AAAaaargh! I didn't read properly. Please ignore my rant.
I rather believe that the change should be done in OPENSSL_realloc(), so future uses elsewhere do not get into the same trouble. [levitte - Thu Aug 1 11:23:34 2002]: > [[EMAIL PROTECTED] - Thu Aug 1 09:20:27 2002]: > > > On Wed, 31 Jul 2002, Jon Peatfield wrote: > > > > > Looking through the rest of the 0.9.6e patch I can't see any > other > > > cases where realloc() is called like this but I might well have > missed > > > some. I'm hoping that someone who understands the code better > will > > > confirm/check this. > > > > Might it not be better to add a check to OPENSSL_realloc (inside > an #ifdef > > if necessary/prudent for speed reasons) to check for this > condition on > > SunOS4 and any other similar platform? That way, you could be > sure this > > (pretty unusual) case would be caught everywhere... > > I'm not sure I udnerstand what you say. It seems like you say that > the 'if (ret->data == NULL) ...' check be wrapped in a conditional > so it will only be performed on some operating systems. If that's > what you say, I'm totally against. Checking for NULL is cheap, and > it's much easier to do that for all than try to guess what operating > systems have realloc() do the same check (trust me, we already have > enough checks like that elsewhere, we really don't need yet another > one). > > I intend to apply the given patch. It's easy to understand and to > the point. -- Richard Levitte [EMAIL PROTECTED] ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]