Re: Unable to bootstrap gcc snapshot of 20001211 on Dynix/ptx

[Russ Allbery <rra at stanford dot edu>]

Robert Lipe <robertlipe@usa.net> writes:

> But isn't a system far more likely to have memmove as described in ISO C
> library ago than bcopy anyway?  Replacing the calls to bcopy still seems
> like a better path.

If there are really systems without memmove and with defective versions of
bcopy, it's not that hard to reimplement a simple, stupid, and unoptimized
version of memmove that should work on most systems.  Relies on comparing
two pointers to different objects to determine whether to copy forward or
backward, which isn't strict ANSI C, but it should work on all the systems
that don't have memmove in libc already (an ever-decreasing count).

I'd write a quick one and just tack it on to the end of this mail, if it
weren't for the various copyright hassles since I don't have paperwork.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>