Platform Team Meeting 25 April 2006

Agenda

• Question of the week: What would make a good "summer of code" project?
• 0.7alpha2
  • feature freeze: in two days (April 27th)
  • release: May 11th
• 0.7alpha3
  • how is the plan looking for everyone?
  • eggs -- porting parcels
• pyicu and pylucene patches -- do we need CLA's?

Notes

• Question of the week: What would make a good "summer of code" project?
  • Heikki: tinderbox work, bonsai work, m2crypto summer of code -- would they be mentoring
    • automated wx gui tests
  • Grant: activity viewer ui, contacts or tasks
  • Bear: natural language date parsing
    • python framework for natural language framework (natural language toolkit)
  • Ted: pile of file converters (pick your favorite file format)
    • open office
    • now up to date
    • outlook pst
  • Phillip:
    • rewrite of generic function engine he prototyped in prep for python3
      • extend the state of the art -- a student could get a decent paper
    • package management tool for setuptools
      • yum or similar type of package management interfaces
      • command line tool that could be the basis to run from a menu in chandler
      • may be more of a developer tool
      • useful for talkback?
    • make it possible to have dynamic linked libraries for eggs (might fall off of 0.7 list)
  • Brian K:
    • wrap native os html renderers accessible via python
    • language parser (cool needed feature for a calendar app that will get props and attention)
• 0.7alpha3: what is missing from the table?
  • i18n regression testing plan and developer check in plan
    • we can write a functional test at the sprint
  • domain model changes to support performance -- make sure the filters can work without loading the items
    • someone on the apps side needs to own some of this work
  • indexing and search probably will be a topic in alpha3
    • indexing is a cause of performance regressions lately
  • talkback
• cosmo dependencies/requirements
  • sharing format
    • cosmo will need to be able to parse it
    • an improved sharing format will help us get away from two resources per item
  • free/busy
    • cosmo does not support the full freebusy report in 0.3
    • 0.4 will support the report
    • alpha3 will need it!!!
• porting parcels
  • should we make the repository an egg? (with chandlerdb)
    • benefit: would help if we wanted to distribute the repository outside of chandler
    • infrastructure is not quite ready, but we should do this when it is ready
  • grant's isync project is an egg
  • is it a goal to make all components eggs? yes, but we haven't specified a timeframe
    • benefit to migrating more components to eggs: nice to have to know about dependencies
  • eggify vobject, zanshin?
• CLA's for PyICU^?, PyLucene
  • we have a CLA -- Ted will send a note to general about the CLA philosophy
  • prove that people who contribute to code have the right to do so
  • Heikki will need it for m2crypto
  • hazmats unit tests fall into the category of work we'd want a CLA for (might have one)
• inactive people -- project can revoke commit access -- commit accesses not tied to employment
  • commit is a function of contribution, not employment

Status

Katie

• Progress
  • morgen and sheila wrote posts about 0.6 --> alpha2 migration (so I didn't need to)
  • stamping conversation with alec and bryan
  • chandler-dev summary
  • http://wiki.osafoundation.org/bin/view/Journal/CoreSprintWeekSpring2006#What%20(agenda)
• Plan
  • stamping writeup
  • sharing format
  • start collections spec
  • start interop spec

Brian K

• Progress
  • Cosmo - memory / I/O profiling for .3
  • Sent i18n road map email
  • Sent various i18n design related emails
  • fixed three alpha 2 bugs related to mail. One is still pending peer review before checking.
  • Had employee review with Lisa
  • Initail research on flights and hotels for Bitcon in June

• Plan
  • Finish outstanding Cosmo 0.3 issues
  • Write up thoughts on Email in Chandler. Can we host our own server?
  • Grab Alpha 2 bugs from the list
  • Regression test i18n features set in Chandler
  • Book trip to Taiwan for Bitcon including Hotel, Airfare, and Conference registration fee

Grant

• Progress
  • 5710 right click continues after ctrl is lifted on Mac
  • 5715 "Go to date" should handle out of range dates
  • 5598 Recurrence and timezone pref don't deal with edge cases
  • 5747 TimeZoneInfo.defaultInfo should get set up in installParcel(), not lazily
  • Resolved a couple of bugs as DONTFIX, e.g. 5673.
  • Isolated a BadRefError triggered by the fix to 5598, fixed by AndiVajda.
  • In process: 5604, 5371, 5112.
  • Phone screened a summer intern candidate
  • Actually finished my peer reviews.
• Plan
  • 0.7a2 bugs
  • Post-alpha2: Write up a chandler-dev@ email for 5407 Alarms should pop up when Chandler is not running

Ted

• Progress
  • Bug:3531
  • Closed discussion on Bug:1745 - deferred to Alpha3 due to scope of changes, including impact on sharing compatibility
  • Organizing Google Summer of Code
  • Portal Taxonomy
• Plan
  • 3 Days PTO
  • Organizing Google Summer of Code
  • Portal Taxonomy

Morgen

• Progress
  • Working with Andi on view merging as it relates to sharing
  • Fixed the Flickr test to not consider timeouts a test failure
  • Fixed sharing issues caused by switchover to Text 'body'
• Plan
  • Re-enable the username/password dialog for subscribing to shares with no matching account
  • Background sync testing
  • Clean-up new sharing code
  • Add facility for shutdown-time sync interrupt
  • Add an interrupt menu item

PJE

• Progress
  • Enhance setuptools to support dynamic choice to build eggs from local source if available (to support a variant of Heikki's Dev-list proposal re: automatic egg builds in Chandler 0.7a3)
  • More work on TestStartups to ensure cascading failures on OS X don't result in tinderbox hangs
  • Check in setuptools to Python 2.5 trunk
  • Have multi-day argument with folks opposed to including setuptools in the Python 2.5 trunk (who apparently didn't pay any attention to Guido's previous call for comments and displaying it in foot-high letters on his PyCon slides, and therefore waited until I'd already done all the integration work to complain...)
  • Take setuptools back out of the Python 2.5 trunk
  • Began writing additional docs requested by Python-Dev to address the more reasonable objections
• Plan
  • Finish the aforementioned setuptools docs, and try to put together a new plan for setuptools inclusion in Python 2.5a3 (May 27 deadline), or blow the whole thing off for a couple years to Python 2.6.
  • Release setuptools 0.6b1
  • Start writing the dump/reload proposal, after a little more thinking

Andi

• http://kumu.osafoundation.org/~vajda/logs.cgi