2015.46 Production Today

Grapheme \reu\nification

The past week saw the demise of Carriage Return / Line Feed as a pair of graphemes. It is now a single grapheme in Rakudo Perl 6 as promised by jnthn++ last week:

say "\r\n".chars # 1

But the goodies didn’t stop at this. Handles now have both an nl-in as well as an nl-out attribute to replace the old nl attribute. And the nl-in can contain multiple strings that can be considered an end of line. The default is ("\r\n","\n","\r"), thereby adding support for reading all standard text-files. There are more developments in this area, but still too much in flux to mention here and now. Most likely next week there’ll be more to tell about this.

Quote Of The Week

I wanted to give a quick thanks to everybody who has worked on #perl6. Our first perl6 code went into production today 🙂

More quotes from the ensuing discussion:

It uses supplies, which made the whole thing a breeze to write

and:

It solved a real pain of a problem for us in a very easy way.

What can I say?

More Than 50K Tests Passing

The Perl 6 spec test-suite now passes more than 50000 tests with Rakudo on MoarVM.

All tests successful.
Files=1070, Tests=50009, 245 wallclock secs (12.62 usr 4.31 sys + 1406.00 cusr 131.27 csys = 1554.20 CPU)

This is all part of the effort of getting more test coverage in light of the Christmas release of Rakudo Perl 6!

Dogfooding

Leon Timmermans++ came with an extensive pull-request that introduces a new asynchronous test harness completely written in Perl 6! It seems still a bit rough around the edges, but if it delivers on its promise to improve parallel testing on Windows systems, I’m pretty sure this will get merged real soon!

Blog Posts

Two posts this week!

Noticeable Fixes And Additions

  • jnthn++ implemented chomp functionality (automatically removing the line ending when reading a line) at the MoarVM level. This made a lot of Perl 6 level hacks unnecessary, which were promptly removed.
  • jnthn++ implemented UTF-8 C-8 encoding. This basically maps the byte values 128..255 to Unicode synthetics. This is intended to be used for all OS level calls for which we do not have a known encoding, like reading directories.
  • ilmari++ implemented a :replacement named parameter to Str.encode to handle byte-sequences that could not be encoded. This was his first contribution to Perl 6, and involved work on MoarVM, nqp and Rakudo!
  • lizmat implemented Str.split(@needles), and :k, :v, :kv, :p and :skip-empty named parameters for Split.str. This sped up the canary in the coalmine test of Text::CSV by a factor of 2.
  • lizmat implemented a Str.comb(Int $n) candidate for looking at a string $n characters at a time. This alternative is about 40x faster than the equivalent /.../.
  • Various speedups: IO::Handle.getc 2x faster, Buf/Blob.new(...) upto 20x, for ^10 { ... } is optimized again.

Ecosystem Additions

Advertisements

Got something to note?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s