Re: freeze for sbcl-2.6.2

And it also cannot be just "If you are using weak hash tables",
because sbcl itself is using them. The original test case doesn't
include any hash-table operations.

On Thu, Feb 26, 2026 at 6:07 PM Stas Boukarev <[email protected]> wrote:
>
> And 2.6.1 has a fix for clrhash on weak hash-tables.
> And as for a scheduled fix, I don't have a fix, only a reversion.
>
> On Thu, Feb 26, 2026 at 6:02 PM Christophe Rhodes <[email protected]> wrote:
> >
> > Less NEWS and more announcement mail, but:
> >
> >   If you are using weak hash tables successfully with a recent SBCL, be
> >   aware that there's an upcoming fix for a problem targeting the next
> >   release.  If you're using them in releases before 2.5.11, consider not
> >   upgrading until that fix has landed, scheduled for early March.
> >
> > Christophe
> >
> > Stas Boukarev <[email protected]> writes:
> >
> > > I'm just not even sure how it can be mentioned in the NEWS:
> > > "hey, if you use weak hash-tables things might crash randomly, or the
> > > GC might enter an infinite loop. So use sbcl 2.5.11 if you care about
> > > these things. But there are other fixes for memory corruptions
> > > included since 2.5.11. But don't despair, these are just the things we
> > > know about"
> > >
> > > On Thu, Feb 26, 2026 at 4:45 PM Stas Boukarev <[email protected]> wrote:
> > >>
> > >> Maybe it can be reverted?
> > >>
> > >> On Thu, Feb 26, 2026 at 11:40 Christophe Rhodes <[email protected]> wrote:
> > >>>
> > >>> I'll call it out in my release message but -- since it's not a
> > >>> regression this month -- I wouldn't want to hold up a release for that
> > >>> (or merge something last-minute-ish) because that looks like the kind of
> > >>> area where it would be easy to come up with something plausible but not
> > >>> complete.  Thank you for finding and reporting it though.
> > >>>
> > >>> Christophe
> > >>>
> > >>> Stas Boukarev <[email protected]> writes:
> > >>>
> > >>> > Would be nice to have this recent, but not current, regression be
> > >>> > solved: https://bugs.launchpad.net/sbcl/+bug/2142714
> > >>> >
> > >>> > On Sat, Feb 21, 2026 at 12:31 AM Christophe Rhodes via Sbcl-devel
> > >>> > <[email protected]> wrote:
> > >>> >>
> > >>> >> Hi all,
> > >>> >>
> > >>> >> It's been a busy month!  184 commits in 25 days: time to pause for a bit
> > >>> >> and take stock.  Please could everyone spend your energies on testing,
> > >>> >> verifying, and dealing with any regressions or critical bugs, for a
> > >>> >> release late next week?
> > >>> >>
> > >>> >> Thank you,
> > >>> >>
> > >>> >> Christophe
> > >>> >>
> > >>> >>
> > >>> >> _______________________________________________
> > >>> >> Sbcl-devel mailing list
> > >>> >> [email protected]
> > >>> >> https://lists.sourceforge.net/lists/listinfo/sbcl-devel


_______________________________________________
Sbcl-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sbcl-devel