GitLab

The oracle is not necessary anymore to run a test.

Moreover, the sut no more needs to (and must not) be named 'sut.ml'
(resp.  'oracle.ml' for the oracle).

Moreover, rename the env var EXTRA_CFILES into EXTRA_SOURCE_FILES, so that users
use it for ocaml programs.

For the time being, it crashes if no valid oracle.ml exists in the
user directory. Moreover, and the sut has to be named 'sut.ml' (resp.
oracle.ml for the oracle). But otherwise, it works fine.

rest of the modules. More precisely, the lucky tools now rely only on
Prog (see below), that is Lucky-agnostic.

The rationale is to ease the plug of a Lutin programs interpreter
afterwards. Indeed, one (ougth to) just need to implement its own
version of the Prog module.

To do that, I've splitted LusProg into :
   LucParse
   LucProg
   Prog

Then, I moved the content of Env_state into LucProg. And all this
had some consequences, that cause more code movement.

Actually, my first idea was to change Run_aut so that the fact that
several programs were run in parallel was hidden. But it turned out
to be a bad idea, because of numeric unsatisfiability + the
interaction between the module that choses a formula, and the one
that draw (performing the numeric satisfiability solving). Indeed,
when a formula appear to be unsatisfiable for numeric reason, we need
to backtrack in via Run_aut.choose_on_formula in one of the
Run_aut.t ; but of course if the list of Run_aut.t is hidden, we are
dead. One solution would have been to complicate a little bit the
Run_aut.t interface ; but my first motivation was actually to
simplify it...

luciole to generate them.

first one should be a boolean, and is the one taken into account in
order to decide if the test fails or not. All the outputs of the
oracle are printed in the rif file after the #ORACLE_OUTPUT pragma.
That can be very convenient to track why the oracle failed.

Or j'étais en train de bidouiller plusieurs trucs sans avoir rien
commité...

Bon, plouf, plouf. Je synchronise et je repars sur des bases saines avec git.

source/luc4c_stubs.c:
        je sauve et je restaure les fegetexcept flags pour tourner autour d'un bug dans
        caml qui se manisfestait quand j'appelais du lucky depuis Alices.
        C'est vraisemblement la source de mes déboires avec Scade/VCS.EXE il fut un temps.

Parent-Version:      1.44
Version-Log:

Fix an embarassing bug in the integer solver (in dimension 1).

Project-Description: Lurette

Parent-Version:      unstable.5
Version-Log:

source/util.ml
  add an entete function in order to decorate generated files.

source/solver.ml
  Add more debug info in case of failure.

source/parse_luc.ml
  better error message

source/env_state.ml
  Fix a bug in the handling of the (undocumented) ~default option.

source/lurettetop.ml
  better error message

install/configure.in
  Add stuff for configuring mac os
  use install instead of cp.

install/Makefile.lurette.in
user-rules
  some improvements(?)

source/Makefile.lucky
source/Makefile.gen_stubs
source/Makefile.lurette_lib
source/Makefile
polka/Makefile.config
polka/Makefile
source/gen_fake_lucky.ml
source/Makefile.gen_fake_lucky
install/xlurette.sh.in
install/xlurette.bat.in
install/set_env_var.in
install/lurettetop.sh.in
install/gen_stubs.sh.in
source/exp.ml
source/luc4ocaml.mli
source/luckyDraw.mli
source/Makefile.luc4ocaml
source/Makefile.luckyDraw
examples/lucky/external_88code/foo.c
examples/lucky/external_code/call_external_c_code.luc
examples/lucky/external_code/Makefile
VERSION
install/set_env_var.bat.in
examples/Makefile
examples/ocaml/crazy-rabbit/rabbit-dir-obstacle.luc
examples/ocaml/crazy-rabbit/Makefile
examples/ocaml/crazy-rabbit/rabbit.ml
examples/rml/Makefile
examples/lucky/other/Makefile
examples/luckyDraw/c/luckyDrawC_stubs.c.m4
examples/luckyDraw/c/essai.c
examples/luckyDraw/c/Makefile
examples/luckyDraw/ocaml/Makefile
examples/xlurette/heater/Makefile
examples/xlurette/Gyro/Makefile
utils/lucky_cpp
utils/luckytolatex
source/Makefile.release
source/version.ml
source/Makefile.luc2c
examples/xlurette/fault-tolerant-heater/Makefile
examples/xlurette/tram/Makefile
examples/xlurette/Sildex/Makefile
xlurette/makefile
xlurette/xlurette_glade_main.ml
lnsw
source/luc4c_stubs.h
source/luc4c_stubs.c
source/luc4c.ml
examples/lucky/C/Makefile
source/Makefile.luc4c
examples/lucky/lustre/Makefile
  Figthing with library stuff.
  Add headers (time-stamps)
  + other minor changes

examples/luckyDraw/c/luckyDrawC_stubs.h
  call the caml_startup() init procedure + various twix.

source/luc2c.ml
  Add a --luciole option

Project-Description: Lurette

Parent-Version:      1.43
Version-Log:         empty
Project-Description: Lurette

Parent-Version:      1.42
Parent-Version:      1.42
Parent-Version:      unstable_1_43.6
Parent-Version:      unstable_1_43.10
Parent-Version:      unstable_1_43.11
Version-Log:

Some progress in the scade tcl gui connection.

Add code to be able to parse type definitions that are coming
from the Scade tcl Gui. Thanks to that, we do not need to parse
generated C header files anymore.

Not yet finished.

Add a lot of files that were not prcs-ed.

source/util.ml
source/parse_luc.ml
source/lustreExp.ml
   «from_char_pos_to_line_and_col» that is useg to guess the line and col
number
   from the char number, was returning bad line numbers because cpp adds lines
   preceeded by «#». Therefore, we add boolean flag to this function so that it

   does not count such lines if the flag is set to true.

source/lurettetop.ml:
   Sim2chro and gnuplot were called with a bad arg in some cases (i do not
remember
   which ones. Probably on win32.

source/parse_c_scade.ml:
   Enhance the debug messages wrt line numbers.

source/lurette_exe.c
   Add ifdef-ed debug dump (in a log file).

install/configure.in:
Makefile.common.source:
Makefile
source/Makefile.show_luc
source/Makefile.lucky
source/Makefile
   Minor changes.

polka/caml/polka_lexer.mll:
   Allow ident starting by «_».

polka/Makefile.config
install/set_env_var.in
install/lurettetop.bat.in
   Minor changes.

xlurette/Scade/LuretteGUILauncher.tcl
xlurette/Scade/GenMake_l4sim.tcl
  The lurette dir in the Scade hierarchie is now «lurette/bin» instead of
  «lurette/win32/bin» (ditto for «lib», and «include»).

  Some fix after my visit to Esterel (thanks to Aubanel).

source/Makefile.luc4ocaml
   Add mli files.

xlurette/Scade/Makefile
   Add a few rules to copy various files at various places.

source/liblutin_c.c
   This a change from Pascal :
   « The interval algorithm was completely false ».

examples/Makefile
   Typo.

examples/ocaml/crazy-rabbit/rabbit.out
examples/lucky/other/test4.rif
examples/lucky/other/test4.res
examples/xlurette/heater/test3.rif0
   A few non-reg tests broke because of the changes to lutin_lib.c.

examples/luckyDraw/ocaml/Makefile
examples/xlurette/Gyro/Makefile
source/Makefile.release
   Some changes.

Project-Description: Lurette

Parent-Version:      unstable_1_43.10
Version-Log:

j'ai rajouté une une option --lustre a luc2c qui permet d'appeler
du lucky depuis lustre.

Project-Description: Lurette

Parent-Version:      unstable_1_43.9
Version-Log:

Project-Description: Lurette

Parent-Version:      unstable_1_43.8
Version-Log:         empty
Project-Description: Lurette

Parent-Version:      unstable_1_43.7
Version-Log:

Implement a luc4c API so that lucky programs can be called from C.
This is meant to be used be luc2c.

nb : not yet working.
There is a seg fault when launched from the scade tck gui. Strangely,
it seems to work when run from the command-line simulation executable.

Add code to be able to parse type definitions that are coming
from the Scade tcl Gui. Thanks to that, we do not need to parse
generated C header files anymore. This change comes from a manual merge
with the main branch.

source/parse_luc.ml
   Enhance the parser error msgs.

source/gen_stubs_common.ml
   Implement a typdef parser for type definition that are coming
   from the Scade tcl Gui (this change comes from a manual merge
   with the main branch).

polka/caml/polka_lexer.mll
   Allow ident starting by _.

source/gen_fake_lucky.ml:
   add a log file for gen_fake_lucky, which is useful for debugging,
   but also for users.

source/type.mli:
source/type.ml
   When generating, C code, translate ocaml float into C double instead
   of C float.

source/gnuplot-rif.ml
   Allow true, false, True, and False as RIF data.

source/liblutin_c.c:
   This a change from Pascal :
    The interval algorithm was completely false .

source/graphUtil.ml
source/graphUtil.mli
   Add a top_sort function that performs a topological sort in a graph
   with no cycle.

source/luc2c.ml
   More work on that.

source/luc_exe.ml
source/lurette.ml
source/lucky.ml
source/util.ml
source/parse_c_scade.ml
install/configure.in
OcamlMakefile
Makefile
source/Makefile
source/luc4ocaml.ml
source/Makefile.release
   Some Trivial changes.

NEW FILES:
source/luc4c_stubs.h
source/luc4c_stubs.c
source/luc4c_caml.ml
source/luc4c.mli
source/luc4c.ml
   The luc4c API files.

examples/lucky/C/test.rif.exp
examples/lucky/C/test.res
examples/lucky/C/foo.luc
examples/lucky/C/call_foo.c
   A few non-reg tests broke because of the changes to lutin_lib.c.

examples/lucky/C/Makefile
   Some changes.

Project-Description: Lurette

Parent-Version:      unstable_1_43.4
Version-Log:         empty
Project-Description: Lurette

Parent-Version:      unstable_1_43.3
Version-Log:

Ditto.

Project-Description: Lurette

Parent-Version:      unstable_1_43.2
Version-Log:

Ditto.

Project-Description: Lurette

Parent-Version:      0.133
Version-Log:

Various changes related to the packaging of lurette.

Project-Description: Lurette

Parent-Version:      0.132
Version-Log:

source/gen_stubs.ml:
source/ocaml2C.idl:
   use double instead of floats at the C level.

Project-Description: Lurette

Parent-Version:      0.130
Version-Log:

Remove the need of the ocaml compilers to be able to use lurette.
The idea is to link the final executable withh the C compiler instead
of using ocamlopt.

source/lurette.ml:
source/gen_stubs.ml:
  The stubs files that are generated are ow completely different.
  The process is much more simple BTW. Now, idl files are no more
  generated dynamically. The only files that are generated are C files
  that interfaces the sut and the oracle with lurette.

/source/lurettetop.ml:
   Add the USER_TESTING_DIR env variable.

Project-Description: Lurette

Parent-Version:      0.105
Version-Log:

source/lurettetop.ml:
   Add a --prefix <string> option that is added before the call
   to lurette (usefull, eg, for timing purpuses).

source/lurettetop.ml:
source/lurette.ml:
source/command_line.ml:
ihm/xlurette/*
  Add a ---show-step options that let one control whether or not
  the step number is printed.

Project-Description: Lurette

Parent-Version:      0.57
Version-Log:

Clean up Makefiles for lurette and ima_exe.

Make sure that gen_stubs does not recompile the fake oracle
(always_true.lus) if it is not necessary.

Project-Description: Lurette

Parent-Version:      0.56
Version-Log:

Add a --output options to specify the output file name.

Project-Description: Lurette

Parent-Version:      0.54
Version-Log:

Allow equalities between formula by translating [A = B] into
[Or(And(A,B), And(Not(A), Not(B)))].

Fix a bug where lurette had a bad behaviour whenever a toss
was performed on a domain which size was bigger than max_int
or max_float.

Fix a bug where lurette was not handling env with both boolean and
numerical as output vars.  Also add a few comments in solver.ml along
the way.
Adds a local bool as an output of heater in order to test the fix
above.

Project-Description: Lurette

Parent-Version:      0.53
Version-Log:

Cosmetic change:

Encapsulates the various fields of Env_state.

Project-Description: Lurette

Parent-Version:      0.51
Version-Log:

Remove the necessity to provide an oracle when the user does
not have any assertion to check. To do that, I parse the sut
file and automatically generates a fake oracle that answers always
true.

Project-Description: Lurette

Parent-Version:      0.47
Version-Log:

Add support for (boolean and numeric) if-then-else.

Remove the eval.ml module as the replacement of pre and input
vars is now done in Solver.formula_to_bdd.

Add a new module gne.ml that provides a garded normal expressions
data structure that is used to handle if-then-elses.

Also introduce a new field to env_state called bdd_tbl_global which
is used to cache the result of formula_to_bdd for bdds that do not
depend on pre nor input vars. The field bdd_tbl which cache the other
reults is now cleared at each step because it is useless to store results
that depends on pre and inputs.

Project-Description: Lurette

Parent-Version:      0.39
Version-Log:

Change the naming convention of prevars. Now, _pretoto is noted
_pre1toto, and _pre_pre_pretoto is noted _pre3toto.

Also fix a bug in the default initialization of ranges : min_float
is the smallest positive float, not the smallet one !!

Project-Description: Lurette

Parent-Version:      0.38
Version-Log:

Use the new cudd interface based on camlidl (mlcuddidl).

Now, bdds do not leaks anymore.  Remove debugging stuff that was used
to track that problem.

Project-Description: Lurette

Parent-Version:      0.35
Version-Log:

Add support to handle numerical contraints. The numerical stuff is
untested, but the boolean part works as before.

source/rnumsolver.mli:
source/rnumsolver.ml:
    New files implementing a numeric solver based on intervals.

source/*.ml:
    Change the var_type type from a string to a sum type containing
    the lower and upper bounds of numeric domains.

    Do not maintain a var_name to index correspondance, but a formula to
    index one. This is to be able to handle numerical constraint in bdds.
    This field is not set at init time anymore, but in formula_to_bdd.

    Add <>, <, and <= to the type of formula.

Project-Description: Lurette

Parent-Version:      0.33
Version-Log:

Fix a bug when computing the power of two.

Project-Description: Lurette

Parent-Version:      0.32
Version-Log:

Avoid recomputing everytime the list of output vars by storing it
in an env_state field.

Project-Description: Lurette