1.001 - 18 Aug 2009 - Jonathan Yu <frequency@cpan.org>
+ Remove generated files from the repository
+ Use a simple string version, per David Golden's article:
http://www.dagolden.com/index.php/369/version-numbers-should-be-boring/
+ No longer use Perl::Critic to test
+ Remove modules available in core since Perl 5.6
from requires
1.0.6 - 25 Jul 2009 - Jonathan Yu <frequency@cpan.org>
+ Rewrote POD a bit
+ The last variables using TEST_AUTHOR now use
RELEASE_TESTING|AUTOMATED_TESTING
+ Overhaul test scripts, remove author tests from
Recommends (Alias' journal #5375)
+ Provide a more generic README so I don't need to
bother updating it.
+ Minor cleanups in the LICENSE
+ Removed CAVEATS from README
+ Updated copyright statements
+ Require Internet (TEST_INTERNET) to do the signature
test, since it tries to download the key from a
keyserver
+ Slight reformatting, nothing significant.
+ No longer enforce "one dot only" in filenames. This
breaks compatibility with OpenVMS.
+ Minor formatting changes
+ Instead of joining the array, just pass an array ref
(otherwise it is split into an array anyway) [Build.PL]
+ Clarified license for the case where public domain is
not permitted (ie, Germany)
+ Applied patch from RT#46071 -- thanks to Taro Nishino. See:
http://rt.cpan.org/Ticket/Display.html?id=46071
1.0.5 - 12 May 2009 - Jonathan Yu <frequency@cpan.org>
+ Add the Perl portability header ppport.h to MANIFEST
+ New simplified license - Public Domain (fallback on
MIT, Artistic-2+, GPL-2+ only)
+ Removed requirement for minimum 'version' version
+ Added ppport.h for portability to older Perls (this
hasn't been a problem so far, but `perl ppport.h'
suggested it due to SvUV)
+ Added ExtUtils::ParseXS as a dependency; modules were
failing to build on systems that don't have it
installed, because the typemap isn't properly included
by xsubpp/Module::Build. See:
http://osdir.com/ml/lang.perl.modules.module-build/2005/msg00501.html
+ Made some changes mostly related to splint warnings:
o Declared isaac() as static because it isn't used
outside rand.c
+ Minor reformatting
+ Changed line to: if (ctx->randcnt-- == 0) { to make
intent more obvious
1.0.4 - 06 May 2009 - Jonathan Yu <frequency@cpan.org>
+ Changed the valgrind test from an author test to
its own special test - one must set TEST_VALGRIND=1
to test the module through valgrind. This must be
done until we have better suppression support
upstream in Test::Valgrind.
+ Updated t/02valgrind.t so that tests are not planned
(so Test::Valgrind can plan)
+ Removed version-specific stuff from the README
+ Reformatted the POD file
+ Updated support information in META.yml and Build.PL
+ Added some support information
+ Reformatted rand.c a bit; moved variable definitions
to the top
+ Added .tmp to MANIFEST.SKIP
+ Made this a dynamic config (instead of using META.yml)
+ Added ExtUtils::CBuilder to META.yml
+ Added Build.bat to MANIFEST.SKIP for Win32 builds...
1.0.3 - 23 Apr 2009 - Jonathan Yu <frequency@cpan.org>
+ Rewrote docs a bit
+ Removed Pod::Usage dependency
+ Added some example files for benchmarking the module and
charting out the distribution generated by the module :-)
+ Added Config to META.yml prerequisites (why not?). It's a
core module so it wouldn't be missing anyway, and Build.PL
would fail outright if it was.
1.0.2 - 21 Apr 2009 - Jonathan Yu <frequency@cpan.org>
+ Add Build.PL detection of int size
+ Version bumped to 1.0.2
+ A flag now controls whether 'int' or 'long' is used as the
type: USE_INT. If Build.PL detects that int is 32bits wide,
then it will use that type. Otherwise it'll use long, which
is more portable.
+ A flag will control if the portability code is used. Right
now this involves masking all numbers with "& 0xffffffff"
to cut off the high bits - this code isn't necessary for
32-bit variable widths, so we might squeeze a bit more
performance out this way
1.0.1 - 07 Apr 2009 - Jonathan Yu <frequency@cpan.org>
+ Changed the bootstrap loading to use VERSION instead
of VERSION->numify, so that they match (prevents compile
errors with "object version v1.0 does not match bootstrap
parameter 1.000000" (apparently causes failures in 50% of
machines so far)
1.0 - 06 Apr 2009 - Jonathan Yu <frequency@cpan.org>
+ Initial release