Rascal

Built-in functions do not appear twice in "functions" any more, removal of a compiler option that was problematic with egcs, exception-catching, and new output precision configuration and query functions for all predefined flavours of Rascal.

With C-XSC modules floating point number string parsing results in the different data types directly, minimizing round-off errors. If a number is being added to a highly accurate data type, it will be parsed as such a data type. Added "variables" and "functions" procedures. Updated documentation, added interval compares and corrected some tests.

An automated test environment was introduced. Command-line options and rc-files
were added. Minor bugs were corrected, and missing functions in various
datatypes were added.

Implicit casts determination method has been
corrected and improved. Implicit cast from long
integer and fractions now correctly default to
double and not to string. An experimental test
case to prevent such fatal bugs in releases has
been created.

A long integers (big numbers) module and long integer fractions are available. sqrt from double and integer now maps to sqrt of complex if needed. Complex number output uses units i or j if one of these variables are defined as the imaginary unit. Numbers like "2." without digits after the dot are now interpreted as doubles. Determinants and Inverse of matrices can now be computed as inv(A) and det(A). Additional spaces in function definitions are now ignored.