Re: [STUMP] trouble requiring swank

[Eric Abrahamsen]

On Wed, May 16 2012, Trevor Murphy wrote:

> On Tue, May 15, 2012 at 12:00 PM,  <address@hidden> wrote:
>> Date: Tue, 15 May 2012 12:09:09 +0800
>> From: Eric Abrahamsen <address@hidden>
>> To: address@hidden
>> Subject: [STUMP] trouble requiring swank
>
>> I really have no idea what would be causing this -- I've only got one
>> sbcl on my system, and it's definitely 1.0.56, so I don't know what
>> would be trying to use sbcl 1.0.54. I'm able to require sb-introspect
>> from a regular repl?
>
> Hey, Eric.
>
> Fellow Archer here. One thing that's bitten me before is assuming that
> updating the sbcl package also updates the lisp processor used by
> stumpwm. It took me two similar swank issues to figure this out. I
> think the following explanation is right, but I'm no expert.
>
> When you compile stumpwm, you're using your current sbcl binary and
> dumping a copy / image of that compiled result for xinit to call
> later. This copy / image handles its own lisp processing from then on
> and never talks to the sbcl binary again (as far as I can tell).
>
> Thus if you compiled stumpwm under sbcl 1.0.54, then updated sbcl and
> swank, you might get this error. The solution in that case is to
> recompile stumpwm against the sbcl you have now. I just do yaourt -S
> stumpwm-git every time I see that sbcl has updated.

Recompiling stumpwm did it! Thanks very much for the hint. I knew that
lisp programs often result in the creation of independent binaries, but
it had never crossed my mind that that's what was going on here.

Off to hack!

Thanks,
Eric