[slime-devel] multithreaded SBCL 0.9.2.x and slime

[Gábor Melis]

On Monday 18 July 2005 10:03, Friedrich Dominicus wrote:
> I probalbly posted this before. Does anyone have encountered problems
> with that combination I'm using SBCL 0.9.2.51 and the CVS-Slime.
> the message I got is:
>    implicitly creating new generic function STREAM-READ-CHAR-WILL-HANG-P
> ;; loading
> #P"/home/frido/.slime/fasl/sbcl-0.9.2.51-linux-x86-64/swank.fasl"
> WARNING: These Swank interfaces are unimplemented:
>  (ACTIVATE-STEPPING CALLS-WHO
>                     DISASSEMBLE-FRAME
>                     INSPECT-FOR-EMACS
>                     RESTART-FRAME
>                     SLDB-BREAK-AT-START
>                     SLDB-BREAK-ON-RETURN
>                     WHO-BINDS
>                     WHO-CALLS
>                     WHO-MACROEXPANDS
>                     WHO-REFERENCES
>                     WHO-SETS
>                     WHO-SPECIALIZES)
> T
> * ;; Swank started at port: 33038.
>
> 33038
> *
> Process inferior-lisp hangup
>
>
> Is there anything I can do to get a multithreaded SBCL running with
> Slime?
>
> Regards
> Friedrich
> _______________________________________________
> slime-devel site list
> slime-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/slime-devel


I'm sorry for replying so late. This problem was discussed on sbcl-devel some 
time ago. AFAICT it is present since the sbcl pthread merge (0.9.1.59) under 
some kernels.

The immediate cause is emacs sending a SIGHUP to sbcl that does not have a 
handler and promptly exits which is the right behaviour. My tests showed that 
even if :sigio or :fd-handler is used, the first thread exit makes emacs
send sighup.

A possible workaround is:

(setq inferior-lisp-program "nohup sbcl")

or writing a little wrapper script around sbcl if you want *inferior-lisp* 
keep its sanity:

#!/bin/sh
sbcl

But probably it is a kernel problem after all: for me the problem went
away after a 2.6.8 -> 2.6.11 upgrade.

Cheers, Gábor