mu/archive/2.transect/compiler7

47 lines
1.4 KiB
Plaintext

== 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 _ stack|heap|global)
(array _ n) # on stack or global
(ref _)
(ref array _) # by definition always on the heap
addresses to global can be saved and manipulated as usual
addresses on stack can't be saved to heap
addresses on heap can't be saved to global (use ref)
addresses to stack or heap can't be included in compound types
or used across a call
or used across a label
<reg x> : (address T stack|global) <- advance <reg/mem> : (array T), <reg offset> : (index T)
<reg x> : (address T heap) <- advance *<mem> : (ref array T), <reg offset> : (index T)
arrays require a size
(ref array _) may not include a size
Arguments of type 'address' are required to be on the stack or global. Can't
be on the heap.
So we need duplication for address and ref arguments?
Argv has type (array (address (array char) global))
variables on stack, heap and global are references. The name points at the
address. Use '*' to get at the value.