[Info-vax] Whither VMS?

[VAXman- at SendSpamHere.ORG]

In article <hao601$lqf$1 at naig.caltech.edu>, glen herrmannsfeldt <gah at ugcs.caltech.edu> writes:
>VAXman- wrote:
>(snip regarding null terminated vs. count field strings)
> 
>< With the counted example, long strings might be implemented with
>< long moves with the odd remainder being byte moves.  That would
>< be much more efficient for large data copies than the byte moves.
>
>OK, to be a little closer to the subject of this group, VAX
>has the MOVC instruction which can move up to 65535 characters
>with one instruction, assuming the length is known.
>
>I suppose LOCC could be used for null terminated strings, 
>to find the null character.  That would be in a loop in case
>the string is longer than 65535.  MOVTUC could probably do
>strcpy(), though with the memory references for the translate
>table (which translates everything to itself) it is probably
>slower than otherwise.  That would still need to be in a loop.
>For strncpy() it could be used with the length, and then MOVC
>with zero source length to pad the result out to the specified
>length.
>
>It seems to me that VAX is a little better for counted strings
>than null terminated strings.
>
>Any processor that doesn't set flags on character load/store
>will require extra tests for strcpy().


Consider the MOVCx a macro or routine.  Internally, it's not
so simple.  THere is a pseudo-code operation in the VAX arch-
itecture reference manual (EY-3459E-DP) pg. 147-150.  Three
ages!

-- 
VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)ORG

  http://www.quirkfactory.com/popart/asskey/eqn2.png
  
  "Well my son, life is like a beanstalk, isn't it?"