Re: Dumper issue, revisited; invalid realloc/free

[Eli Zaretskii]

> Date: Wed, 4 Feb 2015 15:02:15 -0500
> From: Rich Felker <address@hidden>
> Cc: address@hidden
> 
> glibc is maintained by a consensus-based community these days, and I
> don't think sacrificing the ability to detect serious memory
> corruption that likely indicates an exploit attempt for the sake of
> satisfying emacs' invalid assumptions about malloc would be popular.

It can be done without sacrificing anything.

> The same checks can be made on the emacs side before calling the
> underlying realloc/free

Yes, they can.  Both alternative require some knowledge of the other
side, so they are both equivalent to some degree.