tecomp 0.18.1 R202

tecomp is a command line compiler that currently works like an Eiffel interpreter. It compiles Eiffel source code into an internal representation and executes it in its virtual machine.
tecomp runs on Unix and Windows machines (64/32 bit, big/little endian) and on all machines which have gcc and the corresponding make utilities installed (requirements see below).
NOTE: tecomp is licensed and distributed under the terms of the GNU General Public License (GPL) and GNU Library or Lesser General Public License (LGPL).

Main features:
Kernel library:
- available classes: ANY, CHARACTER, CHARACTER_8, BOOLEAN, INTEGER, INTEGER_8/16/32/64, NATURAL, NATURAL_8/16/32/64, STRING, COMPARABLE, STD_FILES, FILE, POINTER, REAL, TYPE, PART_COMPARABLE, INTERVAL, TUPLE
- available features of ANY: io, default_create, type, out, print, is_equal, default_is_equal, copy, default_copy, cloned, default_cloned, is_deep_equal, deep_cloned.
- all conversions between basic types are implemented
- missing features of ANY: onces
- not yet available classes: HASHABLE, STORABLE, MEMORY, EXCEPTIONS, ARGUMENTS, PLATFORM, ONCE_MANAGER, ROUTINE, PROCEDURE, FUNCTION, PREDICATE
Language mechanisms:
available:
- object creation
- object test
- detachable types
- old expressions
- creation expressions
- conversion
- feature call
- feature redefinition
- developer defined expanded types
- deferred features
- polymorphic assignment (dynamic bind)
- assertions (preconditions, postconditions, class invariants, checks, loop variants, loop invariants)
- aliases (with standard operators only)
- rename, redefine, export, undefine, select
- once routines (Result cannot yet be expanded)
- ace files to control class assembly
- switchable assertion monitoring (per cluster)
- public and private features (but no selective export)
- void safety
- automatic boxing
- constant attributes
- TUPLEs
- garbage collection

Enhancements
- Version 0.18 implemented target conversion for binary operators.
- The implementation bypassed type validity checks with binary operators. Therefore invalid code could be executed which might cause runtime crashes.