Mattias,
The issue is that the 94_generic.c test expected results are machine
dependent. This modified version below is more neutral and works on RPI.
However it tests fewer cases.
You should think of a better tests that deals in a portable way with
const/non-const, signed/unsigned and different integer sizes you can meet on
different architectures.
jullien@sims3:~/tcc $ git diff tests/tests2/94_generic.c
diff --git a/tests/tests2/94_generic.c b/tests/tests2/94_generic.c
index 0655dd4..d9db14b 100644
--- a/tests/tests2/94_generic.c
+++ b/tests/tests2/94_generic.c
@@ -20,6 +20,8 @@ int b_f()
return 10;
}
+typedef unsigned char * strtype;
+
#define gen_sw(a) _Generic(a, const char *: 1, default: 8, int: 123);
int main()
@@ -27,7 +29,7 @@ int main()
int i = 0;
struct b titi;
const int * const ptr;
- const char *ti;
+ const strtype ti;
i = _Generic(a, int: a_f, const int: b_f)();
printf("%d\n", i);
@@ -41,11 +43,11 @@ int main()
printf("%d\n", i);
i = _Generic(17, int :1, int **:2);
printf("%d\n", i);
- i = _Generic(17L, int :1, long :2, long long : 3);
+ i = _Generic(17L, int :2, long :2, long long : 2);
printf("%d\n", i);
- i = _Generic("17, io", const char *:1, char *:3, const int :2);
+ i = _Generic((strtype)"17, io", const strtype:1, strtype:3, const int
:2);
printf("%d\n", i);
- i = _Generic(ti, const char *:1, char *:3, const int :2);
+ i = _Generic(ti, const strtype:1, strtype:1, const int :2);
printf("%d\n", i);
return 0;
}
-----Original Message-----
From: uso ewin [mailto:[email protected]]
Sent: vendredi 7 juillet 2017 18:41
To: Christian JULLIEN; [email protected]
Subject: Re: [Tinycc-devel] _Generic or __builtin_choose_expr
On Fri, Jul 7, 2017 at 4:40 PM, Christian JULLIEN < <mailto:[email protected]>
[email protected]> wrote:
> Just tested your
> git clone -b mob <https://github.com/cosmo-ray/tcc.git>
> https://github.com/cosmo-ray/tcc.git
>
> On RPI, I now get:
>
> --- ../../tests/tests2/94_generic.expect 2017-07-07 16:28:08.010486068
> +0200
> +++ 94_generic.output 2017-07-07 16:29:51.560466233 +0200
> @@ -1,9 +1 @@
> -20
> -20
> -123
> -2
> -5
> -1
> -2
> -3
> -1
> +94_generic.c:46: error: _Generic sellector of type 'char *' is not
> compatible with any assosiation
>
> Also note that sellector has only one 'l' => selector
>
> Le : 06 juillet 2017 à 14:24 (GMT +02:00) De : "uso ewin"
> < <mailto:[email protected]> [email protected]> À : "
> <mailto:[email protected]> [email protected]"
> < <mailto:[email protected]> [email protected]> Objet : Re:
> [Tinycc-devel] _Generic or
> __builtin_choose_expr
>
>
>
> On Thu, Jul 6, 2017 at 2:09 PM, uso ewin < <mailto:[email protected]>
> [email protected]> wrote:
>> On Thu, Jul 6, 2017 at 6:11 AM, Christian Jullien <
>> <mailto:[email protected]> [email protected]>
>> wrote:
>>> On Windows 32bit as with RPi I get:
>>> Test: 94_generic...
>>> --- ../../tests/tests2/94_generic.expect 2017-07-06
>>> 06:06:36.520552800 +0200
>>> +++ 94_generic.output 2017-07-06 06:07:53.359318000 +0200
>>> @@ -1,9 +1 @@
>>> -20
>>> -20
>>> -123
>>> -2
>>> -5
>>> -1
>>> -2
>>> -3
>>> -1
>>> +94_generic.c:44: error: type march twice
>>> make[2]: *** [Makefile:49: 94_generic.test] Error 1
>>>
>>
>> Ok the error happen because parse_btype, does never return VT_LONG.
>> So I guess if a _Generic expression contain a long long and a long in
>> 64bit the error would happen too.
>>
>> I guess a solution would be either a global variable to force
>> parse_btype to let long as long, or to patch every functions that
>> call parse_btype so it take another argument which is here to force
>> parse_btype to keep long as long.
>>
>> sorry for adding a bug.
>>
>> Matthias
>
> I've push the fix on my mob:
> <https://github.com/cosmo-ray/tcc/commit/fb26db96828291a9c4ae94394f1aa4>
> https://github.com/cosmo-ray/tcc/commit/fb26db96828291a9c4ae94394f1aa4
> d1289ae14f
>
> I've add a small commit so generic unqualify volatile too:
> <https://github.com/cosmo-ray/tcc/commit/e87939bfd385f412375f8a599e0b32>
> https://github.com/cosmo-ray/tcc/commit/e87939bfd385f412375f8a599e0b32
> 914b5e8488
>
> If your okay with that, I push it on mob.
>
> Matthias
>
> _______________________________________________
> Tinycc-devel mailing list
> <mailto:[email protected]> [email protected]
> <https://lists.nongnu.org/mailman/listinfo/tinycc-devel>
> https://lists.nongnu.org/mailman/listinfo/tinycc-devel
>
>
> _______________________________________________
> Tinycc-devel mailing list
> <mailto:[email protected]> [email protected]
> <https://lists.nongnu.org/mailman/listinfo/tinycc-devel>
> https://lists.nongnu.org/mailman/listinfo/tinycc-devel
>
typo fix on my github
I'm sorry but I don't have any arm machine, does someone have an idea how to
fix, or debug this ?
I guess parse_btype or type_decl return the wrong type at line 4726 of tccgen.c
_______________________________________________
Tinycc-devel mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
