Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

Perl 6 Design Minutes for 10 September 2008

Journal written by chromatic (983) and posted by brian_d_foy on 2008.10.10 8:52   Printer-friendly
The Perl 6 design team met by phone on 10 September 2008. Larry, Allison, Jerry, Patrick, Nicholas, and chromatic attended.
Allison:
  • debugging the MMD branch
  • we've dgone from 900 failing tests to 47 in the past two days
  • working my way through failing tests has had a compounding effect
  • chromatic has been helping
  • I shifted the nastiest failure from a termination in the middle of nasty code with no warnings to a segfault
  • and he fixed that
  • we're very close to merging the branch

c:

  • probably after the next release

Allison:

  • yes, probably

Jerry:

  • a segfault is better?

c:

  • much easier to fix

Allison:

  • this was a PGE problem
  • it pretended it was successful
  • it output nothing
  • it was so much easier to debug then

Nicholas:

  • die early, die often

Patrick:

  • I can't tell you how many times I had PGE output nothing
  • zero?!

Allison:

  • we're having a Parrot developer day in November in Moutain View

Jerry:

  • finally had time to do a little development this week
  • then I had hardware trouble, and couldn't commit anything

Larry:

  • between my day job and my grandson visiting over the last week, not a lot of spare time
  • took several hundred Parrot-specific tests that Moritz put into t/spec
  • fudged them all
  • went through several permutations of how to write the tests
  • realized that there was a missing operator
  • no way to do a pattern object for smart matching that's equivalent to a junctional and but guarantees the order of evaluation
  • that's almost exactly what a where infix operator would do
  • a type on the left and a constraint on the right of it
  • you want to test against a subtype name and a constraint in order
  • lots of places want that sort of pattern
  • we had no easy way to express that short of defining a subset type, which is hard to use for one op
  • or 700 ops in that file
  • there's now a where operator that parses just like the junctional &, but in guaranteed order
  • otherwise just random spec cleanups
  • answering questions
  • lots of people thinking about the nature of lists and iterations now
  • did a little catchup on email on p6l
  • woefully behind on my own personal email
  • doubtless there are bugs in that
  • oh, ended up adding a Perl 5 grammar to STD
  • copied and pasted the Perl 6 regex rules and cut out two thirds of them
  • maybe just need to borrow a test from the Perl 5 suite

Patrick:

  • PGE uses one
  • just uses the test file directly

Larry:

  • not fudgeable

Patrick:

  • it doesn't use the perl5: flag
  • PGE doesn't understand that flag
  • it'll fail miserably

Patrick:

  • added code to Rakudo to go to standalone PIR
  • generates PIR that you can run directly from Parrot

c:

  • that should help you help me

Patrick:

  • that should help a lot of people help you help me
  • updated the Failure type in Parrot
  • warns if you use it while undefined
  • originally I put it in to die fatally
  • that broke the test suite
  • a few people are going through the test suite to fix the evaluation of undef Failures
  • created a couple of functions in Rakudo to make it easier
  • finished my final report for the Mozilla Foundation
  • Richard and Jesse are happy with that
  • I'll publish that later today or early tomorrow
  • that grant is over now
  • trying to figure out list interpolation and how things fit together
  • Rakudo has some nasty bugs there
  • I feel like I'm missing something there
  • it seems like there's another type in there

c:

  • added a new feature to Rakudo last week
  • fixed a couple of bugs to get a lot more tests passing
  • keeping up with the incoming patch and bug queue
  • Christoph is good at helping
  • also trying to fix bugs and help the MMD branch along
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.