Just want to know if weak-p hash tables are required for cmucl? When using slime with cmucl on Mac OS X, there's a note about that because weak-p hash-tables only exist on platforms with gencgc support (x86, solaris).
Thanks,
Ray
* Raymond Toy [2005-09-25 23:52+0200] writes:
Just want to know if weak-p hash tables are required for cmucl? When using slime with cmucl on Mac OS X, there's a note about that because weak-p hash-tables only exist on platforms with gencgc support (x86, solaris).
I think it's not required. The weak hash tables are used to remember the results in the REPL and for presentations. Non-weak hash tables should work instead, but aren't as nice, because the tables are never cleared and the recorded objects could be pretty big.
You can enable/disable the stuff with swank:*record-repl-results*.
So, I think the default hash tables are good enough as long as swank-cmucl.lisp can be compiled without aborting.
Helmut.
Helmut Eller wrote:
- Raymond Toy [2005-09-25 23:52+0200] writes:
Just want to know if weak-p hash tables are required for cmucl? When using slime with cmucl on Mac OS X, there's a note about that because weak-p hash-tables only exist on platforms with gencgc support (x86, solaris).
I think it's not required. The weak hash tables are used to remember the results in the REPL and for presentations. Non-weak hash tables should work instead, but aren't as nice, because the tables are never cleared and the recorded objects could be pretty big.
You can enable/disable the stuff with swank:*record-repl-results*.
So, I think the default hash tables are good enough as long as swank-cmucl.lisp can be compiled without aborting.
Yes, it compiles just fine and runs just fine. I just get the warning about unsupported weak-p tables, and was wondering if it were really important or not.
Thanks!
Ray