[cells-devel] Celtk and SBCL (and linux)

[Ken Tilton]

On 5/24/06, Stu Glaser <stuglaser at gmail.com> wrote:
>
> I'm happy to report that the one-button-window test appears to run
> correctly.  I've attached the appropriate patch.
>
> (The patch works from the current version in CVS, so it includes the
> previous changes)
>
> Additionally, menu-button-test, spinbox-test,
>
> lotsa-widgets almost works.  When I selected File|Quit, it quit my lisp
> process too.  Other than that, it appears okay.


It /is/  okay.  :) Just yesterday I was thinking, "I really should change
that before some poor user hits it." <g> That buttons invoke a Tcl "exit",
which is fine when talking to Tcl over a pipe. If it makes you feel better,
I killed myself that way yesterday when testing. :)


I can't find gears,


Gears got moved into its own module under Cells. I did not want to commit
Celtk to any one binding set to OpenGL, though Celtk does load the Togl
widget. Reason? I do not like the new cl-opengl bindings on c-l.net, but I
assume it will become the standard, so I will not force /my/ OpenGl bindings
on anyone. They happen also to be called cl-opengl.

and ltktest-cells-inside doesn't work.


I am fighting it out with that over here as well. Not sure if we are talking
about the same fight, of course. Oh, and I have not committed to CVS in
quite a few days and things are really getting tossed around, so... let's
talk again after the next CVS commit. I am sure we will need a second cycle
of patches and then things should stabilize.


One major change: SBCL has a special type for pointers.  This type does
> not work in hashtables, so the integer address must be stored there.


OK, I will look at that.

Hey, thanks for bringing Linux/SBCL on-line.

kenny
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cells-devel/attachments/20060524/85dd85b6/attachment.html>