LADSPA - Fix various floating point errors

[zonkmachine]

Fixes NaN in project from closed issue #3685
Also, some fixes salvaged from #3313

[tresf]

@zonkmachine thanks! Just a reminder that once merged, we should PR against https://github.com/swh/ladspa with the same fixes.

[zonkmachine]

Last fixes. Adapted coding style to the rest of the file and swapped fmaxf() with std::max().
I'll merge this later. It's basically just one-liners and well tested.

Just a reminder that once merged, we should PR against https://github.com/swh/ladspa with the same fixes.

I'll try and build those pluggins first. Last time I looked at it I failed miserably.

[tresf]

I'll try and build those pluggins first. Last time I looked at it I failed miserably.

This should ease your struggles: swh/ladspa#55

Besides the Travis-CI capabilities, it fixes a few compilation problems I ran into while working on #3931. They haven't been merged yet so they're sitting as duplicates in the Travis-CI PR.

Pay close attention to the apt-get ... commands in travis.yml, it has the dependencies you'll need incase you don't already have them on your system.

Lastly, make install prefix=$HOME/foo is important if you don't want to run the install as root and avoid the default /usr/lib/[ladspa] prefix.

[tresf]

PR submitted swh/ladspa#57. Second set of eyes appreciated.