problems with end of line hyphen protrusion

[eroux]

The calculation of end of line hyphen protrusion introduced for #845 seem to give strange results in some cases (well, I admit I suspected it was not perfect because of the 50 value I had to give it), for instance in this example (which I don't really manage to "minify"):

on the second line, the space between the note and the custos is too large compared to the first line, and there is no reason for it. The bug disapears with \gresethyphenprotrusion{100}.

Also, this example:

\documentclass[12pt]{article}
\usepackage{fontspec}
\usepackage{geometry}
\geometry{a4paper}
\usepackage{gregoriotex}
\usepackage{libertine}
\begin{document}
\grechangecount{endofsyllablepenalty}{-500}%
\grecreatedim{spacebeforeeolcustos}{0 cm plus 0 cm minus 0 cm}{scalable}%

50:

\gabcsnippet{(c3) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) ammmmmmmmmmmmm(g)g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) (g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) Sal(g)va(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g1(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g)}

100:

\gresethyphenprotrusion{100}

\gabcsnippet{(c3) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) ammmmmmmmmmmmm(g)g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) (g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) Sal(g)va(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g1(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g)}

0:

\gresethyphenprotrusion{0}

\gabcsnippet{(c3) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) ammmmmmmmmmmmm(g)g(g) g(g) g(g) g(g) g(g) g(g) g(gg) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) (g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) Sal(g)va(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g1(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g) g(g)}

\end{document}

shows that the computation does not work as intended...