[Openmcl-devel] Is it safe to make hash tables with non-standard tests?

[Ron Garret]

On Jul 14, 2013, at 6:45 PM, Tom Emerson wrote:

> WRT symbols, fair enough. I'd assumed #'string= for strings, not necessarily things that can become strings.

Those are called string designators.

> I guess wrapping in symbol-name is a portable way to do it.

Yes, I considered that, but there are reasons I don't want to do it.  If you really want to know tell me and I'll describe my design constraints.  It's a long story.

> But as Gary says, if you're not worried about portability then #'string= will be what you want.

I don't care about portability.  I do care about efficiency, security. and memory leaks.  So, for example, I don't want to intern my symbols because I fear they won't be garbage collected.  (Empirically, CCL seems to be pretty good about collecting useless symbols even if they are interned, but it still makes me queasy.)  I have to handle data from untrusted sources, and EQUAL is subject to denial-of-service attacks if someone manages to, say, inject a circular list as a key.

Stuff like that.

rg