HEADSUP: malloc changeover

Giorgos Keramidas keramida at ceid.upatras.gr
Fri Jan 13 12:45:57 PST 2006

On 2006-01-13 15:18, Kris Kennaway <kris at obsecurity.org> wrote:
>> I have built & run at least four versions since January 11, two with
>> DEBUG_FLAGS='-g' for both userland and kernel and two without.  As
>> Jason said, we should probably keep looking for potential problems
>> both in the new malloc() and the programs that use it[...]
> There are several ports that have build failures due to application
> bugs.  I think I emailed some of the maintainers already, but there
> are more that are going to be (re-)exposed by some more recent changes
> than that patchset I was testing.

I know :/

I've spent the relatively limited spare time I had the past 2 days
trying to track down a segfault & core dump in the build of a CVS
snapshot of Emacs.  Strangely enough, building with an older version of
malloc and then running the same binary with the new version works fine,
so this looks like a bug in the multi-stage build of Emacs instead of a
really serious malloc bug.  I've posted a request for help on
emacs-devel and continue trying to nail this down to something more
specific than "it crashes when s->data[nbytes]" is accessed.

More information about the freebsd-current mailing list