standards/175811: libstdc++ needs complex support in order use C99

David Schultz das at
Fri May 31 03:38:13 UTC 2013

On Thu, May 30, 2013, Stephen Montgomery-Smith wrote:
> On 05/30/2013 12:13 PM, Steve Kargl wrote:
> > What I find appalling is reading "people are tired
> > of the situation with libm, so I'm  going to commit
> > some atrocious hack".   The proper response should be
> > "so I'm going to help implement and test the missing
> > functionality".  It's unfortunate that only a few
> > individuals are working to fix libm, but such is
> > life. 
> I don't think the problem is that there are too few individuals.  I
> think the problem is that the standards are set too high.  I presented
> numerically accurate complex arc-trig functions a long time ago, and I
> became increasingly frustrated at the lack of progress.
> I am pleased that it got committed a few days ago.
> But I feel that the change requests, particular the style change
> requests, became too much.  I dutifully complied with the many style
> changes, but it became overwhelming.

Bruce is very meticulous and has a lot of good feedback, but it's
important to understand that Bruce's standards are not the minimum
standards for committing a change.  Bruce doesn't commit directly
anymore in any case.  I don't think I have ever committed a change
that Bruce could find no flaws in, including patches submitted by
Bruce himself. :) It's okay to commit some working code first and
then improve it later.

More information about the freebsd-numerics mailing list