On Wed, 05 Nov 1997 17:10:40 +1000 TONY NIXON 54964
<Tony.nixon@eng.monash.edu.au> writes:
>>For example, clrf STATUS will clear all the page bits, not affect the
>>TO and PD bits, set the Z bit, *and not change C or DC*.  It says so
>>in the Microchip book and also does so on a real chip.  A movwf
>STATUS
>>*will* write to C,DC, and Z, because movwf doesn't inherently affect
>>any status bits.
>
>In my data book...
><snip>
>For example, CLRF STATUS will clear all bits except for TO and PD and
>then set the Z bit and leave STATUS register as 000UU100.

An old data book (1994 paper "Microchip Data Book") agrees with you.  A
newer one (16X84 pdf on the 1996 CD) has the section about the STATUS
register rewritten.  It says:

"If the STATUS register is the destination for an instruction which
modifies the Z, DC, or C bits, then the write to these three bits is
disabled."

"For example, CLRF STATUS will clear the upper three bits and set the Z
bit.  This leaves the STATUS register as 000uu1uu (where u=unchanged)"

My experience with a 1997-made 16F84 has been that CLRF STATUS doesn't
clear the C bit.  I don't know if the old silicon works differently or if
it is a mistake in the old book.

The description in the new book is that if the instruction can change one
or more of the bits, hardware "locks-out" writing to all three.  I would
assume this is true for rotate as well as clear, and, or, add, etc.