Develop and Download Open Source Software

Shed Skin

Project Release infomations and Project Resources. Note that these informations are from this projects Freecode.com page and the downloads themselves may not be hosted with SourceForge.JP.

Project Release Information

2012-01-17 05:40
Several bugs were fixed. itertools.product was greatly optimized. str.join and str.__ne__ were optimized. A new 'image scaling' example was added (hq2x). The optional enumerate 'start' argument is now supported. Extension module warnings were improved. A new option -L was added, to add a library directory. FLAGS can now reside in /etc/shedskin, and lib/ in /usr/share/shedskin.
2011-09-12 01:54
Scalability was again greatly improved. Performance of file I/O and complex numbers was greatly improved. Three new modules are supported. Six new example programs were added. There were many other bugfixes and improvements.
2011-06-22 01:27
Support was added for the "struct" and "array" modules. Three new examples were added, for a total of 57. The C64 example was updated, and is now more than 3,000 lines (sloccount). An option was added to display exception tracebacks (shedskin -x). Many bugs were fixed.
2011-03-27 05:25
Extension module support was improved. Several builtins were optimized: zip, min, max, map, filter, reduce, and pow. A warning was added for non-uniform tuples of length greater than 2. Internals for looping and sequences were cleaned up. Two new examples were added: a quantum monte carlo simulator and an rsync implementation.
2010-12-21 06:40
A new Windows package was added. The warning system was fixed again (this time with tests). There were many minor bugfixes and some optimizations. Two new examples were added.

Project Resources

http://freecode.com/urls/3aa64e0a36c65f41537e4d34b3bbd3cd
http://freecode.com/projects/shedskin

Project Description

Shed Skin is an experimental (restricted-)Python-to-C++ compiler. It accepts pure but implicitly statically typed Python programs and generates optimized C++ code. The result can be further compiled to stand-alone programs or extension modules. For a set of 63 non-trivial test programs, at more than 16,000 lines in total (sloccount), measurements show a typical speedup of 2-200 times over CPython. Not all Python features are supported, and only a subset of about 25 library modules, such as re and random.