of colormap entries.

And it's been a while since 0.7 anyway.

such that texture objects need to be recreated.

If there's an 8th numeric field in a .wf file, take it to be a focallen distance.

Use Toshiyuki Takahei's FLHACKs.  Reorder functions so we don't
need quite so many of them.  Restructure msg() and warn() likewise.

Set an initial display context number so Gview can track future window changes.

if some were fixed-oriented and others screen-facing.
Portability tweak from Toshi -- be even more pessimistic about stack size.

or (new) little-endian binary.  Good for Toshi's constellation figures,
which are in little-endian binary format.

gets changed (as between stereo/mono, or window-detachment).
Pass the news to the texture layer so we don't lose textures.
Or at least so we can arrange not to lose em, not sure this works yet...

Add hooks for Toshiyuki Takahei's FLHACK FLTK stubs.

it seems to get the data completely wrong (most X's = 0, etc.).
If someone ever figures out why, I'd be curious to know.

Move colorpatch::draw()'s body here.

if sqrtf() isn't available.  This is 'way better than using -Dtanf=tan etc. in Makefile.

Just leave Steve Marx's insistence on FLTK >=1.1.x.
Check for values.h.
If we seem to be running on MacOS X, add more special tests:
   Search -lmx for sqrtf() etc., rather than a bunch of -D'efines for float functions.
	If they (or at least sqrtf) are not present, geometry.h now #define's em.
   Insist on {fltkdir}/FL/mac.r so we can use Rez to tag partiview as a GUI app,
	otherwise it can never get keyboard/mouse focus.

needed on MacOS X.