Slash Boxes
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 ]

autrijus (1505)

  (email not shown publicly)

Journal of autrijus (1505)

Monday June 20, 2005
04:07 PM

Day 137: Final day in Leo's: Control structures!

[ #25290 ]
Turns out I didn't get enough time to roll a release tonight -- I'll leave for Toronto in 5 hours. Oh well. :-/

To reduce confusion, I've renamed PAST in Pugs's source code to PIL (Pugs Intermediate Language). It's easy to remember, and fits well with the PIL -> PIR flow. :)

The good news is that leo and I figured out how to implement closures and return/leave, both for his new efficient runloop and simulated with exception handlers for the current runloop. Building on this model, and solid abstractions built over the last week, we see a large influx of PIR checkins today:

  • Globals: @*ARGS, $*PROGRAM_NAME, $*IN, $*OUT, $*ERR
  • Iteration: for 1..10 -> $x { say $x }
  • Code exit: return and leave
  • Loop: loop(;;){}, while{} and until{}
  • Bare, pointy block and anonymous subroutine literals.
  • The entire is now compiled and loaded into PIR.

We're now very close to get make pirsmoke and make pirtest returning some useful results for Pugs/PIR. ParTcl currently supports 2.89% of the entire Tcl test suite (or 6.88% if you take out tests for the single clock command); we'll see how well Pugs/PIR fares tomorrow.

Oh, and we're three days till the announcement of ICFP 2005 contest. I wonder if I can convince my fellow Hackathoners into entering the competition with Perl 6... It sounds insane, but hey, wouldn't it be really amazingly cool, if "Perl 6 is the language of choice for discriminating hackers!" gets announced in ICFP 2005? :-)

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
More | Login | Reply
Loading... please wait.
  • To be fair, a large number of the Tcl tests fail because of unimplemented commands. For instance, out of the 12313 tests, 7137 of them belong to the clock command, which is unimplemented. Taking out those tests increases the percentage to 6.88%. There are a fair amount of bugs, but there is also a fairly substantial start.
    • Yup, it's great work. I've reflected the numbers above. It's by no means disrespectful of your work -- indeed, Pugs/PIR currently pass only 1% or so of the tests, because had not yet bootstrapped. Even if it has, I doubt we'll have a 6.88% coverage at this moment... :-)
  • I never really liked the “PAST” acronym anyway – I don’t like the unfortunate allusion to “the past [].”