C11 conformance of casinl-like functions.
Bruce Evans
brde at optusnet.com.au
Wed Feb 8 11:38:48 UTC 2017
On Wed, 8 Feb 2017, mokhi wrote:
> I recently saw bunch of PRs opened about C11 lack of conformance in
> FreeBSD on Bugzilla, complaining cosinl, acosinl, ... not implemented.
I think you mean acosl, asinl, ...
These were implemented quite well in 2012-2013, but not quite finished,
and not committed. Only the float and double version were committed.
The raw versions are still available in
https://people.freebsd.org/~stephen/catrig*.c. These have rotted and
require some editing. Compare with the committed parts to see most
of the necessary editing.
> I've searched about these and I found documents about them too[2][3].
> Do you think we should implement them? Or standards doesn't have
> suggestions on these?
>
> If yes (you think we should implement them), would you suggest simply
> `strong aliasing symbols of FUNC_l to FUNC`?
They cannot be implemented by aliasing (except on arches which don't
have real long doubles). Bad versions can be implemented using wrappers.
Bad versions are worse than none IMO.
> or implementing FUNC_l
> from scratch?
It would be hard to do better than the 2012-2013 implementation.
Bruce
More information about the freebsd-numerics
mailing list