fixed wavechooser compiling,
change matrix to reset internel sequencer on reset instead of too first pattern
fixed posh sampler pitch changing on reset.

fixed scope and meter accordingly to properly update data and size.

made pause an info variable so jack(and other audio plugins) can sanely
handle outputing null/0 while paused, fixed jack change buffer and samplerate,
Pause/Play and Reset should now work properly.

critical fix for saving comments

as a result SSM will now always try and follow Jack settings while attached.

at any given point, so that changes to the audio subsystems can
be applied live(needed for jack).

this is also usefull in that it can give the ability for a Reset
button on the toolbar along side a play/pause button for being able
to at any point completely kill audio and reset to begining(usefull
for recording takes where you want it to start at specific point w/out
the delays echos etc overlapping for the first iteration.

and safely handle differing order sections/idents and missing etc.
This is so that it can handle driver specific settings in the future(eg. won't die on ALSA settings if alsa driver doesn't exist/missing entry etc).

This adds the ability to select multiple devices either by a ctrl-mouse drag
or by ctrl-click on individual devices title to do finer selection
and deselection, and then ctrl-right-click on the canvas to produce a menu
with the options to cut/copy/paste/delete selection or to merge an existing
patch entirely into the patch. Pasting or merging will auto-select new devices
so any old devices will lose selection.

To remove selection simply click anywhere on canvas.

Current problems -

Multiple Select drag is buggy and may select items not in the selection area.

Multiple OSS plugins are buggy so merging/pasting an OSS device can cause crashing.

This removes the For loop effectively cutting CPU usage in half.
Currently the actual behaviour apears to be exactly like that of the old
however it is possible that those values that go higher then the old table
method(more then 6 octaves up for eg.) will behave slightly different.
Thanks go to Kenny Graunke for working most of this out.

this adds a minimal sound to the filters if it falls below the range the spike usually happens(for me). This prevents huge CPU overhead when using multiple Formants and Filter Plugins on slow lines.