purify 4.1 and egcs still appear to be incompatible.

[Jason Merrill]

>>>>> Brad M Garcia <bgarcia@fore.com> writes:

> If I take all my classes that use this default copy constructor to copy
> a single bogus byte, and specifically give them a copy constructor (one
> which does nothing), I should stop getting these purify errors?

> And furthermore, when I compile with optimization, the resulting code
> should be _no_ different either way?

Not necessarily.  A class with a user-defined copy constructor, even one
that does nothing, has different semantics from one with a trivial,
implicitly generated copy constructor.  For one thing, gcc passes such
classes by invisible reference instead of in registers.

A better solution is to make g++ handle empty classes better.

Jason