Re: [Sbcl-commits] master: Reduce empty hash-table struct size for default rehash params
Stas Boukarev <[email protected]> Sun, 1 Mar 2026 19:22:38 +0300
| Newsgroups | gmane.lisp.steel-bank.devel |
|---|---|
| Message-ID | <CAF63=11w_bYChDRUcaSYkjG-6dWJTvZh3JCw6h-s0hhKoXtuvA@mail.gmail.com> |
If equalp is transformed at compile time then it returns T for weak hash tables too. On Sun, Mar 1, 2026 at 7:21 PM Douglas Katzman <[email protected]> wrote: > > Interestingly this bug already existed in two other ways: > 1) general-hash-table never called sb-int:hash-table-equalp, so it was already wrong. > 2) depending on whether SBCL purposely wanted to impose a new criterion regarding the :weakness aspect, a weak table can't be EQUALP to a non-weak one. We could define it to be the case since weakness isn't part of the spec. > I'll fix at least (1) along with the new problem, but is (2) an accident, or by design do you think? _______________________________________________ Sbcl-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/sbcl-devel