On Sun, Nov 07, 2004 at 02:31:48PM -0800, Chisheng Huang wrote:
Before 2004-11-07, I could fire up CMUCL with my own lisp core image, pcore, this way: M-x dired <RET> directory-containing-pcore <RET> C-u M-x slime <RET> <SPC> -core pcore <RET>
After I did a CVS update on 2004-11-07, doing the same thing gets me the following error message: Cannot find core file pcore Based on lisp binary path `/usr/local/cmucl/bin/lisp'
Process inferior-lisp exited abnormally with code 1
Doing a CVS update with -D "2004-11-06" got me back a version of Slime that won't bark at my "C-u M-x slime <RET> <SPC> -core pcore <RET>"
Is this change on 2004-11-07 a bug or feature?
I don't believe I changed anything like this in my commit today, and I see identical behavior between a 2004-11-06 and 2004-11-07 checkout. Here is what I think is happening:
The first time you run M-x slime, it makes the *inferior-lisp* buffer with the pwd of the directory of the current buffer. So if you were to try your experiment above and make sure you're in the right directory before *inferior-lisp* is created, it will work.
However, *inferior-lisp* isn't normally deleted when the Slime connection is closed. Afterwards, M-x slime will reuse the same *inferior-lisp* with the same old directory.
This can be confirmed by playing around with running: C-u M-x slime RET pwd RET
The obvious workaround to your problem is to make sure to kill the *inferior-lisp* buffer before starting Slime.
Hope that helps, -bcd