quit IRC19:21
*** mcdonc_ has joined #zope3-dev19:27
efgeJ1m: would there be a problem if a storaged defined serials independently of tids? Are they ever compared or one used for the other?19:31
*** MJ has joined #zope3-dev19:32
*** natea|away is now known as natea19:35
*** tarek has quit IRC19:39
*** retsu has quit IRC19:42
AJCsee you guys!19:42
*** AJC has quit IRC19:42
*** zagy has quit IRC19:55
*** deo has quit IRC19:59
*** j-w has quit IRC20:04
*** mcdonc_ has quit IRC20:05
*** efge has quit IRC20:06
*** nixnutz_ has quit IRC20:09
*** regebro has joined #zope3-dev20:21
regebroHiya all!20:21
regebroI'm a bit confused.20:21
*** vlado has quit IRC20:22 imports from zope.component.tests.placelesssetup.20:22
regebroBut to the best I can figure out... it's deprecated!?20:22
regebroOK, just a miss I guess, then.20:25
regebroWhy on earth is placelesssetup starting with doing a cleanup!? Sigh.20:34
regebroMy tests first loads some stuff, then installs all necessary products, the does a cleanup and then a load_site. Then placeless setup does a new cleanup, meaning I then need too, once again, do a new load_site.20:35
regebroAnd I'm not even sure THAT works. :-/20:35
regebroOK, that seems to work at least. :-)20:38
*** mgedmin has joined #zope3-dev20:40
*** deo has joined #zope3-dev20:46
*** tiredbones has quit IRC20:55
*** tiredbones has joined #zope3-dev21:01
*** Aiste has quit IRC21:23
*** MacYET has quit IRC21:32
*** agroszer has quit IRC21:40
*** dman13 has joined #zope3-dev21:44
*** ignas has quit IRC21:47
*** Aiste has joined #zope3-dev21:56
*** mkerrin has quit IRC22:11
*** TwistedBen has joined #zope3-dev22:21
*** MacYET has joined #zope3-dev22:34
*** MacYET has quit IRC22:40
*** _anguenot has quit IRC22:40
*** regebro has quit IRC22:44
*** tarek has joined #zope3-dev22:48
*** romanofski has joined #zope3-dev23:17
*** jenner has joined #zope3-dev23:18
benjihi, jenner23:25
jennerhi benji23:27
jennerDoes anyone have a decent example of formlib usage?23:28
*** dman13 has quit IRC23:32
*** alga has quit IRC23:35
benjijenner, the formlib doctests are good, read the first section carefully though, because it points to the easier-to-use section later in the document23:36
jennerbenji: thnx23:36
*** romanofski has quit IRC23:40
*** romanofski has joined #zope3-dev23:41
*** romanofski has joined #zope3-dev23:42
*** efrerich has joined #zope3-dev23:44
*** jinty has joined #zope3-dev23:47
*** sashav has joined #zope3-dev23:49
*** jenner has quit IRC23:57
*** zbir has quit IRC23:57

Generated by 2.15.1 by Marius Gedminas - find it at!