RE: [boost] Codeguarded regression tests

From: John Maddock (John_Maddock_at_[hidden])
Date: 2001-04-28 06:38:13

>However, after much investigation into the code, I don't believe that
>codeguard is correct.  It's possible that the inlining in the header can
>cause problems: array nothrow new is an inlined call to nothrow new which
is
>an inlined call to malloc, but delete is *not* an inlined call to free.
>
>So there is no actual heap corruption with pool.  But thanks for running
the
>test!

Yes you're right, if there is a bug, it's in Borland calling malloc
directly instead of via the regular new.

- John Maddock
http://ourworld.compuserve.com/homepages/john_maddock/