Notes:

with regards to the build procedure: sparc config file needs help;
-Bstatic for LDFLAGS. otherwise, the linker tries to do the shared
objects trick, and most of the functions in Muse.cdefs will not be
linked in, hence will not be visible in the .SYM file.

we need to include the PlxVideo subtree in the distribution.  i'll try
to set this up real soon now ...

also, we still need a Tools dir, Demos dir, etc. like we started for
the 2.1 tree. and we need the version of 'gconfig' from the galateadev
locker.

BUG LIST
========

1. all $N strings in parse.y must become $N.s; sun yacc generates bad code

2. in activate_element, in the TEXT case, the
   second arg to GetWidgetID() is e->map, not 'm'

3. the Scribe widget dies in _doscan() when it encounters:

   "(David DelloRusso \050left\051 and Paul Nordquist,)"

   in report.PS, which is trying to show the string "(left)".
   
   why is this happy on other machs?

***Didn't fix this.

4. GetTextStringPointer() tries to use XtGetValues() to get the
   entire string in the widget; this returns 0 on sparc, causing
   program death. 

   It seems the problem was due to using XtNuseStringInPlace. Once
   the widget manages the string memory, the problem goes away.

   I also put back, but haven't tested, the stuff to edit files
   on disk. You would have to call XawAsciiSave(widgetID) to force
   the changes to the file, however. This should be tested with
   "No Recuerdo".

5. somehow, figed.dat kills the DECWindows window manager, dxwm (!!!).

***Didn't fix this. No clues.

6. in fn_misc.c, added "newelement" message for builder.

7. when packages with scrollbars get resolved more than once,
   it causes an infinite loop. 

   problem was due to duplicate reps added to the dim[i]->reps list.

8. added u_SetSelection(str) and u_GetSelection(), which
   now do set/get XA_STRING data for the XA_PRIMARY selection.

9. got rid of problem where each resolve_packages() set the rpdcount
   and disccount back to 0.

10. added remote still image service. most new code is in 'videostuff.c';
    means you have to make new sym links in each tree, and you have to
    add videostuff.[co] to the SRCS and OBJS lists in the Makefile. the
    feature is enabled by setting your GALATEA environment variable,
    indicating to Muse that the video source is remote. Muse tries to contact
    a Parallax X server at getenv("GALATEA") or using the "-ghost" option.
    if successful, Muse looks for a prop called "MUSE_PIXMAP_ID", put there
    by 'ldmusepm' as described below. all still image request will go the
    remote galatea; digitize into the reserved pixmap; dither the resulting
    XImage if necessary; and put it into Muse still's screen with XPutImage().
    if the destination is a mono workstation, you get a Floyd-Steinberg image.
    if it's a non-Parallax 8-bit display, you get a gray-scale image. if it's
    a Parallax, Muse tries to do the YUV trick. this may or may not work with
    the R4 server. Although this works great on SPARC and decmips machines,
    it may be intolerably slow for others. Let me know how bad it is, because
    i have a good idea on how to make the response time more constant between
    different performance-level machines.
   
11. wrote 'ldmusepm', to enable remote still image service as outlined above.
    works just like 'ldmusecm', except the resource is a pixmap instead of
    a colormap. to enable remote image service, you also need to do "xhost +"
    on the image server machine.

12. various other things which bugged me. a biggie is that the EvScript parser
    will no longer croak if you redefine a Cfunc. the new definition overwrites
    the old. all future invocations, no matter what env they were defined in,
    will use the new calling format. so, it's NOT a good idea to override a
    common definition (like printf) in this way, esp. if you env. could be used 
    with other environments that might expect a different definition.

------------

new for march 8

* dspvideo.c

	fixed a nit on remote video service; if you're on a Parallax,
	but you're getting remote video, you don't want to do any
	digitization locally; just put up the label.

	if you're paying attention, that means that remote video
	applications (like from Rotch through campus cable) MUST
	use a 'forw' line in their galatead.conf to access remote
	ANALOG video. to get a digital version, they either set the
	GALATEA environment variable or specify -gserv ... on the
	command line.

* resolve.c
	
	tried to fix up more nits to determine whether or not to
	do remote image service.
	
* make_Xtk.c

	fixed the UnSetTextSelection...() function. have to
	re-install the original translation table because i can't
	figure out how to override the sequence installed by the
	SetTextSelection...() function.

=======================

new for march 11

* create_gcs.c

	added a background attribute to all of the GC's. this was
	the ultimate cause of dithered B&W images not displaying
	properly on Suns.

* showpix.c

	fixed the same bug in this utility program.