[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Announcing a re-implementation of CLX for CLISP, coded in C
- To: email@example.com
- Subject: Announcing a re-implementation of CLX for CLISP, coded in C
- From: firstname.lastname@example.org
- Date: Sat, 12 Oct 1996 12:34:00 +0200
As the next semester begins, I decided to release this code. I am hacking on
it since December last year.
[This is just a trimmed version of the README file:]
What is this?
Hadn't we CLX for CLISP for centuries? Yes, but this is a re-implementation
of CLX written in C for speed.
To use this package you will have to recompile CLISP; more precise:
build a new link set. If you could make use of an i386/Linux/ELF binary, you
may want to grasp the binary distribution.
Benefits vs. CLX coded in Lisp
- it is faster
This mainly for two reasons. The CLISP compiler is not very good, when it
comes to do all these bit operations and pushing bytes back and forth as
in the CLX code. Second reason is, that i.e drawing a line do not produce
garbage using this CLX implementation in contrast to MIT-CLX.
- it is virtually smaller
Currently I was able to blow up code size to somewhat near 120k, but it is
likely that it will shrink in the future. MIT-CLX needs approx 700k.
- I was able to run Garnet successfully, all the demos and gilt and lapidary.
- WINDOW, PIXMAP, etc... are now CLOS classes. This is the right way to go,
and I wanted to be able to compile CLUE. But CLUE does not run, because it
uses too many internals of MIT-CLX
- I plan to offer access to the SHAPE and to the multi buffer extension. The
most basic function for the SHAPE extension is already there. [look at the
It will be available, as soon as Marcus Daniels or Bruno Haible will put it
into the ftp directory at ma2s2.mathematik.uni-karlsruhe.de in
/pub/lisp/clisp/packages; These files should be there:
Just the README file. Also stating what is incomplete.
Binary distribution based on clisp-1996-07-22, as the file name says.
Both the binary and the source distribution contain two little demos
for a quick check if the code runs.
So if you do X hacking with CLISP, consider using this package. The more
people use it, the more bug reports I get, the more stable and portable will
this code get. So please take a look at this code, I need beta testers. It's
actually more gamma than beta, I do regular Garnet hacking myself using this
package. When doing heavy graphics, it even out performs CMUCL!
So go and bury me with bug reports and further suggestions,