Seeming g++-specific "feature deficit"...

[Jonathan Wakely]

On 4 October 2014 14:02, Linda A. Walsh <gcc@tlinx.org> wrote:
> Marc Glisse wrote:
>>>
>>> So why not under gnu c++?
>>
>>
>> Because that is a non-standard function, obviously...
>
> ----
>         So valarray isn't really part of the the
> STL and that's why many STL functions won't work with
> it?

It was not part of the STL, and is in the Numerics clause of the
standard, not the Containers clause.


>>
>> http://en.cppreference.com/w/cpp/numeric/valarray
>> Look for "begin".
>> You can write begin(group) but not group.begin().
>
>
>
> I can see I'd grouped valarray in with the STL
> containers... and thought it was part of the same
> paradigm.  I'm just beginning to see how much of the
> STL  functions won't work on it...
>
> Ick, I'll have to remember how much valarray doesn't
> fit into the STL standard container group... it really
> wasn't really that obvious how different it was from the
> rest for someone just being introduced to all of them
> at once.


As I'm pretty sure I've said before, you seem to be trying to use
valarray where it isn't suitable.