Sat Apr 18 18:50:26 CEST 2009
I find there are a lot of subtleties due to the hash-table based
(mutable) name structure. To solve this properly will require quite
Where to start?
Let's first take the files apart, and figure out what is necessary at
compile time to perform pattern matching. Hmm.. too much junk
still.. Let's eliminate the run-time pattern first.
What is it used for? Easy to see by changing the accessor interface:
asm-prototype -> asm-proto
It is only used in the error handler. Let's take it out completely.
proto->asm-error-handler : assembler.ss
Ok.. Now, what information is necessary at run time about an
instruction assembler? For testing purposes it might be linked to its
disassembler, but that's about it.. All other info is best checked at
compile time (i.e. in pattern matching).