Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The PDP-6 and -10 hardware absolutely did have true bitfield pointers as a major feature of the instruction set. A single instruction ("ILDB") would do the equivalent of Rn=* cp++ (where Rn is a register, and cp is a char* ). Even better, the pointer itself specified how many bits were being addressed, and the ++ part of the operation would move forward that many bits. See "Byte instructions" at http://hakmem.org/pdp-10.html


I know how the byte instructions worked, they were copied to the Lisp Machine. You still don't get a single pointer to part of a machine word.


I don’t know why you would say that if you didn’t program a -6 or a -10. Byte pointers could e e incremental to move though memory just like word pointer, using the IBP instruction (increment byte pointer). If the byte spanned a word it would increment the word address.

Http://pdp10.nocrew.org/docs/instruction-set/Byte.html




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: