Re: [bug-mes] ARM mes on wip-arm - Other test failures

[Jan Nieuwenhuizen]

Danny Milosavljevic writes:

Hi Danny,

> What is it that lib/tests/scaffold/7l-struct-any-size-array-simple is
> testing?

I had a long look and it doesn't make much sense to me.  The use of `-1'
together with char and packed suggests testing for overflows, one field
overwriting the other; or sign extension bugs (-1 char becomes 255 int).

Looking at the test both of these hypotheses do not make sense...so i'm
a bit confused.

> Would it be OK to change it via the following patch--or does it break
> what the test was supposed to accomplish?

I'm a bit hesitant about changing and keeping a test that we don't
understand.  I'm inclined to either throw it away, or leave it
unchanged.  In any case, I don't have a strong opinion here.

I think it would be nice if esp. all tcc specific tests would be smaller
and test more orthogonal features.  Of course we need to test feature
interaction too but hmm.  WDYT?

> diff --git a/lib/tests/scaffold/7l-struct-any-size-array-simple.c 
> b/lib/tests/scaffold/7l-struct-any-size-array-simple.c
> index cbfab19a..c68bf89b 100644
> --- a/lib/tests/scaffold/7l-struct-any-size-array-simple.c
> +++ b/lib/tests/scaffold/7l-struct-any-size-array-simple.c
> @@ -42,9 +42,9 @@ main ()
>    tab14[1].a = -1;
>    tab14[1].b = -2;
>    tab14[1].c = -3;
> -  tab14[1].d = -4;
> +  tab14[1].d = 4;
>  
> -  if (tab14[1].d != -4)
> +  if (tab14[1].d != 4)
>      return 1;
>  
>    return 0;
>

-- 
Jan Nieuwenhuizen <address@hidden> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com