[comp.windows.x] MITs Xapollo

hanche@imf.unit.no (Harald Hanche-Olsen) (03/07/91)

We have been trying to build and operate the Xapollo (R4) server from
the MIT source, with rather mixed success.  When run in gpr mode, it
operates fairly well, except for certain startup problems that I will
make the subject of a separate posting.  However, on monochrome
displays we want to run it in mfb mode, as our experiments reveal this
to be much faster.  There is one small problem...

Some clients display both foreground and background as black, whereas
others appear to work normally.  A little poking reveals that those
which try to specify colors by name or rgb values turn out all black,
while those who rely on WhitePixel and BlackPixel work OK.  For
example, xterm and xclock are all black if foreground and background
are specified, but normal otherwise.

So we wrote a minimal client which calls XQueryColor(d,
DefaultColormap(d,s), ...) and prints out the resulting rgb values for
pixel values 0 and 1.  When run against the Xapollo server in gpr
mode, the output is

pixel 0: rgb = 0/0/0
pixel 1: rgb = 65535/65535/65535

but if run against the server in mfb mode, it is

pixel 0: rgb = 0/0/0
pixel 1: rgb = 0/0/0

which seems crazy and ought to explain the odd behaviour of the
server.  Only question then is, why is this happening?  I tried
reading the server code, but found I could not make much sense out of
it.  If this is a known bug or similar problems have been seen by
others, could someone please tell me?

(The boring details: Machine is Apollo DN3000 with 19'' monochrome
screen, operating system is SR10.3 (bsd4.3), compiler used is cc 6.8,
source used is standard X11R4 distribution with fixes 1-18 applied).

- Harald Hanche-Olsen <hanche@imf.unit.no>
  Division of Mathematical Sciences
  The Norwegian Institute of Technology
  N-7034 Trondheim, NORWAY