Re: [[patch] valarray (was Re: valarray_copy implementation question)

[Christian BRUEL <christian dot bruel at st dot com>]

I think I understand your point. dest[b] has 0 size because there were 0 
matching elements, correct ? your definition of size of dest[b] is the 
number of true elements.

What I found really strange with this interpretation is that the 
legality of the program depends on a runtime behavior

for instance if the boolean's value is not known at compile time, like with

...
b[0] = foo(); // with extern bool foo (void);
dest[b] = src;
...

Is this legal or not ? that's quite disturbing :-( I know a couple of 
proprietary and commercial testsuite that fail on this kind of test. 
(having n true elements < at the size of src).

Thanks so much and sorry for the noise.

Christian


Gabriel Dos Reis wrote:
> On Thu, Nov 26, 2009 at 5:58 AM, Christian BRUEL <christian.bruel@st.com> wrote:
> > what about
> >
> > #include <valarray>
> > using namespace std;
> >
> > int main(void)
> > {
> >  int n = 1;
> >
> >  valarray<int>  dest(n);
> >  valarray<bool> b(false, n);
> >  valarray<int>  src(1, n);
> >
> >  dest[b] = src;
>
> Christian,
>
>   the array expression `dest[b]' has length 0; the array expression `src'
> has length 1.  That violates the equal length precondition.
>
> >  return 0;
> > }
> >
> > is it a user error also ?
>
> yes; it is user confusion; the same as we discussed before.
>
> -- Gaby