[Bug libstdc++/54757] FAIL: ext/random/beta_distribution/cons/default.cc (test for excess errors)

["daniel.kruegler at googlemail dot com" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54757

Daniel KrÃgler <daniel.kruegler at googlemail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel.kruegler at
                   |                            |googlemail dot com

--- Comment #3 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-09-30 09:44:27 UTC ---
Hi,

> There is an implementation of hypot, so I'm wondering if we can't do  
> better.

You mean in the libc? It's possible because as you can see the autoconf test
handles all the C99 mathematical functions together and if only one is missing
the check fails. Now, really, I don't think we want to go fine grained to the
level of the single function, but if for example we figure out that a number of
targets has available the entire subset of functions required for <random> and
<ext/random> we could separate those. My point is that we should try to find
patterns, like the long double functions are causing troubles and we can
separate those. As I said, for those features - isn't just about math - I would
rather not fragment everything down to single function, if possible.

> Testing suggestion.

Great, if it works, let's go with that for now. Then, longer term, if you have
ideas about the above, please also feel free to send a message to the library
mailing list.