[slime-devel] Not using Slime's debugger

Teemu Likonen tlikonen at iki.fi
Mon Feb 6 16:18:10 UTC 2012


* Helmut Eller [2012-02-05 20:52:41 +0100] wrote:

> * Teemu Likonen [2012-02-05 18:38] writes:
>> I want to use my own debugger function. Normally I do this by setting
>> a custom function designator to cl:*debugger-hook*. With Slime I
>> can't do this automatically because swank sets it to
>> #'swank:swank-debugger-hook.
>>
>> The Slime manual says that ~/.swank.lisp can be used for setting
>> swank:*global-debugger* to nil. Even if I do this, cl:*debugger-hook*
>> is still set to swank's function.
>
> Works for me: the gobal value of *debugger-hook* is nil. SLIME still
> dynamically binds it while processing it's own requests.

Ah, separate binding, indeed. In *inferior-lisp* buffer *debugger-hook*
variable is NIL and in Slime-REPL buffer it's swank's debugger function.

>> Is there a way to set custom debugger automatically with Slime?
>> Mostly I use SBCL.
>
> Not that I know of.

OK, thanks. I ended up redefining swank:swank-debugger-hook function in
~/.swank.lisp file. It's a bit ugly but does what I want.

My "debugger" just prints the condition type, available restarts and
error message. Then it invokes ABORT restart automatically. I don't need
Slime's debugger because SBCL does not give much useful information
there anyway (even with the highest debugging level).





More information about the slime-devel mailing list