john.russell@canremote.uucp (JOHN RUSSELL) (07/22/90)
None of the answers so fa{ has addressed the question of "space
efficient". Any solution that uses "if" stands a chance of generating
branch instructions. Any solution that explicitly returns 0 or 1 will
generate a "clear" instruction, and maybe a "move" instruction with a 16
or 32-bit operand.
The question might also have been checking for knowledge of a particular
chip instruction set, eg. on 68K a moveq.l #1 instruction allows you to
explicitly return a 1 without wasting 32 bits to hold the operand.
I won't post my solution (after all, Microsoft is presumably still
asking this question), but is compiles to 3 instructions plus function
linking stuff; one of the instructions has a 16-bit operand, the others
are register-to-register. (This is on a 68K without any optimization.)
John
---
* Via ProDoor 3.1R jak@sactoh0.UUCP (Jay A. Konigsberg) (07/25/90)
In article <15303b1aa0bf26a8c5d4@canremote.uucp> john.russell@canremote.uucp (JOHN RUSSELL) writes: >None of the answers so fa{ has addressed the question of "space >efficient". > That isn't the only question. Its also a matter of portability. >I won't post my solution (after all, Microsoft is presumably still >asking this question) > I think you are missing a basic point. There is no "best" solution. Rather, its a matter of understanding the pro's and con's of different approaches. -- ------------------------------------------------------------- Jay @ SAC-UNIX, Sacramento, Ca. UUCP=...pacbell!sactoh0!jak If something is worth doing, its worth doing correctly.