From a47c924e7524b6d398b195d98fbab8cd129858c5 Mon Sep 17 00:00:00 2001
From: Connor Olding <cloningdonor@gmail.com>
Date: Wed, 13 Jan 2016 11:18:36 -0800
Subject: [PATCH] implement ascii/asciiz directives, basic string lexing

---
 README.md       | 11 +++++----
 lips/Dumper.lua |  3 ++-
 lips/Lexer.lua  | 59 ++++++++++++++++++++++++++++++++++++++++++++++---
 lips/Parser.lua | 26 +++++++++++++++++-----
 4 files changed, 86 insertions(+), 13 deletions(-)

diff --git a/README.md b/README.md
index 89cd659..e65730b 100644
--- a/README.md
+++ b/README.md
@@ -145,14 +145,17 @@ include an external assembly file as-is at this position.
 lips will look for the included file
 in the directory of the file using the directive.
 
+* `.ascii "some\ntext\0"`  
+writes a string using its characters' ASCII values.
+a few escapes are currently supported: `\ " a b f n r t v 0`
+
+* `.asciiz "some\ntext"`  
+same as ascii, but with a null byte added to the end.
+
 ### Unimplemented
 
 * FLOAT: writes a list of 32-bit floating point numbers until end-of-line.
 this may not get implemented due to a lack of aliasing in vanilla Lua,
 and thus accuracy issues.
 
-* ASCII: writes a string using its characters' ASCII values.
-
-* ASCIIZ: same as ASCII, but with a null byte added to the end.
-
 * INCBIN: write an external binary file as-is at this position.
diff --git a/lips/Dumper.lua b/lips/Dumper.lua
index faacf02..adb497c 100644
--- a/lips/Dumper.lua
+++ b/lips/Dumper.lua
@@ -1,5 +1,6 @@
-local insert = table.insert
 local floor = math.floor
+local format = string.format
+local insert = table.insert
 
 local data = require "lips.data"
 
diff --git a/lips/Lexer.lua b/lips/Lexer.lua
index 922f454..c5843c2 100644
--- a/lips/Lexer.lua
+++ b/lips/Lexer.lua
@@ -1,10 +1,25 @@
 local byte = string.byte
 local char = string.char
 local find = string.find
+local format = string.format
+local insert = table.insert
 local open = io.open
 
 local data = require "lips.data"
 
+local simple_escapes = {
+    ['0']   = 0x00,
+    ['\\']  = 0x5C,
+    ['"']   = 0x22,
+    ['a']   = 0x07,
+    ['b']   = 0x08,
+    ['f']   = 0x0C,
+    ['n']   = 0x0A,
+    ['r']   = 0x0D,
+    ['t']   = 0x09,
+    ['v']   = 0x0B,
+}
+
 local function readfile(fn)
     local f = open(fn, 'r')
     if not f then
@@ -208,12 +223,48 @@ end
 function Lexer:lex_string(yield)
     -- TODO: support escaping
     if self.chr ~= '"' then
-        self:error("expected opening double quote")
+        self:error('expected opening double quote')
+    end
+    self:nextc()
+
+    local bytes = {}
+    while true do
+        if self.chr == '\n' then
+            self:error('unimplemented')
+            self:nextc()
+            yield('EOL', '\n')
+        elseif self.ord == self.EOF then
+            self:nextc()
+            self:error('unexpected EOF; incomplete string')
+        elseif self.chr == '"' then
+            self:nextc()
+            break
+        elseif self.chr == '\\' then
+            self:nextc()
+            local simple = simple_escapes[self.chr]
+            if simple then
+                insert(bytes, simple)
+            else
+                self:error('unknown escape sequence')
+            end
+            self:nextc()
+        else
+            insert(bytes, byte(self.chr))
+            self:nextc()
+        end
+    end
+
+    yield('STRING', bytes)
+end
+
+function Lexer:lex_string_naive(yield) -- no escape sequences
+    if self.chr ~= '"' then
+        self:error('expected opening double quote')
     end
     self:nextc()
     local buff = self:read_chars('[^"\n]')
     if self.chr ~= '"' then
-        self:error("expected closing double quote")
+        self:error('expected closing double quote')
     end
     self:nextc()
     yield('STRING', buff)
@@ -222,7 +273,7 @@ end
 function Lexer:lex_include(_yield)
     self:read_chars('%s')
     local fn
-    self:lex_string(function(tt, tok)
+    self:lex_string_naive(function(tt, tok)
         fn = tok
     end)
     if self.options.path then
@@ -293,6 +344,8 @@ function Lexer:lex(_yield)
             else
                 yield('DIR', up)
             end
+        elseif self.chr == '"' then
+            self:lex_string(yield)
         elseif self.chr == '@' then
             self:nextc()
             local buff = self:read_chars('[%w_]')
diff --git a/lips/Parser.lua b/lips/Parser.lua
index f3e9bc3..528abc8 100644
--- a/lips/Parser.lua
+++ b/lips/Parser.lua
@@ -1,5 +1,5 @@
-local insert = table.insert
 local format = string.format
+local insert = table.insert
 
 local data = require "lips.data"
 local overrides = require "lips.overrides"
@@ -57,6 +57,15 @@ function Parser:number()
     return value
 end
 
+function Parser:string()
+    if self.tt ~= 'STRING' then
+        self:error('expected string')
+    end
+    local value = self.tok
+    self:advance()
+    return value
+end
+
 function Parser:directive()
     local name = self.tok
     self:advance()
@@ -84,13 +93,20 @@ function Parser:directive()
             self.dumper:add_directive(line, name, self:number())
         end
         self:expect_EOL()
-    elseif name == 'HEX' then
-        self:error('unimplemented')
     elseif name == 'INC' then
-        -- noop
+        -- noop, handled by lexer
+    elseif name == 'ASCII' or name == 'ASCIIZ' then
+        local bytes = self:string()
+        for i, number in ipairs(bytes) do
+            self.dumper:add_directive(line, 'BYTE', number)
+        end
+        if name == 'ASCIIZ' then
+            self.dumper:add_directive(line, 'BYTE', 0)
+        end
+        self:expect_EOL()
     elseif name == 'INCBIN' then
         self:error('unimplemented')
-    elseif name == 'FLOAT' or name == 'ASCII' or name == 'ASCIIZ' then
+    elseif name == 'FLOAT' then
         self:error('unimplemented')
     else
         self:error('unknown directive')