use Perl

With nothingmuch's lead, it seems that a good portion of development is now occuring in SubEthaEdit space. That is exciting news -- here's hope that MoonEdit can work on OSX some day, or (preferably) DocSynch or any of the many SEE-clone projects can finally deliver us from the multiplayer cat (IRC) age into the multiplayer notepad age.

Today saw 50+ commits by 10+ committers, which is at least twice the number I've been used too. Maybe I'll get used to it gradually... *grin* Anyway, here's some of the developments happened today:

• Stevan, nothingmuch & co worked on a test cataloging script, an important technology that will cross-reference the Synopses with the coverage in unit tests. It will let us better visualize the untested/unimplemented parts of Synopses better, hopefully with minimal burden on test writers.
• malaire tested, implemented, refactored, and simplified the index primitive.
• Darren Duncan turned Algorithm::Dependency into a Perl 6 module in just a few hours, including the entire test suite. Wow!
• lightstep, our prolific lambdafolk, cleaned up the source tree and hacked the interactive Shell; now lexical variables is preserved across lines, so we can now do this:

  pugs> my $x = 1
  1
  pugs> $x
  1

• As more and more people are building Pugs with GHC 6.4, lightstep, theorbtwo and I worked on silencing warnings and fixing dvergin's build failure.
• mugwump started working on a Kwid spec; ingy moved core kwid documentations into ext/Pugs-Documentation/.
• scw cleaned up some tests and elaborated on the precedence problem around ?? ::. This leads to...
• ...I finally checked in the code to parse ?? :: correctly, thanks to scw's analysis and the advice from kosmikus. Here is a short write-up of the approach.
• Thomas Yeh reported that pugscc was broken on case-sensitive filesystems, because I misspelled MainCC.hs as MainCc.hs. I fixed it, and it now works; we can finally make stand-alone AIX executables from Perl 6 code to deploy into production!
• I have implemented the true primitive, which is to ? what not is to !. Everyone on #perl6 disliked the operator name; after some discussion on p6l, Larry suggested so, which I happily implemented.
• I fixed last so it no longer conflicts with return, and simplified the control flow considerably. All hail callCC!
• gaal checked in many, many TODO tests; nothingmuch cleaned up many of them, and many besides. We have 1659 tests now.
• ingy makes Pugs installable on systems that did not install Perl5 into a path that contains perl.
• Pugs was numifying "0xFF" as 255; that was arguably the wrong behaviour. After some discussions with pmichaud, we settled on only numifying digits and the dot, ignoring everything else; that means "1e2f3g4" is now numified as 1, not as 100. Larry has since expressed the DWIMmier view, so that may change tomorrow. :)

So, that's it for today. See you tomorrow!