Re: [Squeak-VMdev] Re: [Squeak-Commits] squeak/platforms/unix/misc/rfb - New directory

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Hi Cees,

> Did I miss an announcement somewhere?

No.  I haven't told anyone (outside of the people who were originally
involved) about this.  I did it a while back (when there was a certain
interest in running Squeak on machines a long way from our desktops) and
always intended to pull it into the regular distribution but simply never
had time.  (Apart from anything else it requires surgically removing the
entirety of the X11 code into a separate `VM component' and then
homogenising the interface to allow the graphics back-ends to be pluggable
and interchangeable.  You can guess just how much of a delicate operation
this is, and to be sure it was done completely correctly I'd have to
concentrate really hard on it for many hours.)

> Which endianity issues remain?

Something teeny and trivial in the encoding algorithm.  The symptom is
that groups of four pixels are reversed when running between two
little-endian machines.  (But the colours are fine, which is kind of
unsettling in a way.)

> I'm on a wrong-endian platform (i386), so if there are issues I'm
> interested in fixing them.

When I get serious about 3.3 I'll pull all this stuff into the regular
distribution and fix the problem then.  (My head will already be hurting
so much from making sure the X11 circumcision was done correctly to make
this additional task negligible.)  OTOH, you can almost certainly make a
working VM from the sources I pointed you at and see the problem for
yourself.  (The sources are very, very stale -- so don't even speculate
about trying to reuse anything outside of RFB from them.)  OTOH, if you
chekc out the symtpoms for yourself and identify the rogue line(s) in the
encoding routines then I would be very grateful for the info.

I think I also saw a crash once when connecting between two machines of
different byteorders, so there might be the odd [nh]to[hn][sl] missing in
the handshake and/or DES authentication.

Regards,
Ian