While the list was sleeping, I adapted Helmut's allegro backend for abcl. It is not quite ready for prime time, but is somewhat usable. Actually, it's already at version 2 (or should I say 0.02): originally, I had to patch swank.lisp, but thanks to Peter Graves, who tracked down pretty printing bugs in abcl, this is no longer necessary. Nevertheless, it's probably too early to include it in the official slime sources, but I thought I'd let you know about it anyway. It's temporarily available here:
http://www.math.bme.hu/~asimon/slime-abcl.tgz
The tar file includes swank-loader.lisp (so back up yours if you untar it in the slime directory) and swank-abcl.lisp. It doesn't work with the latest abcl release, only with the cvs version. (In fact, it's probably better to wait a few hours until anon cvs at sourceforge picks up the latest sources.)
Andras
Andras Simon andras@renyi.hu writes:
The tar file includes swank-loader.lisp (so back up yours if you untar it in the slime directory) and swank-abcl.lisp. It doesn't work with the latest abcl release, only with the cvs version. (In fact, it's probably better to wait a few hours until anon cvs at sourceforge picks up the latest sources.)
I played a bit with swank-abcl and it's quite funny to see SLIME running on a JVM :-)
It seems that debugging via SLIME doesn't work yet. I also had the impression that the connection breaks when multi-line strings are transmitted.
Nice work!
Helmut.
On Wed, 9 Jun 2004, Helmut Eller wrote:
I played a bit with swank-abcl and it's quite funny to see SLIME running on a JVM :-)
Thanks for giving it a try!
It seems that debugging via SLIME doesn't work yet. I also had the
It should, in a minimal way: the *sldb* buffer pops up, I can scroll the backtrace and choose a restart or quit.
impression that the connection breaks when multi-line strings are transmitted.
As in, say
CL-USER> (print "<very long string here>")
or something more complicated? I'm asking because this seems to work here.
Andras
Andras Simon andras@renyi.hu writes:
It seems that debugging via SLIME doesn't work yet. I also had the
It should, in a minimal way: the *sldb* buffer pops up, I can scroll the backtrace and choose a restart or quit.
Doesn't work here. The debugger prints the condition and the restarts in the terminal. That's with ABCL 0.0.3.14+.
impression that the connection breaks when multi-line strings are transmitted.
As in, say
CL-USER> (print "<very long string here>")
or something more complicated? I'm asking because this seems to work here.
Try
CL-USER> (make-list 100)
Helmut.
On Thu, 10 Jun 2004, Helmut Eller wrote:
Andras Simon andras@renyi.hu writes:
It seems that debugging via SLIME doesn't work yet. I also had the
It should, in a minimal way: the *sldb* buffer pops up, I can scroll the backtrace and choose a restart or quit.
Doesn't work here. The debugger prints the condition and the restarts in the terminal. That's with ABCL 0.0.3.14+.
With today's slime version, it fails here, too, although in a different way. But that's not surprising, because swank-backend seems to have changed a lot in the past few days. I can't reproduce the problem with the 2004-06-07 version.
So this may be an abcl version problem. Like slime, abcl changes rapidly, and version numbers don't mean a lot (all that is guaranteed is that they increase monotonically :-) Or perhaps we're using different emacs versions (mine is 21.2.1).
impression that the connection breaks when multi-line strings are transmitted.
As in, say
CL-USER> (print "<very long string here>")
or something more complicated? I'm asking because this seems to work here.
Try
CL-USER> (make-list 100)
No problem here, even with the newest slime.
Andras
On Thu, 10 Jun 2004, Andras Simon wrote:
On Thu, 10 Jun 2004, Helmut Eller wrote:
Andras Simon andras@renyi.hu writes:
It seems that debugging via SLIME doesn't work yet. I also had the
It should, in a minimal way: the *sldb* buffer pops up, I can scroll the backtrace and choose a restart or quit.
Doesn't work here. The debugger prints the condition and the restarts in the terminal. That's with ABCL 0.0.3.14+.
With today's slime version, it fails here, too, although in a different way.
Now I know what causes _that_ failure; but a proper fix will probably have to wait until tomorrow.
Andras
-
Andras Simon -
Helmut Eller