[slime-devel] Re: It's not just me! [was: CLISP on Windows?]

Dave Roberts ldave at droberts.com
Mon Jun 27 05:13:38 UTC 2005


So I was able to get this working with CLISP 2.33.1. The firewall was a
non-issue. The first time I fired it up, it basically worked. Windows
popped up a dialog informing me of the TCP connection and basically
asking what I wanted to do. I enabled it for Emacs/CLISP and things
worked well.

The problem, where ever it lies, is something that changed between CLISP
2.33.1 and 2.33.2. From what I can tell, something is failing and
preventing the server socket from getting created. I inserted debug
routines into all the various create-server sort of routines in SWANK
and they aren't getting called when Swank first starts. The only thing I
can surmise is that some CLISP routine called earlier is returning an
error that Swank doesn't like and it's just bailing out. The death is
quiet. No messages and no Lisp debugger loop. It basically just loads
everything and never gets around to starting the server.

To be honest, I have no idea what changed between 2.33.1 and 2.33.2. I
was just using 2.33.2, I think like you, because it was the latest
available tarball on the CLISP web site.

BTW, whatever the problem is here is specific to CLISP. The same issue
occurs whether using SLIME 1.0 or 1.2.1 tarballs.

-- Dave


On Sun, 2005-06-26 at 20:10 -0700, Peter Seibel wrote:
> So, as I mentioned in another message, I have the Windows firewall
> completely turned off. Dorking around with slime.el it seems that the
> problem is that in slime-start-swank-server the call to 
> 
>   (comint-send-string process 
>     (format "(swank:start-server %S :external-format %s)\n"
>     file encoding))
> 
> is having no effect. In fact, it doesn't seem to be happening at
> all. I put something in the elisp that loads slime to trace
> comint-send-string and I put a call to message right before the call
> to comint-send-string and I see the message but no sign of any
> activity from comint-send-string. However if, in the *inferior-lisp*
> buffer I call swank:start-server with the same arguments as it was
> supposed to be called with here, a) SLIME then finishes conecting and
> starts up and b) I see the trace of comint-send-string. So does anyone
> have any idea why this isn't working? FWIW, I have cygwin installed on
> this box but I'm not (as far as I know) using it in Lispbox--I have a
> .bat file that launches Lispbox and I'm using the plain vanilla binary
> distributions of Emacs and CLISP.
> 
> -Peter
> 
> 
-- 
Dave Roberts <ldave at droberts.com>




More information about the slime-devel mailing list