corrections on loggers + spec in AST
start to write error messages
prepare for the modern error handling of Menhir
fix almost all warnings
- Refactored Error exception and messages- Bugs in partial evaluation for equalities among bool constants and a nice recursive call generating a stack overflow! Now solved- Setup a timeout for z3 in seal- Better log for seal
Array access: solved issues in C backend when basic operations in array access dimensions. Also better handling in EMF, ie further normalization through new equations
cleaning debug logs
more explanation in case of failure. Still dirty
Cocospec: parsing, normalizing and processing machines for contracts.
Some progress on compiling cocospec contract.Contract resolution still need to be done as well as dealing with the machine code level and so on.
Unified compilation of lusi and lus filesDifferent parsers yet but shared process.In case of lusi input the C backend is bypassed since the .h is generated from the lusic and no C code should be generated since it may overwrite existing manually written code...
- Dep type with a tuple has been replaced by a record type- Modules now is more integrated and performed the building of the type/clock env. previously some computation were performed twice by different functions. Some of these functions have been moved from compiler_common to modules
- Global type env and clock env now availble as a global reference (Global module)- Adapted the parsing of specification with a cocospec compatible one- The data structure of contracts is now almost cocospec compatible- Lustrec-test has been updated to use the newest syntax
- Makefile: solved dependency problem when compiling include lusi- Renamed type declarations as lustre_types and machine_code_types
Recursive resolution of dependencies
[general] large modification: added machine types, a second typing phase dealing with machine types (eg uint8) typing was transformed as a functor and parametrized by basic types (int/real/bool) it can also be applied multiple times on the same program
Refactor error printing.
- Added a field lustre_eq to machine instruction in order to record the originating lustre equation- EMF backend now impose the optimization level to be set to 0 in order to avoid equation elimination that would render traceability difficult- Options.ml has been split into Options.ml / Options_management.ml. Options.ml only contains references and no functions
Providing means to have specification as dynamic checks.[bug] seems to crash with EMF backend
Improved include folders behaviors:- allow multiple -I dir, will be used in order (first one declared is first used)
- when declaring a global library #open <foo>, foo is first checked in local folders, than in global one (install path). This does not apply to local libraries opened with #open "foo".
trying to improving pretty printing. Not so perfect for the moment
Merge branch 'unstable' into merging_pluginsNon regression results were similar to master branch
Solved a bug in the compilation of asserts. Now different behavior depending on the backend:functional: keep it as isnon func: introduce a fresh local var v and replace assert(e) by v=e; assert (v);
full merge of salsa/mpfr and master
several bugs/oddities corrected: - bizarre date stamping (when generating .lusi files) - bug in loading libraries, forbidding several (transitive) imports of the same library. - plugin scopes was wrongly always activated
major branche merging salsa/mpfr with trunk
Major revision due to severe limitations and bugs of inlining capabilities: - destination dir should now work properly - lusic files now have a version number, to avoid nasty segfaults when loading lusic files created by an older compiler version - inlining should now work with generic nodes and generic array library...
correction of bugs: - a small problem in the parser - regarding the handling of destination directory, source directory, current directory, etc. It seems to be working now. A nice chasing after weird behaviors...
git-svn-id: https://cavale.enseeiht.fr/svn/lustrec/lustre_compiler/trunk@430 041b043f-8d7c-46b2-b46e-ef0dd855326e
Added a construct for Dependencies (was a tuple before) and a boolean attribute stateful
git-svn-id: https://cavale.enseeiht.fr/svn/lustrec/lustre_compiler/trunk@380 041b043f-8d7c-46b2-b46e-ef0dd855326e
- Added major feature: Lustre V6 automata !!! - one automata example added - changed the reset condition in node calls (now a simple bool expr) - bug corrected in clock calculus - bug corrected in traceability info - added field in variables to test whether they are original...
This is a major revision: - added interface files (.lusi) in the language, that can be compiled on their own, giving an object file (.lusic) and a header file (.h) - modular code generation, from Lustre to C level included. - nice amount of code refactoring...
This is a major revision: - added interface files (.lusi) in the language, that can be compiled on their own, giving an object file (.lusic) and a header file (.h) - modular code generation, from Lustre to C level included. - nice amount of code refactoring