Add "selcounts()" function to count particles matching given SelOp pattern.

of command/history box.  Needs to be set before they're created,
apparently -- ctrl-K trick doesn't work so well.

Doesn't seem to work, though -- I still see the regular old arrow.  Why not?

able to track collections rather than single particles (not yet though).
Trails are relative to tracked-point position, so can be non-inertial.

Warp pointer under X.
And, justify numeric labels so they're always visible.

Offer selsrc2dest(), seldest2src().

Compute fake aggregate temp/luminosity for CM nodes.
Oops, marks need their own sel[] array!
Allow adjusting H-R diagram alpha values too.

And, report match counts from "sel", "thresh", "emph".

Pay attention to emph/see settings.
Get selection working properly.

    selsrc(), seldest(), selname(), seltoken().
"thresh" now accepts "-s 'destexpr'".
Add new selthresh SelOp.

Only use SEL_DEST-initialized pick etc. info.

Next: implement "see" in drawspecks().
Add "picked" callback which gets notified of all pick results.

Make specks_picker() work for either Gview or Plot windows.

Add set-selection features: SelOp bitmasks, etc.

Add set-selection features, including interaction tracing.

This really needs to be done in an intermediate superclass!
All these changes really in previous revision; just missed typing in comments then.

so we can use same function for both.

Hack for gcc 3.0's stdio (#ifdef NEWSTDIO).