bug#37140: 27.0.50; calc gives "Wrong type argument: number-or-marker-p, inf" to many operations

[Eli Zaretskii]

> From: Mattias Engdegård <mattiase@acm.org>
> Date: Thu, 29 Aug 2019 12:32:42 +0200
> Cc: 37140@debbugs.gnu.org, huszaghmatt@gmail.com, npostavs@gmail.com
> 
> +#if defined __FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__
> + #error Emacs cannot be built with -ffinite-math-only
> +#endif

Thanks, but I question the wisdom of signaling an error in some
arbitrary place.  If we don't want such an Emacs to run at all, we
should test this at startup and refuse to continue.  Errors such as
the above could, under some unfortunate circumstances, cause an
endless loop of errors being signaled, with the user unable to do
anything but kill the process (and potentially lose edits).