Now now, causing a NULL follow crash is a feature. It negates the effect of the leak :)
On Wed, Apr 23, 2014 at 2:21 PM, Ted Unangst <t...@cvs.openbsd.org> wrote: > CVSROOT: /cvs > Module name: src > Changes by: t...@cvs.openbsd.org 2014/04/23 14:21:23 > > Modified files: > lib/libssl/src/crypto/bio: b_sock.c > > Log message: > if realloc failed, BIO_accept would leak memory and return NULL, causing > caller to crash. Fix leak and return an error instead. from Chad Loder
