\pgfmathfloatparse is a copy of \pgfmathparse when /pgf/fpu=false

[schtandard]

When the fpu library is not "installed", i.e. \pgfmathparse and friends have their normal behavior, \pgfmathfloatparse is a copy of \pgfmathparse, i.e. it does not parse floating point numbers.

\documentclass{article}

\usepackage{tikz}
\usetikzlibrary{fpu}

\begin{document}

\pgfmathparse{1.234e-56}\pgfmathresult

\pgfmathfloatparse{1.234e-56}\pgfmathresult

\pgfkeys{/pgf/fpu}

\pgfmathparse{1.234e-56}\pgfmathresult

\pgfmathfloatparse{1.234e-56}\pgfmathresult

\end{document}

outputs

0.0
0.0
1Y1.234e-56]
1Y1.234e-56]

but should output

0.0
1Y1.234e-56]
1Y1.234e-56]
1Y1.234e-56]

[hmenke]

Well, it does parse floating point numbers, it just doesn't output them as FPU floats. I'm not sure what the correct behaviour should be here, because FPU floats are pretty much useless when the fpu is not enabled.

[schtandard]

Does it?

\pgfmathfloatparse{1.234e-56 * 5.25e57}\pgfmathresult

does not work with fpu disabled either, so it can't just be the output formatting.

I don't agree that the fpu functions are useless separately. They can be used to parse floating point numbers without disturbing the usual drawing functions (when the floating point numbers are not to be used for drawing directly). In the current state, one has to locally activate fpu for each calculation.

I thought that this was the whole point of having the float versions of all the math macros. If \pgfmathparse and pgfmathfloatparse do the same thing regardless of if fpu is loaded or not, what's the point of having both names?