Version 1 (modified by particle, 4 years ago)

notes from the parrot developer summit meeting

parrot virtual developer summit: 11 april 2010

* review of the 2.3 release

  • deprecation policy is better than before. some would prefer opt-in to opt-out
  • need better testing of packages before supported releases

* process questions

good:

  • shorter deprecation policy.
  • new weekly status meeting time works well.
  • gsoc works well for getting new parrot contributors.
  • weekly roadmap review is helpful, keep it going.

bad: we set too many weekly priorities. fix:

  • make weekly priorities specific tasks
  • completable by 1-2 people in a week
  • one won't be set unless 2+ people have volunteered to work on it
  • know this fix works if we complete a higher percentage of our weekly priorities

bad: milestone priorities are hard to manage far out in time fix:

  • stop the busy work of maintaining longer-term roadmap items
  • use a long-term roadmap for big plans, not small tasks.
  • every week, review weekly, monthly, and quarterly priorities.
  • schedule people on weekly priorities.
  • know this fix works if we complete a higher percentage of our weekly priorities, and spend less time maintaining roadmap items

the team has committed to these fixes for the 2.6 quarterly cycle.

bad: stale tickets in trac fix:

  • reduce ticket count by periodic focused triage/fix sessions
  • use ticket system only for bugs, put wishlist items on wiki
  • open separate tickets for testing bugs and their fixes, close original bug report with pointer to new ticket
  • put wishlist items on wiki page
  • define stale tickets as 'no comments in 6 months', make trac report available

{LOST THE THREAD HERE, NEED BETTER SUMMARY OF FIX AND SUCCESS CRITERIA}

* error reporting (line numbers)

  • high priority for rakudo
  • testing framework needed
  • in progress: cotto, chromatic, tene have worked or will work on this

* block exit handlers, TT#1523

  • need more info from rakudo team
  • may require pdd/spec work
  • in progress: allison, whiteknight working on it, will contact jnthn

* lexical implementation

  • need requirements from rakudo

* performance/memory usage in general

* plumage (project priority, how to boost resources, how far along is it? what does it need?)

  • project priority for parrot
  • not enough developers working on it
  • will not add to parrot core, will keep as separate project
  • fix: seek out funders, which may unblock developers

* NFG strings

  • gsoc proposal in
  • hope it's funded, discuss later if not

* concurrency

  • gsoc proposal in
  • threads currently work, but are limited
  • not a pre-rakudo * priority
  • will focus on this after gc has been addressed

* prototype-based OO

  • PMC and Class unification will not be extensive
  • other class models will not be excluded
  • PMCProxy will go away
  • will add Prototype PMC to parrot core and notify rakudo team

* HLL interop, TT#556, TT#557, TT#558

  • sorear working on an implementation
  • may require spec changes based on work in progress
  • meeting with stakeholders (sorear, tene, pmichaud, japhb) would be beneficial

* moving the build system off Perl5

  • not urgent, just general long-term direction
  • need smaller tasks, wiki page seems a good place

* chmod files * tar library * zlib library

  • good features, add to wishlist

* git conversion

  • not planned during this quarter
  • revisit at next quarterly meeting

* security subsystem

  • parrot priority, largest blocker for PL/Parrot
  • not before 2.9
  • specific tasks can be completed sooner
  • coordinate with PL/Parrot team

* runcore support, can we trim the list?

  • discuss in upcoming weekly status meeting, to be deprecated at 2.3

* next big task: GC/Concurrency/Lorito/?

  • garbage collection is the next big task
  • possibly followed by: concurrency, lorito, jit

* make sure gsoc students aren't blocked

  • use weekly tasks to support them
  • have them attend weekly status meeting