view release on metacpan or  search on metacpan

Changes  view on Meta::CPAN

	Tue Nov 13 2012
	- Long-running tests now check for $ENV{AUTHOR_TESTING} = 1,
	  and skip if it doesn't.
	Tue Nov 6 2012
	- Worked on bitonic() code. Use existing pack/unpack code to
	  determine power-of-two value, this replaces the
	  greater_power_of_2_less_than() function.
	- Made loop a little more perlish by using the '..' operator.
	- Added 'bitonic' to the keyword list.
	- Added a get_options parameter in Build.PL. Looking towards
	  an option to automatically skip long-running tests in
	  non-development situations.
	Tue Oct 30 2012
	- Fall-back algorithm for 'best' option is now Batcher's Merge
	  Exchange algorithm instead of Bose-Nelson.
	- Checks for $inputs < 2 in the algorithm functions are pointless
	  since the calling, exported function already checks for it.
	- "use integer" means that int($n/2) may be written more simply
	  as $n/2.
	Sat Oct 27 2012
	- Added two new 'best' networks for input sizes 18 and 22,
	  discovered by Sherenaz Waleed Al-Haj Baddar and submited
	  by Doug Hoyte.
	- Added the bitonic() code and test file as provided by Doug Hoyte.
	Thu Oct 25 2012
	- New function nw_sort_stats() to return information about
	  sorting network exchange counts (based on Doug Hoyte's
	  idea) and other future ideas.
1.24
	Sun Oct 14 2012
	- Emergency change to svg.t -- I'm doing something wrong in
	  the tests, so for now they are skip()ped.
1.23
	Thu Oct 11 2012
	- Doug Hoyte requested the addition of the sorting network
	  equivalent of the bubble sort, for demonstration purposes
	  only, and sent a patch to implement it and a test file to
	  add to the t/ directory. Free code is hard to turn down,
	  even if it is for bubble sort.
	Tue Oct  2 2012
	- Added a new test, svg.t. It is modeled on pod.t, and uses
	  Test::XML::Easy. If I had had it as part of the tests
	  before, I wouldn't have let the badly merged code in the
	  earlier versions get released.
1.22
	Sun Sep 30 2012
	- Something odd happened to the SVG code post version 1.08,
	  leading to unparsable output (could be the result of the
	  hardware failure, could be merge error on my part).
	  Took the SVG graph code from version 1.08, and re-applied
	  the changes to the graph code between then and version 1.21.
	  The graphs come out correctly now.
	Fri Sep 28 2012
	- Took horizontal and vertical coord code out of nw_eps_graph()
	  and nw_svg_graph(), and made them into their own functions.
	- I really had an odd style for initializing variables from @_
	  back in 2003. Made it less odd.
	- Small inefficiency with the column count in nw_text_graph() fixed.
1.21
	Sun Sep 24 2012
	- Added color debugging code in eg/svg.pl (it lists the
	  color settings in XML comments).
	- Updated eg/text.pl to take command line options.
	- Moved graph and text option handling to their own functions,
	  graph_segment() and text_segment(), for future non-messy coding.
	Fri Sep 22 2012
	- Removed prototypes.
	- Found and removed some left-over marker code in the SVG output.
	- Moved SVG graph color code to its own function, svg_color().
	Mon Sep 10 2012
	- Modernized parts of the code -- $VERSION is now an 'our'
	  variable instead of a 'use VARS' variable, for example.
	- Added 'create_license => 1' to the Build.PL list of attributes.
	- Added 'dynamic_config => 0' to the Build.PL list of attributes.
	- Added LICENSE to the MANIFEST file.
	- Yanked an "$flag_internal" variable (and its debugging code)
	  in favor of Smart::Comments.
	- Bumped the $VERSION value up a point.
1.20
	Sun May 20 2012
	- Really bad bug introduced in version 1.08 (text graph
	  options were checked for in the wrong hash) has been squashed.
1.10
	Wed Mar 28 2012
	- Finally organized changes (I had a hardware failure), and updated
	  the MANIFEST file to include the new JSON formatted META files.
	  Ready for release to CPAN.
	Tue Jan 31 2012
	- Perl version requirement changed to 5.8.3 (from 5.6.0). Support
	  for 5.6.0 in various perl development tools is being dropped,
	  and 5.8.3 seems to be a safe minimum, at least for the current
	  year. Change made in Build.PL and Networksort.pm.
	- Changed "repository" key in Build.PL from a hash to a string,
	  as seems to be currently required under the version of
	  Module::Build that I'm developing with (the hash form looks
	  to be for a later version that I cannot install yet).
	Tue Jan 11 2011
	- Found an error in the SYNOPSIS portion of the POD. Function
	  nw_format() had an incorrect argument.
	Thu Sep 23 2010
	- Experimented by adding a 'keywords' hash to Build.PL. We'll see if I
	  did it correctly.
	Wed Sep 15 2010
	- Update the references. Joe Celko's article in dbazine.com is gone
	  (dbazine.com seems to be under new ownership), but he used the
	  same material in the third edition of his book, so I refer to
	  that now.  Dr. Batcher's web site at Kent State has changed slightly,
	  making a change in the links necessary.
1.09	Sat May 22 2010
	- Error in POD snuck in. While I was fixing this, added some more
	  documentation on group options.
	- Added another choice for "grouping" in nw_comparators().
1.08    Mon May 17 2010
	- Added an option to nw_comparators() as a result of a communication by
	  Daniel Stutzbach. Up until now, nw_comparators simply returned the
	  comparator in the order that the algorithms created them. Daniel
	  Stutzbach pointed out that these were often not in the best order and
	  would cause a super-scalar pipeline to stall. He also pointed out that
	  just reading the list of comparators grouped for graphing removes that
	  problem (the function that arranges them for graphing, nw_group(), does
	  in fact "parallelize" the comparators so that they'll look nice on the
	  graph). So there's now an option to re-group the list. It's not the
	  default because the default behavior has been there since the beginning
	  of the module.
	  See the documentation for nw_comparator() for more information.
	- Undid use of mark attributes for SVG graphing. Too many inconsistencies
	  between different SVG viewers for me to be sure that I'm doing the right
	  thing.
	- While undoing that change, simplified the code that generates the
	  group/color attributes.  I decided that individually assigning the
	  color attribute to each element was probably a less error-prone mechanism.
	- Updated svg.pl and eps.pl in the eg directory to make use of the
	  color options.
	- Fixed the nw_group() example. It was using the default 'bosenelson'
	  algorithm, but the output was from the 'batcher' algorithm. Oops.
	- Added grouping.pl to the eg directory.