Increase MAXSTUFF yet again.

emphasis tag.
Use new sfStrDrawTJ (with transform & justification) for text.
Yeow, handle polymax properly -- default == infinity!
Add "vcmap" for per-datafield colormaps.
Enable GL_SMOOTH mode to get texture colors to work right (huh? why do we need to?).
"tfm c"/"tfm w" choose cam vs. world (default) location for object tfms.

have "o2w" transforms that are really interpreted relative to camera space.
Also start adding per-object clip support; this requires change to
subcameras, supplying projection parameters subc_lrbt()
rather than a (clip-dependent) projection matrix.

where ``a = b;  if(a == b) { ... }'' can be false!

rather than a string.

popen() needs "r" not "rb" -- no such thing as binary mode.
kira_open() gets original filename if we can't find it, for error messaging.

Star type is number-encoded, sigh:
	0 = unknown
	1 = ms	Main Sequence
	2 = cd	Compact Dwarf??
	3 = gs	??
	4 = bh	Black Hole
	5 = sg	Supergiant?
	6 = hb	Horizontal Branch

Now, the only #ifdef KIRA/WARP needed in partibrains.c are
the data-command initialization routines.  All others,
including control-command parsing and specialized drawing,
is now via a function table.

Don't try to allocate in local memory -- use shmem after all.

Put colorseq/sizeseq/threshseq invalidation in kira_parti, not here,
since we needn't do it when warping.
By default, clock range is whatever it had been set to before.

Add seed counter to struct stuff.

Initialize all dynamic-data stuff explicitly.

we're not using them yet.

get_dsp_context ad MAXDSPCTX.

Ellipsoids allow 3-component (Rx Ry Rz) orientation numbers too.
st->dyndata doesn't mean we should trash the anima[][] specklist.
Need to keep it intact for warping.

So if we're off the end, or if a null warp is in effect,
don't mess with it.

differential rotation function.  See "warp" command in partibrains.c.

Finally parse snapshot arguments properly...?