Re: gparamspecs.c param_double_validate() doesn't support NaN/Inf?

[Andrew Paprocki <andrew ishiboo com>]

On Sat, Jan 10, 2009 at 7:06 AM, Brian J. Tarricone <bjt23 cornell edu> wrote:
> But I don't disagree that NaN or +/- inf should be allowed, assuming
> a restriction such as the above isn't in place for the particular
> instance of GParamSpecDouble.

This brings up an interesting question.. if the GParamSpecDouble
minimum/maximum are -G_MAXDOUBLE/G_MAXDOUBLE, should NaN/Inf be
allowed, otherwise not? That would allow a patch to add the
functionality without adding new fields to the GParamSpecDouble
structure. What does everyone think? Explicit, modifying the
structure, or implicit based upon the existing values?

-Andrew