| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|\| |
|
| |
| |
| |
| |
| |
| | |
In order to allow introducing structs in parameter definitions the
environment must keep the type information.
Bug 19602
|
| |
| |
| |
| |
| |
| | |
If a declaration of a composite is encountered it is also tested
if the kind is equal.
Bug 19630.
|
| |
| |
| |
| |
| |
| | |
Since some incomplete types are allowed in initialization just
test whether the default initilization exists.
Bug 19601
|
| |
| |
| |
| |
| |
| | |
New types introduced in casts are now inserted into the right
Environment and carried along.
Bug 19614.
|
| |
| |
| |
| | |
This reverts commit c64c4ab2526ad87a3506c9e1fdf31fa1446c16eb.
|
| |
| |
| |
| |
| | |
Allows adding struct definitions in function parameters.
Bug 19602.
|
| |
| |
| |
| |
| |
| | |
Initializers for function variables are not allowed. CompCert now
reports an error and exits.
Bug 19606
|
| |
| |
| |
| |
| |
| | |
Before the initializazion is computed we check wether the type is
incomplete.
Bug 19601
|
| |
| |
| |
| |
| |
| | |
CompCert now reports an error for usage of the va_start macro in
non variadic functions.
Bug 19600.
|
| |
| |
| |
| |
| | |
Parameters also need to be checked for incomplete types.
Bug 19596
|
| |
| |
| |
| |
| |
| | |
Alginof and sizeof applied to incomplete types now exit earlier
with a fatal error.
Bug 19594.
|
| |
| |
| |
| |
| |
| | |
This allows problems in elaboration of the initializers for
variables of void type.
Bug 19577.
|
| |
| |
| |
| |
| |
| |
| | |
Instead of a warning for an empty union CompCert reports an error
and exits. This avoids problems during the generation of
initializers for these.
Bug 19565.
|
| |
| |
| |
| |
| |
| | |
The exception Wrong_attr_arg raised is now catched during the
translation of the wrong _Alignas attributes.
Bug 19568.
|
| |
| |
| |
| |
| |
| | |
Added a check for errors after the elab phases to avoid problems
in the transformations due to broken input programs.
Bug 19504
|
| |
| |
| |
| |
| |
| | |
Return with a expression that is not compatible with the given return type
of a function now causes and fatal error, to avoid problems with later
transformation passes depending on it.
|
| |
| |
| |
| |
| |
| | |
This avoids introducing line breaks during printing of function
types.
Bug 18004
|
|/
|
|
|
|
|
|
|
|
| |
Now each warning either has a name and can be turned on/off, made
into an error,etc. or is a warning that always will be triggered.
The message of the warnings are similar to the ones emited by
gcc/clang and all fit into one line.
Furthermore the diagnostics are now colored if colored output is
available.
Bug 18004
|
| |
|
|
|
|
|
|
| |
In a call such as "f(expr, ..., expr)", if the identifier "f" is not declared, declare it as specified in the ISO C90 standard, namely like "extern int f()" would do it. Previously, the declaration was done "on the side" and not properly recorded in the environments.
The diff is relatively large because the "enter_or_refine_ident" had to be moved up in the file.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, CompCert incorrectly handles 'extern' and function
declarations within a block. For example:
int main(void)
{
{
extern int i;
i = 0;
}
{
extern float i;
i = 10;
}
}
CompCert fails to detect the inconsistent declarations of "i" in the
two blocks, simply because the first declaration is not in scope when
the second declaration is processed.
This commit changes the elaboration of file-scope declarations,
block-scope "extern" declarations and block-scope function
declarations so that they are checked against possible earlier
declarations found in the already-elaborated top-level declarations,
instead of in the current typing environment.
Owing to the lifting of block-scoped extern and static declarations to
the top-level already performed by the elaboration pass, the
already-elaborated top-level declarations give an accurate view of the
declarations of variables with internal or external linkage already
encountered and processed, even if they are not currently in scope.
|
|
|
|
| |
conversion
|
|
|
|
|
|
|
|
| |
This commits handles the case where the argument is passed with a type different from the actual type of the argument, as in
float f (x) float x; { return x; }
"x" is passed with type "double", and must be converted to "float" at the beginning of the function.
|
|
|
|
|
|
|
|
| |
Most of the code can be String.uppercase usages can either be
replaced by a more specialized version of coqstring_of_camlstring
(which is also slightly more effecient) or by specialized checks
that reject wrong code earlier.
Bug 19187
|
|
|
|
|
|
| |
Since casts, sizeof, etc. expressions can introduce new types
we also need to add these to the environment and pass it through.
Bug 17814
|
|
|
|
| |
(§6.7.8), bug 18000
|
|
|
|
| |
(§6.7.8), bug 18000
|
| |
|
|
|
|
| |
strings, bug 18000
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
CompCert now recognizes the C11 _Noreturn function specifier and
emits a simple warning for functions declared _Noreturn containing
a return statement. Also the stdnoreturn header and additionally
the stdalign header are added.
Bug 18541
|
|
|
|
|
|
| |
The code was mostly there for documentation effort. So warning
27 is deactivated again.
Bug 18349
|
|
|
|
|
|
|
|
|
|
|
| |
Since the menhir version we use requires ocaml>4.02 we can also
upgrade the required ocaml version to >4.02 and remove the
deprecate String functions.
Also we now activate all warnings except for 4,9 und 27 for regular
code plus a bunch of warnings for the generated code. 4 and 9 are
not really usefull and 27 is deactivated since until the usage
string is printed in a way that requires no newline.
Bug 18394.
|
|
|
|
|
|
| |
Removed some unused variables, functions etc. and resolved some
problems which occur if all warnings except 3,4,9 and 29 are active.
Bug 18394.
|
|
|
|
|
|
| |
Since we cannot construct a default initializer for void types
we need to exit earlier.
Bug 18004.
|
|
|
|
|
|
| |
In parse_int it was not tested if the value of v is smaller than
zero. This allowed it that certain large integers were accepted
due to wrap around.
|
|
|
|
| |
Regression test added in regression/initializers.c
|
|\ |
|
| | |
|
|/
|
|
|
|
| |
- Added a Cabs.PROTO_OLD constructor to Cabs.decl_type
- Refactored the Parser.vy and pre_parser.mly grammars
- Rewritten the conversion of old function definitions to new-style
|
|
|
|
| |
handled as a function with void parameter.
|
|
|
|
|
| |
C11 allows a typedef redefinition if the types are the same.
We now allow this also and issue a warning and an error if the types are different.
|
|\
| |
| | |
Correction of a few bugs in the pre-parser, added comments.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
was not parsed correctly:
typedef int a;
int f() {
for(int a; ;)
if(1);
a * x;
}
Additionnaly, I tried to add some comments in the pre-parser code,
especially for the different hacks used to solve various conflicts.
|
| |
| |
| |
| |
| |
| | |
Instead of creating separate annotations for the local variables
we call the Debug.add_lvar_scope and we construct a mapping from
function id + scope id to scope information.
|
|\|
| |
| |
| |
| |
| | |
Conflicts:
debug/CtoDwarf.ml
debug/DwarfPrinter.ml
|
| |
| |
| |
| |
| | |
If they only report an error an assertion failure in Reame.ml was
triggered.
|
| |
| |
| |
| |
| | |
Instead of reimplementing the whole scope handling in the debug information
use the existing functionality and fill the scopes explicitly in the functions.
|