Le 31/10/2014 21:56, Emil Lenngren a écrit :
> And this is done in the function JR_object_cast in gb_runtime.c, which is
> called from jitted code when objects are converted to other object types.
>
OK, I see it... mmm...
--
Benoît Minisini
--
And this is done in the function JR_object_cast in gb_runtime.c, which is
called from jitted code when objects are converted to other object types.
2014-10-31 21:55 GMT+01:00 Emil Lenngren :
> The problem is the signature of the _convert functions in the classes.
> Before they were OBJECT* _conve
The problem is the signature of the _convert functions in the classes.
Before they were OBJECT* _convert(CLASS* src, CLASS* class) but now they
are bool _convert(CLASS* src, CLASS* class, VALUE* value). The JIT calls
the function using the first one.
2014-10-31 21:49 GMT+01:00 Benoît Minisini :
>
Le 31/10/2014 21:40, Emil Lenngren a écrit :
> Hi!
> Actually this will also fail:
>
> Dim a As Integer[]
> a = ["1"]
>
> The problem is with the "new" _convert mechanism, introduced in
> http://sourceforge.net/p/gambas/code/4908/. The JIT code expects the older
> API, which therefore crash...
> I
Hi!
Actually this will also fail:
Dim a As Integer[]
a = ["1"]
The problem is with the "new" _convert mechanism, introduced in
http://sourceforge.net/p/gambas/code/4908/. The JIT code expects the older
API, which therefore crash...
I haven't found out how much refactor needs to be done to support
Hi Emil,
The following little project segfaults when the jit compiler is requested.
If you can look at it...
--
Benoît Minisini
test-jit-0.0.1.tar.gz
Description: application/gzip
--
__
