mirror of
https://github.com/notwa/mm
synced 2024-11-05 02:39:02 -08:00
update lips
This commit is contained in:
parent
449c021110
commit
a6f953c0c0
10 changed files with 416 additions and 355 deletions
|
@ -1,29 +1,24 @@
|
|||
local insert = table.insert
|
||||
|
||||
local path = string.gsub(..., "[^.]+$", "")
|
||||
local Base = require(path.."Base")
|
||||
local Token = require(path.."Token")
|
||||
local TokenIter = require(path.."TokenIter")
|
||||
local Statement = require(path.."Statement")
|
||||
local Muncher = require(path.."Muncher")
|
||||
|
||||
local arg_types = { -- for instructions
|
||||
NUM = true,
|
||||
REG = true,
|
||||
VARSYM = true,
|
||||
LABELSYM = true,
|
||||
RELLABELSYM = true,
|
||||
}
|
||||
|
||||
local Collector = Muncher:extend()
|
||||
local Collector = Base:extend()
|
||||
function Collector:init(options)
|
||||
self.options = options or {}
|
||||
end
|
||||
|
||||
function Collector:statement(...)
|
||||
local s = Statement(self.fn, self.line, ...)
|
||||
local I = self.iter
|
||||
local s = Statement(I.fn, I.line, ...)
|
||||
return s
|
||||
end
|
||||
|
||||
function Collector:push_data(datum, size)
|
||||
local I = self.iter
|
||||
--[[ pseudo-example:
|
||||
Statement{type='!DATA',
|
||||
{tt='BYTES', tok={0, 1, 2}},
|
||||
|
@ -33,10 +28,12 @@ function Collector:push_data(datum, size)
|
|||
}
|
||||
--]]
|
||||
|
||||
-- FIXME: optimize the hell out of this garbage, preferably in the lexer
|
||||
-- TODO: consider not scrunching data statements, just their tokens
|
||||
-- TODO: concatenate strings; use !BIN instead of !DATA
|
||||
|
||||
if type(datum) == 'number' then
|
||||
datum = self:token(datum)
|
||||
datum = I:token(datum)
|
||||
end
|
||||
|
||||
local last_statement = self.statements[#self.statements]
|
||||
|
@ -58,13 +55,13 @@ function Collector:push_data(datum, size)
|
|||
insert(s, datum)
|
||||
return
|
||||
else
|
||||
self:error('labels are too large to be used in this directive')
|
||||
I:error('labels are too large to be used in this directive')
|
||||
end
|
||||
elseif datum.tt == 'VARSYM' then
|
||||
insert(s, datum:set('size', size))
|
||||
return
|
||||
elseif datum.tt ~= 'NUM' then
|
||||
self:error('unsupported data type', datum.tt)
|
||||
I:error('unsupported data type', datum.tt)
|
||||
end
|
||||
|
||||
local sizes = size..'S'
|
||||
|
@ -74,74 +71,67 @@ function Collector:push_data(datum, size)
|
|||
if last_token and last_token.tt == sizes then
|
||||
t = last_token
|
||||
else
|
||||
t = self:token(sizes, {})
|
||||
t = I:token(sizes, {})
|
||||
insert(s, t)
|
||||
s:validate()
|
||||
end
|
||||
insert(t.tok, datum.tok)
|
||||
end
|
||||
|
||||
function Collector:variable()
|
||||
local t = self.t
|
||||
local t2 = self:advance()
|
||||
|
||||
local s = self:statement('!VAR', t, t2)
|
||||
insert(self.statements, s)
|
||||
self:advance()
|
||||
end
|
||||
|
||||
function Collector:directive()
|
||||
local name = self.tok
|
||||
self:advance()
|
||||
function Collector:directive(name)
|
||||
local I = self.iter
|
||||
local function add(kind, ...)
|
||||
insert(self.statements, self:statement('!'..kind, ...))
|
||||
end
|
||||
|
||||
if name == 'ORG' or name == 'BASE' then
|
||||
add(name, self:const(nil, 'no labels'))
|
||||
add(name, I:const(nil, 'no labels'))
|
||||
elseif name == 'PUSH' or name == 'POP' then
|
||||
add(name, self:const())
|
||||
while not self:is_EOL() do
|
||||
self:optional_comma()
|
||||
add(name, self:const())
|
||||
add(name, I:const())
|
||||
while not I:is_EOL() do
|
||||
I:eat_comma()
|
||||
add(name, I:const())
|
||||
end
|
||||
elseif name == 'ALIGN' or name == 'SKIP' then
|
||||
if self:is_EOL() and name == 'ALIGN' then
|
||||
if I:is_EOL() and name == 'ALIGN' then
|
||||
add(name)
|
||||
else
|
||||
local size = self:const(nil, 'no label')
|
||||
if self:is_EOL() then
|
||||
local size = I:const(nil, 'no label')
|
||||
if I:is_EOL() then
|
||||
add(name, size)
|
||||
else
|
||||
self:optional_comma()
|
||||
add(name, size, self:const(nil, 'no label'))
|
||||
I:eat_comma()
|
||||
add(name, size, I:const(nil, 'no label'))
|
||||
end
|
||||
end
|
||||
elseif name == 'BIN' then
|
||||
-- FIXME: not a real directive, just a workaround
|
||||
add(name, I:string())
|
||||
elseif name == 'BYTE' or name == 'HALFWORD' or name == 'WORD' then
|
||||
self:push_data(self:const(), name)
|
||||
while not self:is_EOL() do
|
||||
self:optional_comma()
|
||||
self:push_data(self:const(), name)
|
||||
self:push_data(I:const(), name)
|
||||
while not I:is_EOL() do
|
||||
I:eat_comma()
|
||||
self:push_data(I:const(), name)
|
||||
end
|
||||
elseif name == 'HEX' then
|
||||
if self.tt ~= 'OPEN' then
|
||||
self:error('expected opening brace for hex directive', self.tt)
|
||||
if I.tt ~= 'OPEN' then
|
||||
I:error('expected opening brace for hex directive', I.tt)
|
||||
end
|
||||
self:advance()
|
||||
I:next()
|
||||
|
||||
while self.tt ~= 'CLOSE' do
|
||||
if self.tt == 'EOL' then
|
||||
self:advance()
|
||||
while I.tt ~= 'CLOSE' do
|
||||
if I.tt == 'EOL' then
|
||||
I:next()
|
||||
else
|
||||
self:push_data(self:const(), 'BYTE')
|
||||
self:push_data(I:const(), 'BYTE')
|
||||
end
|
||||
end
|
||||
self:advance()
|
||||
I:next()
|
||||
elseif name == 'INC' or name == 'INCBIN' then
|
||||
-- noop, handled by lexer
|
||||
self:string()
|
||||
return -- don't expect EOL
|
||||
I:string()
|
||||
elseif name == 'ASCII' or name == 'ASCIIZ' then
|
||||
local bytes = self:string()
|
||||
local bytes = I:string()
|
||||
for i, number in ipairs(bytes.tok) do
|
||||
self:push_data(number, 'BYTE')
|
||||
end
|
||||
|
@ -149,85 +139,61 @@ function Collector:directive()
|
|||
self:push_data(0, 'BYTE')
|
||||
end
|
||||
elseif name == 'FLOAT' then
|
||||
self:error('unimplemented directive', name)
|
||||
I:error('unimplemented directive', name)
|
||||
else
|
||||
self:error('unknown directive', name)
|
||||
I:error('unknown directive', name)
|
||||
end
|
||||
self:expect_EOL()
|
||||
|
||||
I:expect_EOL()
|
||||
end
|
||||
|
||||
function Collector:basic_special()
|
||||
local name, args = self:special()
|
||||
|
||||
local portion
|
||||
if name == 'hi' then
|
||||
portion = 'upperoff'
|
||||
elseif name == 'up' then
|
||||
portion = 'upper'
|
||||
elseif name == 'lo' then
|
||||
portion = 'lower'
|
||||
else
|
||||
self:error('unknown special', name)
|
||||
end
|
||||
|
||||
if #args ~= 1 then
|
||||
self:error(name..' expected one argument', #args)
|
||||
end
|
||||
|
||||
local t = self:token(args[1]):set('portion', portion)
|
||||
return t
|
||||
end
|
||||
|
||||
function Collector:instruction()
|
||||
local s = self:statement(self.tok)
|
||||
function Collector:instruction(name)
|
||||
local I = self.iter
|
||||
local s = self:statement(name)
|
||||
insert(self.statements, s)
|
||||
self:advance()
|
||||
|
||||
while self.tt ~= 'EOL' do
|
||||
local t = self.t
|
||||
if self.tt == 'OPEN' then
|
||||
t = self:deref()
|
||||
t.tt = 'DEREF' -- TODO: should just be returned by :deref
|
||||
insert(s, t)
|
||||
elseif self.tt == 'UNARY' then
|
||||
local peek = self.tokens[self.i + 1]
|
||||
while I.tt ~= 'EOL' do
|
||||
local t = I.t
|
||||
if I.tt == 'OPEN' then
|
||||
insert(s, I:deref())
|
||||
elseif I.tt == 'UNARY' then
|
||||
local peek = assert(I:peek())
|
||||
if peek.tt == 'VARSYM' then
|
||||
local negate = t.tok == -1
|
||||
t = self:advance()
|
||||
t = I:next()
|
||||
t = Token(t):set('negate', negate)
|
||||
insert(s, t)
|
||||
self:advance()
|
||||
I:next()
|
||||
elseif peek.tt == 'EOL' or peek.tt == 'SEP' then
|
||||
local tok = t.tok == 1 and '+' or t.tok == -1 and '-'
|
||||
t = Token(self.fn, self.line, 'RELLABELSYM', tok)
|
||||
t = Token(I.fn, I.line, 'RELLABELSYM', tok)
|
||||
insert(s, t)
|
||||
self:advance()
|
||||
I:next()
|
||||
else
|
||||
self:error('unexpected token after unary operator', peek.tt)
|
||||
I:error('unexpected token after unary operator', peek.tt)
|
||||
end
|
||||
elseif self.tt == 'SPECIAL' then
|
||||
t = self:basic_special()
|
||||
elseif I.tt == 'SPECIAL' then
|
||||
t = I:basic_special()
|
||||
insert(s, t)
|
||||
self:advance()
|
||||
elseif self.tt == 'SEP' then
|
||||
self:error('extraneous comma')
|
||||
elseif not arg_types[self.tt] then
|
||||
self:error('unexpected argument type in instruction', self.tt)
|
||||
I:next()
|
||||
elseif I.tt == 'SEP' then
|
||||
I:error('extraneous comma')
|
||||
elseif not I.arg_types[I.tt] then
|
||||
I:error('unexpected argument type in instruction', I.tt)
|
||||
else
|
||||
insert(s, t)
|
||||
self:advance()
|
||||
I:next()
|
||||
end
|
||||
self:optional_comma()
|
||||
I:eat_comma()
|
||||
end
|
||||
|
||||
self:expect_EOL()
|
||||
I:expect_EOL()
|
||||
s:validate()
|
||||
end
|
||||
|
||||
function Collector:collect(tokens, fn)
|
||||
self.tokens = tokens
|
||||
self.fn = fn or '(string)'
|
||||
self.main_fn = self.fn
|
||||
self.iter = TokenIter(tokens)
|
||||
local I = self.iter
|
||||
|
||||
self.statements = {}
|
||||
|
||||
|
@ -241,29 +207,27 @@ function Collector:collect(tokens, fn)
|
|||
insert(self.statements, s)
|
||||
end
|
||||
|
||||
self.i = 0 -- set up Muncher iteration
|
||||
self:advance() -- load up the first token
|
||||
while true do
|
||||
if self.tt == 'EOF' then
|
||||
-- don't break if this is an included file's EOF
|
||||
if self.fn == self.main_fn then
|
||||
break
|
||||
end
|
||||
self:advance()
|
||||
elseif self.tt == 'EOL' then
|
||||
-- empty line
|
||||
self:advance()
|
||||
elseif self.tt == 'VAR' then
|
||||
self:variable() -- handles advancing
|
||||
elseif self.tt == 'LABEL' or self.tt == 'RELLABEL' then
|
||||
insert(self.statements, self:statement('!LABEL', self.t))
|
||||
self:advance()
|
||||
elseif self.tt == 'DIR' then
|
||||
self:directive() -- handles advancing
|
||||
elseif self.tt == 'INSTR' then
|
||||
self:instruction() -- handles advancing
|
||||
for t in I do
|
||||
if t.tt == 'EOF' then
|
||||
-- noop
|
||||
elseif t.tt == 'EOL' then
|
||||
-- noop; empty line
|
||||
elseif t.tt == 'LABEL' or t.tt == 'RELLABEL' then
|
||||
insert(self.statements, self:statement('!LABEL', t))
|
||||
elseif t.tt == 'VAR' then
|
||||
local t2 = I:next()
|
||||
I:next()
|
||||
local s = self:statement('!VAR', t, t2)
|
||||
insert(self.statements, s)
|
||||
I:expect_EOL()
|
||||
elseif t.tt == 'DIR' then
|
||||
I:next()
|
||||
self:directive(t.tok)
|
||||
elseif t.tt == 'INSTR' then
|
||||
I:next()
|
||||
self:instruction(t.tok)
|
||||
else
|
||||
self:error('expected starting token for statement', self.tt)
|
||||
I:error('expected starting token for statement', t.tt)
|
||||
end
|
||||
end
|
||||
|
||||
|
|
|
@ -1,3 +1,4 @@
|
|||
local byte = string.byte
|
||||
local floor = math.floor
|
||||
local format = string.format
|
||||
local insert = table.insert
|
||||
|
@ -253,6 +254,10 @@ function Dumper:load(statements)
|
|||
s.length = util.measure_data(s) -- cache for next pass
|
||||
self.pos = self.pos + s.length
|
||||
insert(new_statements, s)
|
||||
elseif s.type == '!BIN' then
|
||||
s.length = #s[1].tok
|
||||
self.pos = self.pos + s.length
|
||||
insert(new_statements, s)
|
||||
elseif s.type == '!ORG' then
|
||||
self.pos = s[1].tok
|
||||
insert(new_statements, s)
|
||||
|
@ -378,6 +383,9 @@ function Dumper:load(statements)
|
|||
end
|
||||
self.pos = self.pos + (s.length or util.measure_data(s))
|
||||
insert(new_statements, s)
|
||||
elseif s.type == '!BIN' then
|
||||
self.pos = self.pos + s.length
|
||||
insert(new_statements, s)
|
||||
elseif s.type == '!ORG' then
|
||||
self.pos = s[1].tok
|
||||
insert(new_statements, s)
|
||||
|
@ -423,6 +431,12 @@ function Dumper:dump()
|
|||
error('Internal Error: unknown !DATA token')
|
||||
end
|
||||
end
|
||||
elseif s.type == '!BIN' then
|
||||
local data = s[1].tok
|
||||
for i=1, #data do
|
||||
self.writer(self.pos, byte(data, i))
|
||||
self.pos = self.pos + 1
|
||||
end
|
||||
elseif s.type == '!ORG' then
|
||||
self.pos = s[1].tok
|
||||
else
|
||||
|
|
96
Lua/lib/lips/Expander.lua
Normal file
96
Lua/lib/lips/Expander.lua
Normal file
|
@ -0,0 +1,96 @@
|
|||
local insert = table.insert
|
||||
|
||||
local path = string.gsub(..., "[^.]+$", "")
|
||||
local data = require(path.."data")
|
||||
local overrides = require(path.."overrides")
|
||||
local Statement = require(path.."Statement")
|
||||
local Reader = require(path.."Reader")
|
||||
|
||||
local Expander = Reader:extend()
|
||||
function Expander:init(options)
|
||||
self.options = options or {}
|
||||
end
|
||||
|
||||
function Expander:statement(...)
|
||||
local s = Statement(self.fn, self.line, ...)
|
||||
return s
|
||||
end
|
||||
|
||||
function Expander:push(s)
|
||||
s:validate()
|
||||
insert(self.statements, s)
|
||||
end
|
||||
|
||||
function Expander:push_new(...)
|
||||
self:push(self:statement(...))
|
||||
end
|
||||
|
||||
function Expander:pop(kind)
|
||||
local ret
|
||||
if kind == nil then
|
||||
ret = self.s[self.i]
|
||||
elseif kind == 'CPU' then
|
||||
ret = self:register(data.registers)
|
||||
elseif kind == 'DEREF' then
|
||||
ret = self:deref()
|
||||
elseif kind == 'CONST' then
|
||||
ret = self:const()
|
||||
elseif kind == 'END' then
|
||||
if self.s[self.i] ~= nil then
|
||||
self:error('expected EOL; too many arguments')
|
||||
end
|
||||
return -- don't increment self.i past end of arguments
|
||||
else
|
||||
error('Internal Error: unknown kind, got '..tostring(kind))
|
||||
end
|
||||
self.i = self.i + 1
|
||||
return ret
|
||||
end
|
||||
|
||||
function Expander:expand(statements)
|
||||
-- fourth pass: expand pseudo-instructions and register arguments
|
||||
self.statements = {}
|
||||
for i, s in ipairs(statements) do
|
||||
self.s = s
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
if s.type:sub(1, 1) == '!' then
|
||||
self:push(s)
|
||||
else
|
||||
local name = s.type
|
||||
local h = data.instructions[name]
|
||||
if h == nil then
|
||||
error('Internal Error: unknown instruction')
|
||||
end
|
||||
|
||||
if data.one_register_variants[name] then
|
||||
self.i = 1
|
||||
local a = self:register(data.all_registers)
|
||||
local b = s[2]
|
||||
if b == nil or b.tt ~= 'REG' then
|
||||
insert(s, 2, self:token(a))
|
||||
end
|
||||
elseif data.two_register_variants[name] then
|
||||
self.i = 1
|
||||
local a = self:register(data.all_registers)
|
||||
local b = self:register(data.all_registers)
|
||||
local c = s[3]
|
||||
if c == nil or c.tt ~= 'REG' then
|
||||
insert(s, 2, self:token(a))
|
||||
end
|
||||
end
|
||||
|
||||
if overrides[name] then
|
||||
self.i = 1
|
||||
overrides[name](self, name)
|
||||
self:pop('END')
|
||||
else
|
||||
self:push(s)
|
||||
end
|
||||
end
|
||||
end
|
||||
|
||||
return self.statements
|
||||
end
|
||||
|
||||
return Expander
|
|
@ -289,7 +289,7 @@ function Lexer:lex_string_naive(yield) -- no escape sequences
|
|||
yield('STRING', buff)
|
||||
end
|
||||
|
||||
function Lexer:lex_include(_yield)
|
||||
function Lexer:lex_filename(_yield)
|
||||
self:read_spaces()
|
||||
local fn
|
||||
self:lex_string_naive(function(tt, tok)
|
||||
|
@ -297,6 +297,18 @@ function Lexer:lex_include(_yield)
|
|||
end)
|
||||
_yield('STRING', fn, self.fn, self.line)
|
||||
|
||||
if self.chr ~= '\n' then
|
||||
self:error('expected EOL after filename')
|
||||
end
|
||||
_yield('EOL', '\n', self.fn, self.line)
|
||||
self:nextc()
|
||||
|
||||
return fn
|
||||
end
|
||||
|
||||
function Lexer:lex_include(_yield)
|
||||
local fn = self:lex_filename(_yield)
|
||||
|
||||
if self.options.path then
|
||||
fn = self.options.path..fn
|
||||
end
|
||||
|
@ -308,25 +320,15 @@ function Lexer:lex_include(_yield)
|
|||
end
|
||||
|
||||
function Lexer:lex_include_binary(_yield)
|
||||
self:read_spaces()
|
||||
local fn
|
||||
self:lex_string_naive(function(tt, tok)
|
||||
fn = tok
|
||||
end)
|
||||
_yield('STRING', fn, self.fn, self.line)
|
||||
local fn = self:lex_filename(_yield)
|
||||
|
||||
-- TODO: allow optional offset and size arguments
|
||||
if self.options.path then
|
||||
fn = self.options.path..fn
|
||||
end
|
||||
local data = util.readfile(fn, true)
|
||||
|
||||
-- FIXME: this allocates a table for each byte.
|
||||
-- this could easily cause performance issues on big files.
|
||||
_yield('DIR', 'BYTE', fn, 0)
|
||||
for b in string.gfind(data, '.') do
|
||||
_yield('NUM', string.byte(b), fn, 0)
|
||||
end
|
||||
_yield('DIR', 'BIN', fn, 0)
|
||||
_yield('STRING', data, fn, 0)
|
||||
end
|
||||
|
||||
function Lexer:lex_expression(yield)
|
||||
|
|
|
@ -6,6 +6,7 @@ local Token = require(path.."Token")
|
|||
local Lexer = require(path.."Lexer")
|
||||
local Collector = require(path.."Collector")
|
||||
local Preproc = require(path.."Preproc")
|
||||
local Expander = require(path.."Expander")
|
||||
local Dumper = require(path.."Dumper")
|
||||
|
||||
local Parser = Base:extend()
|
||||
|
@ -37,53 +38,28 @@ function Parser:tokenize(asm)
|
|||
assert(#tokens > 0, 'Internal Error: no tokens after preprocessing')
|
||||
|
||||
local collector = Collector(self.options)
|
||||
self.statements = collector:collect(tokens, self.main_fn)
|
||||
return collector:collect(tokens, self.main_fn)
|
||||
end
|
||||
|
||||
function Parser:debug_dump()
|
||||
local boring = {
|
||||
tt = true,
|
||||
tok = true,
|
||||
fn = true,
|
||||
line = true,
|
||||
}
|
||||
function Parser:dump()
|
||||
for i, s in ipairs(self.statements) do
|
||||
local values = ''
|
||||
for j, t in ipairs(s) do
|
||||
local tok = t.tok
|
||||
if type(tok) == 'number' then
|
||||
tok = ("$%X"):format(tok)
|
||||
end
|
||||
values = values..'\t'..t.tt..'('..tostring(tok)..')'
|
||||
for k, v in pairs(t) do
|
||||
if not boring[k] then
|
||||
values = values..'['..k..'='..tostring(v)..']'
|
||||
end
|
||||
end
|
||||
end
|
||||
values = values:sub(2)
|
||||
print(s.line, s.type, values)
|
||||
print(s.line, s.type, s:dump())
|
||||
end
|
||||
end
|
||||
|
||||
function Parser:parse(asm)
|
||||
self:tokenize(asm)
|
||||
self.statements = self:tokenize(asm)
|
||||
if self.options.debug_token then self:dump() end
|
||||
|
||||
if self.options.debug_token then self:debug_dump() end
|
||||
self.statements = Preproc(self.options):process(self.statements)
|
||||
if self.options.debug_pre then self:dump() end
|
||||
|
||||
local preproc = Preproc(self.options)
|
||||
self.statements = preproc:process(self.statements)
|
||||
|
||||
if self.options.debug_pre then self:debug_dump() end
|
||||
|
||||
self.statements = preproc:expand(self.statements)
|
||||
|
||||
if self.options.debug_post then self:debug_dump() end
|
||||
self.statements = Expander(self.options):expand(self.statements)
|
||||
if self.options.debug_post then self:dump() end
|
||||
|
||||
local dumper = Dumper(self.writer, self.options)
|
||||
self.statements = dumper:load(self.statements)
|
||||
|
||||
if self.options.debug_asm then self:debug_dump() end
|
||||
if self.options.debug_asm then self:dump() end
|
||||
|
||||
if self.options.labels then
|
||||
dumper:export_labels(self.options.labels)
|
||||
|
|
|
@ -1,31 +1,33 @@
|
|||
local abs = math.abs
|
||||
local insert = table.insert
|
||||
|
||||
local path = string.gsub(..., "[^.]+$", "")
|
||||
local data = require(path.."data")
|
||||
local overrides = require(path.."overrides")
|
||||
local Statement = require(path.."Statement")
|
||||
local Reader = require(path.."Reader")
|
||||
local Base = require(path.."Base")
|
||||
local Expression = require(path.."Expression")
|
||||
local util = require(path.."util")
|
||||
|
||||
local abs = math.abs
|
||||
local signs = util.signs
|
||||
|
||||
local function signs(s)
|
||||
local start, end_ = s:find('[+-]+')
|
||||
if start ~= 1 then
|
||||
return 0
|
||||
end
|
||||
if s:sub(1, 1) == '+' then
|
||||
return end_
|
||||
elseif s:sub(1, 1) == '-' then
|
||||
return -end_
|
||||
end
|
||||
end
|
||||
|
||||
local Preproc = Reader:extend()
|
||||
local Preproc = Base:extend()
|
||||
function Preproc:init(options)
|
||||
self.options = options or {}
|
||||
end
|
||||
|
||||
function Preproc:iter(statements)
|
||||
assert(statements)
|
||||
local i = 0
|
||||
return function()
|
||||
i = i + 1
|
||||
local s = statements[i]
|
||||
if s == nil then return end
|
||||
self.i = i
|
||||
self.s = s
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
return s
|
||||
end
|
||||
end
|
||||
|
||||
function Preproc:lookup(t)
|
||||
if t.tt == 'VARSYM' then
|
||||
local name = t.tok
|
||||
|
@ -110,8 +112,6 @@ function Preproc:check(s, i, tt)
|
|||
end
|
||||
|
||||
function Preproc:process(statements)
|
||||
self.statements = statements
|
||||
|
||||
self.variables = {}
|
||||
self.plus_labels = {} -- constructed forwards
|
||||
self.minus_labels = {} -- constructed backwards
|
||||
|
@ -119,39 +119,29 @@ function Preproc:process(statements)
|
|||
|
||||
-- first pass: resolve variables and collect relative labels
|
||||
local new_statements = {}
|
||||
for i=1, #self.statements do
|
||||
local s = self.statements[i]
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
if s.type:sub(1, 1) == '!' then
|
||||
-- directive, label, etc.
|
||||
if s.type == '!VAR' then
|
||||
local a = self:check(s, 1, 'VAR')
|
||||
local b = self:check(s, 2, 'NUM')
|
||||
self.variables[a] = b
|
||||
elseif s.type == '!LABEL' then
|
||||
if s[1].tt == 'RELLABEL' then
|
||||
local label = s[1].tok
|
||||
local rl = {
|
||||
index = #new_statements + 1,
|
||||
name = label:sub(2)
|
||||
}
|
||||
local c = label:sub(1, 1)
|
||||
if c == '+' then
|
||||
insert(self.plus_labels, rl)
|
||||
elseif c == '-' then
|
||||
insert(self.minus_labels, 1, rl) -- remember, it's backwards
|
||||
else
|
||||
error('Internal Error: unexpected token for relative label')
|
||||
end
|
||||
for s in self:iter(statements) do
|
||||
-- directive, label, etc.
|
||||
if s.type == '!VAR' then
|
||||
local a = self:check(s, 1, 'VAR')
|
||||
local b = self:check(s, 2, 'NUM')
|
||||
self.variables[a] = b
|
||||
elseif s.type == '!LABEL' then
|
||||
if s[1].tt == 'RELLABEL' then
|
||||
local label = s[1].tok
|
||||
local rl = {
|
||||
index = #new_statements + 1,
|
||||
name = label:sub(2)
|
||||
}
|
||||
local c = label:sub(1, 1)
|
||||
if c == '+' then
|
||||
insert(self.plus_labels, rl)
|
||||
elseif c == '-' then
|
||||
insert(self.minus_labels, 1, rl) -- remember, it's backwards
|
||||
else
|
||||
error('Internal Error: unexpected token for relative label')
|
||||
end
|
||||
insert(new_statements, s)
|
||||
else
|
||||
for j, t in ipairs(s) do
|
||||
self:lookup(t)
|
||||
end
|
||||
insert(new_statements, s)
|
||||
end
|
||||
insert(new_statements, s)
|
||||
else
|
||||
-- regular instruction
|
||||
for j, t in ipairs(s) do
|
||||
|
@ -163,21 +153,14 @@ function Preproc:process(statements)
|
|||
|
||||
-- second pass: resolve relative labels
|
||||
self.do_labels = true
|
||||
for i=1, #new_statements do
|
||||
self.i = i -- make visible to :lookup
|
||||
local s = new_statements[i]
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
for s in self:iter(new_statements) do
|
||||
for j, t in ipairs(s) do
|
||||
self:lookup(t)
|
||||
end
|
||||
end
|
||||
|
||||
-- third pass: evaluate constant expressions
|
||||
for i=1, #new_statements do
|
||||
local s = new_statements[i]
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
for s in self:iter(new_statements) do
|
||||
for j, t in ipairs(s) do
|
||||
if t.tt == 'EXPR' then
|
||||
local expr = Expression()
|
||||
|
@ -194,89 +177,4 @@ function Preproc:process(statements)
|
|||
return new_statements
|
||||
end
|
||||
|
||||
function Preproc:statement(...)
|
||||
self.fn = self.s.fn
|
||||
self.line = self.s.line
|
||||
local s = Statement(self.fn, self.line, ...)
|
||||
return s
|
||||
end
|
||||
|
||||
function Preproc:push(s)
|
||||
s:validate()
|
||||
insert(self.statements, s)
|
||||
end
|
||||
|
||||
function Preproc:push_new(...)
|
||||
self:push(self:statement(...))
|
||||
end
|
||||
|
||||
function Preproc:pop(kind)
|
||||
local ret
|
||||
if kind == nil then
|
||||
ret = self.s[self.i]
|
||||
elseif kind == 'CPU' then
|
||||
ret = self:register(data.registers)
|
||||
elseif kind == 'DEREF' then
|
||||
ret = self:deref()
|
||||
elseif kind == 'CONST' then
|
||||
ret = self:const()
|
||||
elseif kind == 'END' then
|
||||
if self.s[self.i] ~= nil then
|
||||
self:error('expected EOL; too many arguments')
|
||||
end
|
||||
return -- don't increment self.i past end of arguments
|
||||
else
|
||||
error('Internal Error: unknown kind, got '..tostring(kind))
|
||||
end
|
||||
self.i = self.i + 1
|
||||
return ret
|
||||
end
|
||||
|
||||
function Preproc:expand(statements)
|
||||
-- fourth pass: expand pseudo-instructions and register arguments
|
||||
self.statements = {}
|
||||
for i=1, #statements do
|
||||
local s = statements[i]
|
||||
self.s = s
|
||||
self.fn = s.fn
|
||||
self.line = s.line
|
||||
if s.type:sub(1, 1) == '!' then
|
||||
self:push(s)
|
||||
else
|
||||
local name = s.type
|
||||
local h = data.instructions[name]
|
||||
if h == nil then
|
||||
error('Internal Error: unknown instruction')
|
||||
end
|
||||
|
||||
if data.one_register_variants[name] then
|
||||
self.i = 1
|
||||
local a = self:register(data.all_registers)
|
||||
local b = s[2]
|
||||
if b == nil or b.tt ~= 'REG' then
|
||||
insert(s, 2, self:token(a))
|
||||
end
|
||||
elseif data.two_register_variants[name] then
|
||||
self.i = 1
|
||||
local a = self:register(data.all_registers)
|
||||
local b = self:register(data.all_registers)
|
||||
local c = s[3]
|
||||
if c == nil or c.tt ~= 'REG' then
|
||||
insert(s, 2, self:token(a))
|
||||
end
|
||||
end
|
||||
|
||||
if overrides[name] then
|
||||
self.i = 1
|
||||
overrides[name](self, name)
|
||||
self:pop('END')
|
||||
else
|
||||
self:push(s)
|
||||
end
|
||||
end
|
||||
end
|
||||
|
||||
return self.statements
|
||||
end
|
||||
|
||||
return Preproc
|
||||
|
|
|
@ -50,4 +50,28 @@ function Statement:validate(n)
|
|||
end
|
||||
end
|
||||
|
||||
local boring = {
|
||||
tt = true,
|
||||
tok = true,
|
||||
fn = true,
|
||||
line = true,
|
||||
}
|
||||
|
||||
function Statement:dump()
|
||||
local values = ''
|
||||
for j, t in ipairs(self) do
|
||||
local tok = t.tok
|
||||
if type(tok) == 'number' then
|
||||
tok = ("$%X"):format(tok)
|
||||
end
|
||||
values = values..'\t'..t.tt..'('..tostring(tok)..')'
|
||||
for k, v in pairs(t) do
|
||||
if not boring[k] then
|
||||
values = values..'['..k..'='..tostring(v)..']'
|
||||
end
|
||||
end
|
||||
end
|
||||
return values:sub(2)
|
||||
end
|
||||
|
||||
return Statement
|
||||
|
|
|
@ -2,11 +2,21 @@ local format = string.format
|
|||
local insert = table.insert
|
||||
|
||||
local path = string.gsub(..., "[^.]+$", "")
|
||||
local data = require(path.."data")
|
||||
local Base = require(path.."Base")
|
||||
local Token = require(path.."Token")
|
||||
|
||||
local arg_types = {
|
||||
local Iter = {}
|
||||
function Iter:__call()
|
||||
return self:next(1)
|
||||
end
|
||||
|
||||
local TokenIter = {}
|
||||
function TokenIter:init(tokens)
|
||||
assert(tokens ~= nil)
|
||||
self.tokens = tokens
|
||||
self:reset()
|
||||
end
|
||||
|
||||
TokenIter.arg_types = {
|
||||
NUM = true,
|
||||
REG = true,
|
||||
VARSYM = true,
|
||||
|
@ -14,17 +24,57 @@ local arg_types = {
|
|||
RELLABELSYM = true,
|
||||
}
|
||||
|
||||
local Muncher = Base:extend()
|
||||
-- no base init method
|
||||
|
||||
function Muncher:error(msg, got)
|
||||
function TokenIter:error(msg, got)
|
||||
if got ~= nil then
|
||||
msg = msg..', got '..tostring(got)
|
||||
end
|
||||
error(format('%s:%d: Error: %s', self.fn, self.line, msg), 2)
|
||||
end
|
||||
|
||||
function Muncher:token(t, val)
|
||||
function TokenIter:reset()
|
||||
self.i = 0
|
||||
self.tt = nil
|
||||
self.tok = nil
|
||||
self.fn = nil
|
||||
self.line = nil
|
||||
self.ended = false
|
||||
end
|
||||
|
||||
function TokenIter:advance(n)
|
||||
n = n or 0
|
||||
if self.ended then
|
||||
error('Internal Error: attempted to advance iterator past end', 2 + n)
|
||||
end
|
||||
|
||||
self.i = self.i + 1
|
||||
self.t = self.tokens[self.i]
|
||||
if self.t == nil then
|
||||
self.tt = nil
|
||||
self.tok = nil
|
||||
self.fn = nil
|
||||
self.line = nil
|
||||
self.ended = true
|
||||
else
|
||||
self.tt = self.t.tt
|
||||
self.tok = self.t.tok
|
||||
self.fn = self.t.fn
|
||||
self.line = self.t.line
|
||||
end
|
||||
end
|
||||
|
||||
function TokenIter:next(n)
|
||||
n = n or 0
|
||||
self:advance(n + 1)
|
||||
if self.t then return self.t end
|
||||
end
|
||||
|
||||
function TokenIter:peek()
|
||||
return self.tokens[self.i + 1]
|
||||
end
|
||||
|
||||
-- now begins the parsing stuff
|
||||
|
||||
function TokenIter:token(t, val)
|
||||
-- note: call Token directly if you want to specify fn and line manually
|
||||
if type(t) == 'table' then
|
||||
t.fn = self.fn
|
||||
|
@ -37,36 +87,25 @@ function Muncher:token(t, val)
|
|||
end
|
||||
end
|
||||
|
||||
function Muncher:advance()
|
||||
self.i = self.i + 1
|
||||
self.t = self.tokens[self.i]
|
||||
self.tt = self.t.tt
|
||||
self.tok = self.t.tok
|
||||
self.fn = self.t.fn
|
||||
self.line = self.t.line
|
||||
return self.t
|
||||
end
|
||||
|
||||
function Muncher:is_EOL()
|
||||
function TokenIter:is_EOL()
|
||||
return self.tt == 'EOL' or self.tt == 'EOF'
|
||||
end
|
||||
|
||||
function Muncher:expect_EOL()
|
||||
function TokenIter:expect_EOL()
|
||||
if self:is_EOL() then
|
||||
self:advance()
|
||||
return
|
||||
end
|
||||
self:error('expected end of line', self.tt)
|
||||
end
|
||||
|
||||
function Muncher:optional_comma()
|
||||
function TokenIter:eat_comma()
|
||||
if self.tt == 'SEP' and self.tok == ',' then
|
||||
self:advance()
|
||||
return true
|
||||
end
|
||||
end
|
||||
|
||||
function Muncher:number()
|
||||
function TokenIter:number()
|
||||
if self.tt ~= 'NUM' then
|
||||
self:error('expected number', self.tt)
|
||||
end
|
||||
|
@ -75,7 +114,7 @@ function Muncher:number()
|
|||
return self:token(t)
|
||||
end
|
||||
|
||||
function Muncher:string()
|
||||
function TokenIter:string()
|
||||
if self.tt ~= 'STRING' then
|
||||
self:error('expected string', self.tt)
|
||||
end
|
||||
|
@ -84,7 +123,7 @@ function Muncher:string()
|
|||
return self:token(t)
|
||||
end
|
||||
|
||||
function Muncher:register(registers)
|
||||
function TokenIter:register(registers)
|
||||
registers = registers or data.registers
|
||||
if self.tt ~= 'REG' then
|
||||
self:error('expected register', self.tt)
|
||||
|
@ -97,7 +136,7 @@ function Muncher:register(registers)
|
|||
return self:token(t)
|
||||
end
|
||||
|
||||
function Muncher:deref()
|
||||
function TokenIter:deref()
|
||||
if self.tt ~= 'OPEN' then
|
||||
self:error('expected opening parenthesis for dereferencing', self.tt)
|
||||
end
|
||||
|
@ -111,10 +150,10 @@ function Muncher:deref()
|
|||
self:error('expected closing parenthesis for dereferencing', self.tt)
|
||||
end
|
||||
self:advance()
|
||||
return self:token(t)
|
||||
return self:token(t):set('tt', 'DEREF')
|
||||
end
|
||||
|
||||
function Muncher:const(relative, no_label)
|
||||
function TokenIter:const(relative, no_label)
|
||||
local good = {
|
||||
NUM = true,
|
||||
EXPR = true,
|
||||
|
@ -132,7 +171,7 @@ function Muncher:const(relative, no_label)
|
|||
return t
|
||||
end
|
||||
|
||||
function Muncher:special()
|
||||
function TokenIter:special()
|
||||
if self.tt ~= 'SPECIAL' then
|
||||
self:error('expected special name to call', self.tt)
|
||||
end
|
||||
|
@ -145,7 +184,7 @@ function Muncher:special()
|
|||
local args = {}
|
||||
while true do
|
||||
local arg = self:advance()
|
||||
if not arg_types[arg.tt] then
|
||||
if not self.arg_types[arg.tt] then
|
||||
self:error('invalid argument type', arg.tt)
|
||||
else
|
||||
self:advance()
|
||||
|
@ -163,4 +202,39 @@ function Muncher:special()
|
|||
return name, args
|
||||
end
|
||||
|
||||
return Muncher
|
||||
function TokenIter:basic_special()
|
||||
local name, args = self:special()
|
||||
|
||||
local portion
|
||||
if name == 'hi' then
|
||||
portion = 'upperoff'
|
||||
elseif name == 'up' then
|
||||
portion = 'upper'
|
||||
elseif name == 'lo' then
|
||||
portion = 'lower'
|
||||
else
|
||||
self:error('unknown special', name)
|
||||
end
|
||||
|
||||
if #args ~= 1 then
|
||||
self:error(name..' expected one argument', #args)
|
||||
end
|
||||
|
||||
local t = self:token(args[1]):set('portion', portion)
|
||||
return t
|
||||
end
|
||||
|
||||
-- TODO: move this boilerplate elsewhere
|
||||
|
||||
local MetaBlah = {
|
||||
__index = TokenIter,
|
||||
__call = TokenIter.next,
|
||||
}
|
||||
|
||||
local ClassBlah = {}
|
||||
function ClassBlah:__call(...)
|
||||
local obj = setmetatable({}, MetaBlah)
|
||||
return obj, obj:init(...)
|
||||
end
|
||||
|
||||
return setmetatable(TokenIter, ClassBlah)
|
|
@ -47,7 +47,7 @@ local function li(self, buffer, dest, im)
|
|||
end
|
||||
|
||||
local overrides = {}
|
||||
-- note: "self" is an instance of Preproc
|
||||
-- note: "self" is an instance of Expander
|
||||
|
||||
local function tob_override(self, name)
|
||||
-- handle all the addressing modes for lw/sw-like instructions
|
||||
|
|
|
@ -14,7 +14,7 @@ local function readfile(fn, binary)
|
|||
end
|
||||
|
||||
local function bitrange(x, lower, upper)
|
||||
return floor(x/2^lower) % 2^(upper - lower + 1)
|
||||
return floor(x / 2^lower) % 2^(upper - lower + 1)
|
||||
end
|
||||
|
||||
local function parent(t)
|
||||
|
@ -25,6 +25,18 @@ local function parent(t)
|
|||
return mt.__index
|
||||
end
|
||||
|
||||
local function signs(s)
|
||||
local start, end_ = s:find('[+-]+')
|
||||
if start ~= 1 then
|
||||
return 0
|
||||
end
|
||||
if s:sub(1, 1) == '+' then
|
||||
return end_
|
||||
elseif s:sub(1, 1) == '-' then
|
||||
return -end_
|
||||
end
|
||||
end
|
||||
|
||||
-- http://stackoverflow.com/a/9279009
|
||||
local loadcode
|
||||
if setfenv and loadstring then -- 5.1, JIT
|
||||
|
@ -77,6 +89,7 @@ return {
|
|||
readfile = readfile,
|
||||
bitrange = bitrange,
|
||||
parent = parent,
|
||||
signs = signs,
|
||||
loadcode = loadcode,
|
||||
measure_data = measure_data,
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue