Re: [bug-gawk] Overflow to Infinity

[arnold]

Eli Zaretskii <address@hidden> wrote:

> > I offer you the opportunity to submit a patch as proof of my open-
> > mindedness and that it's not just a "waste of everybody's time."
>
> Sorry, I didn't mean to annoy, let alone offend.  I meant "waste" in
> the positive meaning: I'd like to avoid asking you and others to
> invest your time in reviewing patches if it is known in advance that
> the basic assumption behind the patches is not acceptable, because
> we'd just go back to arguing on those assumptions.
>
> I think we should wait for your analysis of the output I sent (I'm
> here to help and answer questions, of course), before we move to
> discussing further actions.  Does that make sense?

Your C output distinguishes -nan from +nan but apparently the use of
signbit() doesn't work as I expect on MinGW.

I am going to back down from messing with this for a little while.
Instead I am going to commit my planned changes to a branch for safekeeping.

Eli, I'd appreciate your reviewing things in depth and deciding if you wish
to produce a patch or not. If you do, we'll put it in another branch for
safekeeping and then review both yours and mine later, when the issue
becomes unbearable.

Thanks,

Arnold