The lambdafolk skew joined #perl6 today, offerring many useful performance hints to Pugs, including narrowing type declaration and other helpful tricks. At end of day, we are about 5% faster than yesterday -- that should about conclude this week's optimisation efforts before 6.2.2 ships.
At skew's suggestion, we also have a better make profiled and a working make pugs.prof target, to run the entire test suite in profiled mode to find out bottlenecks.
theorbtwo wrangled with Haddock and emerged victorious -- we will have a nice HTML documentation to accompany 6.2.2! He is now adding function signatures to improve Haddock's output.
During his work on testing default arguments, Corion reported that for @array { say } broke, because $_ is not the implicit argument in the inner block -- @_ is. I have special-cased to turn the block used in for to default to $_ for now, pending further clarification.
pjcj, our Devel::Cover guru, inquired about the state of instrumentation support in Pugs following yesterday's Pos change, and I replied affirmatively. A Perl 6 code coverage tool will be very nice indeed.
iblech repaired the Parrot backend accidentally broken yesterday by the Exp change. pjcj also saved the tree from the brief breakage from Haddocking, and revdiablo applied the patch for him. Limbic_region noticed and fixed a broken next.t. Thanks!
gaal continued to hack on #line 123 "filename" support, and helped me in adding comments to main data structures in AST.hs.
Stevan carried on Schwern's work of yesterday, making Test::Harness report TODOs as well as failures. As this feature is useful beyong Pugs itself, he sent the patch to Andy, so it should be in the next release of Test::Harness.
So, I need to sleep now. See you tomorrow!
