Kanban

Active

Next

local labels (visibility limited to single function)
structs: usage
types: methods
FIX - push.@[ goto.@] ]: in morty ([' and ]` treated as comments) -> drop support for separators ???

To do

peephole optimizer as a separate tool
separate stacks, code and mem ?
memory slices (addr -> slice,offset) ?
all addrs are indirect ?
debug stacks depth
check stacks overflows
trace
stand-alone executable
transpilation to C
cooperative multitasking
docs: all vm instructions (including turbo mode)
macros / inline ?
tail call optimization
fix globals_cnt
constants
iterators
else ?
heapsort.morty (requires else for easy porting)
array operations (similar to numpy / APL) ?
encode word in ASM
special stack shuffling syntax -> ie abcd|cadb or |cabd // ab|aba or |aba
vm text-based mem dump
strings
quote / unquote
ASM - token to line number (for error reporting)
case
little vs big endian
literate programming (compile from markdown)
multiprocessing
add Morty to linguist (https://github.com/github/linguist) so it can be detected by GitHub
https://rosettacode.org/wiki/Mandelbrot_set#Julia

Done

remove const from the language -> optimizer
JS - touch / mouse
fix !draw -> &draw
modules
VM interrupts handlers (run function, halt when it returns)
JS - time frames / draw function
C - shorter VMINFO (TOS IP SD RD HD dt)
remove macro support -> optimizer
FIX - vm.asm usage in morty
quote function (push function addr -> !name )
JS - IO ops - canvas
JS - better code loading
JS - canvas api adapter
JS - shorter VMINFO (TOS IP SD RD HD dt)
vm.js prototype
remove old VM ops in preparation for JS VM
fix: loops.asm (comments, wrong for order)
fix: sieve.asm / sieve.morty (for loop order)
replace -i option with -f (more standard)
forward-declarations or other technique for top-down programming
separate linker
{ ... } ( -- len addr ) push array length on the stack before addr
a[x 1 add] .get array notation (remove lambdas for now, introduce [ and ] as visual separators)
array notation ( .get .set )
more intuitive for ( lo hi incr )
qsort
ackerman.morty
structs: declaration
fix random.asm -O (muli, divi, modi vs OPFREQ)
vm: load cell code from stdin
remove HLA (lang.py prototype + asm is enough)
capture variables in lambdas (lambdas will use call frames)
rewrite benchmarks from HLA to ASM
lang: lambdas
fix: random.morty
fix: sieve.morty
fix: array.morty
rewrite benchmarks from HLA to Morty
compiler CLI (morty -> asm, -i -o)
global references -> globals pointer
static arrays
vm cleanup, separate core from (turbo) extension
primitive output
dynamic statistics (op freq) -> 25% performance penalty
vm text-based portable executable format
conditionals
replicated switch threading
sieve.hla
parametric cell size (16, 32, 64 bits)
loop changes (NO - new word for counting down on existing loops)
define and document file extensions
indefinitive loops
definitive loops
asm text-based compiled format
stack based label resolution
high-level assembler
vm cli
vm - load from binary file
primitive output
vm boot sequence
asm output to binary file
asm character literals
lang - structs
lang - macros