about summary refs log blame commit diff stats
path: root/archive/2.transect/compiler6
blob: 48a7030f654081d49a9ab00d7cffea9a02934456 (plain) (tree)



































                                                                                           
== Goal

A memory-safe language with a simple translator to x86 that can be feasibly written in x86.

== Definitions of terms

Memory-safe: it should be impossible to:
  a) create a pointer out of arbitrary data, or
  b) to access heap memory after it's been freed.

Simple: do all the work in a 2-pass translator:
  Pass 1: check each instruction's types in isolation.
  Pass 2: emit code for each instruction in isolation.

== types

int
char
(address _)
(array _ n)
(ref _)

addresses can't be saved to stack or global,
      or included in compound types
      or used across a call (to eliminate possibility of free)

<reg x> : (address T) <- advance <reg/mem> : (array T), <reg offset> : (index T)

arrays require a size
(ref array _) may not include a size

== open questions
Is argv an address?
Global variables are easiest to map to addresses.
Ideally we'd represent 'indirect' as a '*' and we could just count to make
sure that an instruction never has more than one '*'.
class="w"> break-if->= negate n } return n } fn sgn n: int -> _/eax: int { compare n, 0 { break-if-<= return 1 } { break-if->= return -1 } return 0 } fn shift-left-by n: int, bits: int -> _/eax: int { var i/eax: int <- copy bits { compare i, 0 break-if-<= shift-left n, 1 i <- decrement loop } return n } fn shift-right-by n: int, bits: int -> _/eax: int { var i/eax: int <- copy bits { compare i, 0 break-if-<= shift-right n, 1 i <- decrement loop } return n } fn clear-lowest-bits _n: (addr int), bits: int { var dest/edi: (addr int) <- copy _n var n/eax: int <- copy *dest n <- shift-right-by n, bits n <- shift-left-by n, bits copy-to *dest, n }