TR1 math and documentation patch

[Paolo Carlini]

Smith-Rowland, Edward M wrote:
> This patch fixes some doxygen issues with the TR1 special functions.
>   
Ok, thanks. The doc changes have to go in before the documentation freeze.
> The progress chart is updated to reflect completion of special functions.
>   
The progress chart is already ok, then, or I'm just not getting your
point?!?
> I also checked off a lot of regex.  It looks like this is mostly or all done.
> Folks who know more about this might give this a look.
>   
Unfortunately mostly just a boilerplate, if I remember correctly.
> There are two substantive things: some constification and a change to allow the branch point x=1 in the hypergeometric function.
>   
Generally, we don't constify by-value parameters. Also, frankly, I'm a
little scared by the substantive change, because we learned in the past
that regressions appear easily on less common targets... Are the
existing testcases heavily stressing that case? On hoe many targets did
you regtest?

Paolo.