transcript of #zope (2004-10-19 11am-12am EST)
ZCBrian: I'm standing in for Rob today as MOC / guy-with-the-stopwatch
ZCBrian: Agenda items from the email:
ZCBrian: - Update on Zope 3 (Jim Fulton, 10 minutes)
ZCBrian: - Sprint reports (Jim Fulton, 15 minutes)
ZCBrian: - Zope.org docs update (Mark Pratt, 15 minutes)
ZCBrian: - Zope Bug Days (Chris Withers, 10 minutes)
ZCBrian: - CMF Update (Tres Seaver, 10 minutes)
ZCBrian: So with that, I'll cede the floor to the gentleman from across the table ;)
J1m: Update on Zope 3:
J1m: Still working toward initial (Zope X3 3.0) release. We've had 3 betas
J1m: and 2 release candidates, but errors keep trickling in. Sigh. We'll
J1m: do a 3rd release candidate this week and, hopefully, a final release
J1m: next week.
J1m:
J1m: We (the developers) need much more help from the community with
J1m: testing alpha and beta releases if we want to have more frequent
J1m: releases.
J1m:
J1m: New features and refactoring are happening on the trunk. Stephan
J1m: would like to make an X3.1 release in December or January.
J1m:
J1m: Questions? :)
ChrisW: what refactoring is happening?
ZCBrian: (bet nobody knew Jim could type that fast... ;)
ChrisW: what new features are being added?
J1m: I don't have a list handy ...
ChrisW: hmmm, I guess underlying both of those is "how will I move my X3.0 app to X3.1 when it becomes available?" :-S
J1m: http://svn.zope.org/Zope3/trunk/doc/CHANGES.txt?rev=28211&view=log
J1m: X3.1 should be backward compatable with X3.0.
ChrisW: will it though? ;-)
J1m: yes
ChrisW: :-)
J1m: This is a good lead in for a current standard speach of mine:
J1m: For both Zope 2 and Zope 3, we need to do a better job of:
J1m: - Backward comptability
J1m: - Managing change
J1m:
J1m: This means:
J1m: - We really make changes backward comptable
J1m: - We issue deprecation warnings for "old ways" of doing things
J1m: (e.g. old apis), with specific information on when old way will go
J1m: away
J1m: - When times up, we remove backward comptability code
ChrisW: (and it wasn't even scripted ;-) )
J1m:
J1m: Frankly, I'm a little bit worried about 2.8 in this regard.
J1m: Lots of people have been helping with 2.8, but I fear that, in many cases, the esponse to backward incompatabilities has been to change affected "third-party" code, rather than add backward comptabilty code to 2.8.
J1m: I haven't had enough bandwidth to track this.
J1m:
J1m: We need to get the messages out that:
kosh: overall I am okay with fixing my code to use new features so long as it is clearly documented about what is being removed
J1m: 1, Backward compatability (w deprecation) is important
kosh: and how it needs to be written now
J1m: 2, We need a lot more help with testing
kosh: and so long as the feature itself is not removed
ChrisW: also, deprecatation warnings are only useful if you can tell where they're coming from
J1m: IMO, you should have time to do this.
ChrisW: (eg: NOT like the search index stuff)
J1m: Hm, deprecation warnings generally have tracebacks.
kosh: and if I feature is going to be removed entirely I want to know about it long before it goes in
kosh: and have it clearly documented so I can figure out ways to work around
TresEquis: kosh, that is what a deprecation warning is
J1m: One place I'd like to deviate from Python norms is to make deprecation warnings more verbose.
TresEquis: it says, "This feature is deprecated, and will be removed in version 2.9"
J1m: right
TresEquis: e.g., the verbosity could be pointers to the URL describing the change
kosh: well I had heard that implicit acquisition is largely going to go away which is going to create so much work to fix such that it may be easier to just not bother upgrading
J1m: Although, saying when is important and not usually done for Python itself.
ChrisW: J1m: the probem with the search index ones was that yes, it said where in the code, but ti didn't tell you which object
ChrisW: which was arguably much more important...
J1m: ChrisW, I don't see what the object has to do with it.
ChrisW: so you can get to it and stop it using searchindex ;-)
J1m: kosh, Zope 2 to Zope 3 transition is an entirely different issue.
ChrisW: (like an index burried in the help system's catalog, which certainly isn't obvious ;-) )
J1m: Ah, good point
TresEquis: ChrisW: put a raise "DEBUG" at the top of SearchIndex/__init__.py
J1m: So, deprecation also has to provide tools for fixing data, where necessary.
kosh: in the warning I agree that the full path to the object in question wouuld be very useful
ZCBrian: What if we had a deprecated page on zope.org somewhere so
ZCBrian: that the warning in the code could be brief and point you
ZCBrian: to www.zope.org/deprecated#something
J1m: We've begun making progress in this area with the "generations" framework in Zope 3.
TresEquis: Sure
ZCBrian: then the page could be as verbose as required
kosh: how about deprecated/something
J1m: In general, if a change affects data, we need to provide conversion scripts.
ChrisW: TresEquis: indeed, but most people just ignored the deprecation warning instead, which was easier ;-)
kosh: that way a search for the warning will give you a page that is a closer match to what you are searching for
J1m: Zope 3 has a framework for this now, which will, eventually be ported to Zope 2.
kosh: also as you get more and more things deprecated the page would get very long
ChrisW: ZCBrian: yeah, but that's asking for forever-available urls on zope.org
ChrisW: that can't be a good idea :-S
markbee: nope
ZCBrian: DING! Sprint reports (Jim Fulton, 15 minutes)
J1m: We need a new ZODB feature (planned for ZODB 3.4) to allow us to iterate over all object of a certain kind.
J1m: OK, sprints
J1m: The most recent ones were the castle sprint:
ChrisW: J1m: your new ZODB feature would be unbelievably useful and cool :-)
J1m: http://dev.zope.org/Wikis/DevSite/Projects/ComponentArchitecture/CastleSprint2004
J1m: and the ISAR sprint:
J1m: http://dev.zope.org/Wikis/DevSite/Projects/ComponentArchitecture/IsarSprint
J1m: The castle sprint was less of a sprint than a meeting
J1m: It was an opportunity for a lot of people making CMSs for Zope to get together and do some planning for Zope 3.
J1m: Some work also got done, although this was mostly prototyping or work on the Xickens demo CMS.
J1m:
TresEquis: s/CMS/CM application/ ;)
J1m: The ISAR sprint was a typical sprint except for the fact that it was twice as long (6 days) as most sprints.
J1m: Quite a lot gone done there.
J1m:
J1m: Future Zope 3 sprints that I'm aware of will be a sprint in conjunction with PyCON 2005 and EuroPython 2005.
J1m: I'd also like to have ZODB sprints at these meetings, if there is interest.
J1m: (In general, I'd like to increase community involvement in developing ZODB.)
J1m:
J1m: One thing I'd planned to work on at the PyCon sprint was Zope/Twised integration.
J1m: In particular, I'd like to allow Twisted to be used instead of ZServer.
J1m:
kosh: that I would like
ChrisW: oo, cool...
kosh: I would love to have it so that news things like slashdot could also be accessed via nntp
kosh: and twisted would work with that
J1m: Interestingly, PEP 333, WSGI, provides a short-term opportunity for using both Twisted and mod_python.
J1m: (but just for HTTP.)
J1m:
J1m: Questions?
TresEquis: Anybody evaluated what ZPublisher would need to do to support WSGI?
TresEquis: I know Phillip Eby has done something on it
J1m: I imagine it wouldn't be too hard.
J1m: WSGI like ZPublisher/zope.publisher is based on a CGI model, so there is a close impedence match.
ZCBrian: DING! Zope.org docs update (Mark Pratt, 15 minutes)
markbee: ok
markbee: Let's summarize:
markbee: Sent around some graffles with proposals for the Zope.org Doc reorganization.
markbee: (Got very little feedback).
markbee: I moved to Barcelona and discovered that Telefonica is the worst quasi Telco monopolies -- after one month no phone line and no DSL :-(
markbee: (And yes all other Telco's resell their lines).
markbee: but this week I intend to add a few pages
markbee: and Morten and I will be moving over all of our demo pages to the Wiki that Jim
markbee: set up on Zope.org on Tuesday of next week
markbee: also plan to send out an email this week to the zope mailinglists calling for writers to help
markbee: write up some docs that i would like to put up
markbee: that explain a in more detail
markbee: how the zope community is organized and how they can get involved
markbee: there isn't much newbie friendly docs in that area
TresEquis: I know the 2.7 Zope Book effort could use some forward momentup
markbee: i haven't spoken to ChrisM
markbee: in more than a month
TresEquis: http://plope.com/Books/zb_signup
markbee: so I don't know where the problems are
TresEquis: still shows some unclaimed chapters
markbee: i know the URL :-)
TresEquis: posting for the channel
markbee: i'll be happy to help with that effort as well
markbee: but the first priority is to get the zope.org doc section reorganized
markbee: and make sure that people have a clear intro
markbee: to what kind of resources are in the community and how (if they so choose)
markbee: they can get productively involved
markbee: http://plone2.blogologue.com/Documentation/ContributingDocumentation/
markbee: was an early draft of
markbee: an overview of how people could contribute to various Doc efforts and resources
markbee: if anybody here would like to help
markbee: i recommend you either contact me via email
markbee: or even better read up in the archives
markbee: and look for a link to the latest graffle that i sent around
markbee: anybody have any questions?
markbee: otherwise i'd rather spend my time working on the wiki :-)
ZCBrian: going once... going twice...
ZCBrian: ok - DING! Zope Bug Days (Chris Withers, 10 minutes)
ChrisW: Well, not huge amounts to say really
ChrisW: we hae another bug day coming up at the end of this month
ChrisW: #zope-dev is the channel to be at
ChrisW: it's handy if you have CVS access and find out how to run the tests
ChrisW: but if you don't people can help out on the day
ChrisW: I see discussion over collector statuses stagnated, I have no problem with it re-opening on the alt wiki page
ChrisW: until then, I hope people will respect the status meanings on the current page
ChrisW: makes life a lot easier ;-)
ChrisW: apart from that, would be great to see as many of you as possible on the last friday of each month
ChrisW: questions? :-)
TresEquis: hold on a sec
TresEquis: klm might want to pipe in; I just jostled him
ChrisW: heh ;-)
ZCBrian: you jostled him too hard ;)
* TresEquis looks mortified
* ChrisW wasn't aware ken was so fragile :-S
* klm waves
* ChrisW jostles ken again :-)
markbee: ok
ZCBrian: CMF Update (Tres Seaver, 10 minutes)
klm: tres let me know that the collector status characterization discussion came up again -
klm: i dropped the ball after covering what i saw as the easy ones, intending to get back to it,
klm: and still intend to do so...
klm: i'm pretty sure i won't get to it in the next two weeks, and that i will soon after that.
klm: (anyone out there?)
ChrisW: yep, we're listening ;-)
ChrisW: that's fine, I think the discussion is totally worth having
ChrisW: but I'd appreciate it if we could stick with what's on the "current" page until those discussions are finished...
klm: what do you mean, "stick with what's on the current page"?
klm: are there some disagreements, or do you feel people should be handling things differently than they are?
ChrisW: nope, just clarity
TresEquis: I think ChrisW means to treat the "current" page as the policy being applied until the discussion settles out
ChrisW: yep
ChrisW: rather than people looking at the current page, the alt page, and then actually doing a 3rd thing ;-)
klm: if nobody who's been handling public issues objects, that's fine with me.
ChrisW: k
ChrisW: any more questions?
ZCBrian: Tres, any CMF updates?
TresEquis: Yup
TresEquis: Is it time?
ZCBrian: yup
TresEquis: CMF 1.5 will be out this week
TresEquis: Major new features include:
TresEquis: - CMFActionIcons included in the core
TresEquis: - CMFSetup, a new framework for managing site configuration as filesystem files (XML + scripts)
TresEquis: - CMFUID framework, with tools for controlling generation, assignment, and location tracking around UIDs
TresEquis: - Tons of API cleanup and bug fixing
TresEquis: Most of the effort has been community driven
TresEquis: Kudos to Yvo Schubbe, Jens Vagelpohl, Florent Guillaume, and Gregoire Weber, in particular
TresEquis: The Plone guys may release a 1.4.8, to support some fixes they want in the core
TresEquis: but otherwise, the 1.4 branch will no longer be the "supported release"
TresEquis: Questions?
TresEquis: I guess folks will have noticed that the dogbowl site is now on zope.org/Products/CMF
TresEquis: mostly the releases, but some docs
TresEquis: that's it for CMF updates, then
ZCBrian: ok - unless anyone has any good jokes, I think we're done ;)