- refactorization of typing code (simpler subtyping rules) - simplification of clock calculus (may be still buggy, work in progress) no impact on unclocked programs.
clean handling of undefined node application
bug correction in homomorphic extension
Changed the load of lusi files: imported nodes or function can specify the linking lib and/or use a classical C prototype (without pointers).Parse updated as well as Makefile generation.
changed name generation to avoid conflict with C predefined symbols; added checking for declared but not defined node symbol
- corrected causality bug (cf. previous commit)
- reimplemented computation of dead variables - added computation of a reuse policy (depending on types) - not yet used though, would have to change code generation in order to be useful...
- work in progress for stateless/stateful status computation (to turn conditionals into merges, which yield more efficient C code)
- added struct types declaration - added constant definition with a struct type - added checking for multiple definitions of nodes (behavior was buggy) - better and more uniform error messages for undefined/already defined symbols
We still need struct expressions...
work in progress for struct types...
more steps towards struct types...Cette ligne, et les suivantes ci-dessous, seront ignorées--
M trunk/src/corelang.mliM trunk/src/type_predef.mlM trunk/src/main_lustre_compiler.mlM trunk/src/types.mlM trunk/src/printers.mlM trunk/src/typing.ml...
first steps towards struct types...
- added generation of clock information in interface (.lusi) files- added clock checking between interface and implementation files
- merged test script - added -d support - corrected #open parser problem - corrected interface/implementation (.lusi/.lus) checking for types (not yet for clocks)
Merge inlining branch within trunk.The test target requires branch lustrec/horn as binary lustreh.
Solved some bugs in the lustre printerGeneration of a witness with both the main node and hte inlined main nodeTest script modified to check consistency of the inlining process
better error message for tuple type mismatch
again, debugged tuple subtyping
removed debug printing
corrected wrong subtyping rule for tuple assignment
added subtyping in equations (rhs may be a subtype of lhs)
- work in progress for interface typing
- small bug correction in dimension typing - #open keyword instead of open - dummy generic matrix/vector library interface added - modified examples according to the new syntax
Merge (if it works) of the lustre interfaces branche providing lusi files into trunk
Moved files to trunk in lustre_compiler