You can clone with
HTTPS or Subversion.
structural changes (fixed segfault | args now have to be passed one a…
…t a time)
surprise, surprise - a SIGSEGV \0/
Now uses List reverse(), and ditched Cell for the sdk's one
removed old files
structural change pt.2 (made yajit more modular/see tests/)
change in structure (to be continued)
fixed stack-order bug concerning multiple closure-args
if the order of the passed arguments doesn't matter or mem-addresses …
…instead of int=42 are not a problem - then yajit works for you right know =)
seperated genCode into several sub-functions
fixed mem-leak|i'm hunting the yajit-hates-everything-apart-DWords-bug
Merge branch 'master' of github.com:nddrylliog/yajit
added makefile + i run out of memory \o/
segfault - and i don't know why \o/
Fixed argSizes iteration. Should avoid segfault.
Fill `transTable` in BinarySeq init~withSize.
cleanup + added vals to transTable
added support to specify arg-size with a string like 'ihc' (Int, Shor…
Modified to use gc_malloc + mprotect instead of mmap
Works for me now, too! Using mmap instead of malloc =D. Note that thi…
…s is not really good, since it'll be leaking, afaict.
added a test for closure return values.
converted tabs to 4 spaces.
replaced hardcoded type sizes in yajit core.
made yajit WORK (yeah, WORK) under latest ooc.
I love you showstopper!
Now compiles under the latest ooc - but crashes, haha ^^
added abstraction - doesn't wanna compile
Let the generated function accept further arguments
Added opcdes in order to pass args
Fixed opcode for PUSH_WORD (sorry showstoppy)
Used generics all over the place - great fun =)
added new opcode
Not working code as a comment
Fixed bug in ooc allows arg to be used as-is