TWIP - This week in parrot, ending February 14th, 2009

These articles aren't meant to cover every change that's occurred in parrot in the last week, but to show some highlights for both developers and users, and provide a gateway to more information.

Release Milestones

The next monthly release, expected to be 0.9.1, is due this coming Tuesday, February 17th, 2009. See the call for action.

The following release, expected to be 1.0, is due on March 17th, 2009. See the vision document.


trac2email is still not fully functional. In the meantime, please continue to use the web interface.

There's definitely still confusion about which list is for what, as even the new parrot-dev list is getting patches for rakudo-perl6. If you're not sure where to send something, start here.


As usual, a lot of bug fixes, and some new development. Some items of note:

  • pmichaud updated the utility which converts PBC into a fakecutable (it still uses parrot to run bytecode, but the bytecode and the parrot startup are bundled into a single executable. This resulted in a speedup in generating the perl6 fakecutable from 175s to 16.5s.
    • This work exposed a bug with buffered IO that was resolved by infinoid.
    • Speed issues with buffered IO also re-raised an issue regarding switching between parrot calling conventions and C calling conventions. chromatic has pointed out that (counterintuitively) we'll end up with faster code if we avoid switching context into C and back. Allison fired up a branch; hopefully this can be partially addressed before 1.0.
  • Some churn toward getting parrot a BigNum PMC.
  • PMCs have been retronymmed to PolyMorphic Crackers. Er, I mean Containers.
  • Work has begun on a Parrot Virtual Appliance - should be a bootable cd with working copies of parrot, docs, and more.
  • Proposal about the self keyword in PIR has generated some feedabck this week on the list.

Issue from the Vault

We've been doing a great job closing out issues (especially from the old tracking system). A lot of the mailing list traffic in the past week is fact finding or closing out of previous issues.

Here's one crufty old issue that I hope we can get rid of by mentioning it here. RT #36086 - FreeBSD t/op/trans.t failure when running with jit.

The BSDs aren't one of our core platforms, but we usually do pretty well there. We've had some recent work on NaN by particle that might have already resolved this issue; hopefully closing this one out is as easy as getting an up to date report on make fulltest from a FreeBSD hacker.

Google Summer of Code

Google has posted their FAQ for Summer of Code 2009. Parrot hopes to participate again this year. Last year we entered as a part of the Perl Foundation umbrella organization, but this year will try as an independant organization.


  • Active issues in trac: 168 (up from 158).
  • Active issues in RT: 460 (down from 501).
  • Commits to parrot in the last week: 202.



If you have tuits to spare to help out parrot, here are some simple points of entry:

  • Setup a scheduled task to build parrot and report back via smolder.
  • Read through the documentation and make sure it makes sense and that the examples work. The documentation exists in several places: primarily in the repository and the trac wiki.
  • Help us close out the old wiki and complete the migration to the wiki in trac.
  • Go through old tickets in RT and verify that the reports still make sense. Can you run the code in question? Does the patch still apply? Is the report against an old copy of parrot that we have a recent smolder report for?

If you can't donate time, but are still interested in the success of the project, you can also consider donating through the parrot foundation. Donations are not tax deductible. (FYI, I'm on the board.)


To find out more about parrot, you can follow the same resources I used:

  • The weekly developer meeting on IRC in #parrotsketch on
  • The daily chatter on IRC in #parrot, also on
  • The parrot-dev mailing lists.
  • The Trac timeline, which includes bug reports, svn commits, and wiki updates.

This summary written by Will "Coke" Coleda.