Re: lujit - Malformed number

[Florian Weimer <fw@...>]

* Mike Pall:

> Anyway, since the only popular decimal point besides '.' is ',',
> I could add a heuristic workaround for that. Except that it would
> slow down number parsing if a non-C locale is set (and uglify the
> code a lot). Or write my own replacement for strtod(), which is
> far from trivial, but would solve other compatibility problems.

There is dtoa for this purpose, see: <http://www.netlib.org/fp/>
It's quite portable.

What I would like to see is that n = tonumber(tostring(n)) holds for
all finite numbers, and that tostring(n) returns the shortest possible
representation such that n = tonumber(tostring(n)) holds.  I believe
the Python folks put a bit of effort into that recently (in version
3.1 at least).