Re: [Tinycc-devel] _Generic or __builtin_choose_expr

[uso ewin]

On Fri, Jul 14, 2017 at 7:22 AM, Christian Jullien <address@hidden> wrote:
> It's nice to see new features coming on tcc, but it's also nice if they work
> everywhere
> _Generic author(s) can you please take time to fix failing test?
>
> Test: 94_generic...
> --- ../../tests/tests2/94_generic.expect        2017-07-10
> 23:00:46.045513740 +0200
> +++ 94_generic.output   2017-07-14 07:12:25.965274077 +0200
> @@ -1,9 +1 @@
> -20
> -20
> -123
> -2
> -5
> -1
> -2
> -3
> -1
> +94_generic.c:44: error: type march twice
>
> IMHO, your test depends on how different types are implemented in native
> machine. So it's more a test issue than an implementation issue.
> May be? Using 'portable' types like uint32_t may solve some issues. You may
> also solve const/non const issue signed/unsigned char by forcing an exact
> type OR by giving the same value whether string is const or not.
>
> In any cases, 94_generic should produce NO error on platform tcc supports.
> I'll be on vacation next two weeks and have limited time to investigate
> myself but I'll be glad to test changes on Windows 32/64 and RPi
>

Hello,

Sorry if I take time to answer,
So actually it's not a test issue, because gcc can do difference
between a 'char *' a 'signed char *', and a 'unsigned char *',
gcc doesn't consider int and a typedef of an int as the same type,
and most important it will find the good match to a long.

grischka seems to have fix char * problem,
I've just push a fix for the long on my github,
but it does refactor how VT_LONG is handle so I’m not very confident
on pushing it now on mob without a review
(this is why it take me time to answer).

I've also fix all type.

for the typedef, I guess it might work, but I didn't check further yet.