[38] in Kerberos-V5-bugs
Re: bcopy vs. memcpy
daemon@ATHENA.MIT.EDU (Mark W. Eichin)
Tue Oct 23 10:25:09 1990
Date: Tue, 23 Oct 90 10:16:40 -0400
From: Mark W. Eichin <eichin@ATHENA.MIT.EDU>
To: eichin@ATHENA.MIT.EDU
Cc: bstrand@poplar.cray.com, sommerfeld@apollo.com,
In-Reply-To: Mark W. Eichin <eichin@ATHENA.MIT.EDU> Thu, 18 Oct 90 18:12:45 -0400 <9010182212.AA13679@E40-008-11.MIT.EDU>
My apologies to anyone I've lead astray... I originally posted that
bcopy() does *NOT* do the right thing with overlap on the Sun 4 under
SunOS 4.0. (The code is scary enough just trying to handle alignment
efficiently - it's over 300 words long as it is...) The Sun 3
versiondoes handle overlap.
Brad Strand <bstrand@poplar.cray.com> checked bcopy in the particular
case I thought I remembered as failing, and determined both
experimentally and by examining sources and man pages for SunOS 3.0
through 4.1, that in fact bcopy *does* handle overlap while memcpy
does not. (Unfortunately I cannot check the code that I ran into this
situation with - it was at a previous employer who has some serious
attitudes about code release.) I'll have to assume that the code in
question was using memcpy instead.
_Mark_
ps. Kudos to Brad for breaking it to me politely :-)
