From 7fe860d5d4b70a8da820bd0b3eb8b47ed3c62cff Mon Sep 17 00:00:00 2001
From: Brett Laptop <tri11paragon@tpgc.me>
Date: Mon, 20 Nov 2023 20:33:51 -0500
Subject: [PATCH] sexy new codegen

---
 mips.asm                |  27 +++---
 mips2.asm               | 181 ++++++----------------------------------
 src/bf_mips_codegen.cpp |  49 +++++------
 src/bf_tokenizer.cpp    |  70 +++++++++++-----
 4 files changed, 112 insertions(+), 215 deletions(-)

diff --git a/mips.asm b/mips.asm
index 3cfa3fc..4695ee5 100644
--- a/mips.asm
+++ b/mips.asm
@@ -38,8 +38,11 @@ bf:
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
-	beqz  $t1, BF_CLOSE_1_0
-BF_OPEN_1_0:
+	addi  $t1, $t1, 1
+	sb    $t1, ($t0)
+	lb    $t1, ($t0)
+	beqz  $t1, BF_CLOSE_03e0a731165036472c4dae07_0
+BF_OPEN_03e0a731165036472c4dae07_0:
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
@@ -54,8 +57,8 @@ BF_OPEN_1_0:
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
-	beqz  $t1, BF_CLOSE_2_1
-BF_OPEN_2_1:
+	beqz  $t1, BF_CLOSE_972059b874ce5ef469e931bb_1
+BF_OPEN_972059b874ce5ef469e931bb_1:
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
@@ -95,8 +98,8 @@ BF_OPEN_2_1:
 	subi  $t1, $t1, 1
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
-	bnez  $t1, BF_OPEN_2_1
-BF_CLOSE_2_1:
+	bnez  $t1, BF_OPEN_972059b874ce5ef469e931bb_1
+BF_CLOSE_972059b874ce5ef469e931bb_1:
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
@@ -115,19 +118,19 @@ BF_CLOSE_2_1:
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
-	beqz  $t1, BF_CLOSE_3_1
-BF_OPEN_3_1:
+	beqz  $t1, BF_CLOSE_4f2d57da98e6bb1619411ff9_1
+BF_OPEN_4f2d57da98e6bb1619411ff9_1:
 	subi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	bnez  $t1, BF_OPEN_3_1
-BF_CLOSE_3_1:
+	bnez  $t1, BF_OPEN_4f2d57da98e6bb1619411ff9_1
+BF_CLOSE_4f2d57da98e6bb1619411ff9_1:
 	subi  $t0, $t0, 1
 	lb    $t1, ($t0)
 	subi  $t1, $t1, 1
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
-	bnez  $t1, BF_OPEN_1_0
-BF_CLOSE_1_0:
+	bnez  $t1, BF_OPEN_03e0a731165036472c4dae07_0
+BF_CLOSE_03e0a731165036472c4dae07_0:
 	addi  $t0, $t0, 1
 	addi  $t0, $t0, 1
 	li    $v0, 11
diff --git a/mips2.asm b/mips2.asm
index c15b3d6..24e4f59 100644
--- a/mips2.asm
+++ b/mips2.asm
@@ -1,96 +1,48 @@
+
 .data
-	data_address: 	.word 	0
-	data_pointer:	.word	0
-	newline: .asciiz "\n"
+	data_pointer:   .word   0
 .text
 setup:
 	li $v0, 9
 	li $a0, 30000
 	syscall
 	
-	la $t1, data_address
+	la $t1, data_pointer
 	sw $v0, ($t1)
 	
 	# t0 - current address (data_pointer)
-	# t1 -
-	# t2 - temp
 	move $t0, $v0
 bf:
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 8
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
 	beqz  $t1, BF_CLOSE_1
 BF_OPEN_1:
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 4
 	sb    $t1, ($t0)
 	lb    $t1, ($t0)
 	beqz  $t1, BF_CLOSE_2
 BF_OPEN_2:
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 2
 	sb    $t1, ($t0)
+	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 3
+	sb    $t1, ($t0)
+	addi  $t0, $t0, 1
+	lb    $t1, ($t0)
+	addi  $t1, $t1, 3
 	sb    $t1, ($t0)
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	addi  $t0, $t0, 1
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	addi  $t0, $t0, 1
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	subi  $t0, $t0, 1
-	subi  $t0, $t0, 1
-	subi  $t0, $t0, 1
-	subi  $t0, $t0, 1
+	subi  $t0, $t0, 4
 	lb    $t1, ($t0)
 	subi  $t1, $t1, 1
 	sb    $t1, ($t0)
@@ -109,8 +61,7 @@ BF_CLOSE_2:
 	lb    $t1, ($t0)
 	subi  $t1, $t1, 1
 	sb    $t1, ($t0)
-	addi  $t0, $t0, 1
-	addi  $t0, $t0, 1
+	addi  $t0, $t0, 2
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
@@ -128,65 +79,30 @@ BF_CLOSE_3:
 	lb    $t1, ($t0)
 	bnez  $t1, BF_OPEN_1
 BF_CLOSE_1:
-	addi  $t0, $t0, 1
-	addi  $t0, $t0, 1
+	addi  $t0, $t0, 2
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
+	subi  $t1, $t1, 3
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 7
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
-	li    $v0, 11
-	lb    $a0, ($t0)
-	syscall
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 3
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
-	addi  $t0, $t0, 1
-	addi  $t0, $t0, 1
+	addi  $t0, $t0, 2
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
@@ -202,67 +118,24 @@ BF_CLOSE_1:
 	lb    $a0, ($t0)
 	syscall
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 3
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
 	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
+	subi  $t1, $t1, 6
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
 	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	subi  $t1, $t1, 1
+	subi  $t1, $t1, 8
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
-	addi  $t0, $t0, 1
-	addi  $t0, $t0, 1
+	addi  $t0, $t0, 2
 	lb    $t1, ($t0)
 	addi  $t1, $t1, 1
 	sb    $t1, ($t0)
@@ -271,11 +144,11 @@ BF_CLOSE_1:
 	syscall
 	addi  $t0, $t0, 1
 	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
-	sb    $t1, ($t0)
-	lb    $t1, ($t0)
-	addi  $t1, $t1, 1
+	addi  $t1, $t1, 2
 	sb    $t1, ($t0)
 	li    $v0, 11
 	lb    $a0, ($t0)
 	syscall
+	li    $v0, 11
+	lb    $a0, ($t0)
+	syscall
diff --git a/src/bf_mips_codegen.cpp b/src/bf_mips_codegen.cpp
index 6776bd0..5fe8567 100644
--- a/src/bf_mips_codegen.cpp
+++ b/src/bf_mips_codegen.cpp
@@ -6,24 +6,21 @@
 #include <bf_mips_codegen.h>
 #include <iostream>
 
-std::string preamble = ".data\n"
-                       "\tdata_address: \t.word \t0\n"
-                       "\tdata_pointer:\t.word\t0\n"
-                       "\tnewline: .asciiz \"\\n\"\n"
-                       ".text\n"
-                       "setup:\n"
-                       "\tli $v0, 9\n"
-                       "\tli $a0, 30000\n"
-                       "\tsyscall\n"
-                       "\t\n"
-                       "\tla $t1, data_address\n"
-                       "\tsw $v0, ($t1)\n"
-                       "\t\n"
-                       "\t# t0 - current address (data_pointer)\n"
-                       "\t# t1 -\n"
-                       "\t# t2 - temp\n"
-                       "\tmove $t0, $v0\n"
-                       "bf:\n";
+std::string preamble = R"(.data
+	data_pointer:   .word   0
+.text
+setup:
+	li $v0, 9
+	li $a0, 30000
+	syscall
+	
+	la $t1, data_pointer
+	sw $v0, ($t1)
+	
+	# t0 - current address (data_pointer)
+	move $t0, $v0
+bf:
+)";
 
 void process_print(const std::vector<token_t>& tokens, size_t index);
 
@@ -33,12 +30,12 @@ void codegen(tokenizer& tokenizer, std::ostream& out)
     if (!tokenizer.hasNext())
         throw std::runtime_error("You failed to provide valid BF code");
     // skip past comments
-    if (tokenizer.next().type == bf_token::OPEN)
-        while (tokenizer.hasNext() && tokenizer.next().type != bf_token::CLOSE);
+    if (tokenizer.peek().type == bf_token::OPEN)
+        while (tokenizer.hasNext() && tokenizer.peek().type != bf_token::CLOSE)
+            tokenizer.advance();
     
     tokenizer.print();
     
-    size_t sp = 0;
     while (tokenizer.hasNext())
     {
         auto& token = tokenizer.next();
@@ -48,19 +45,19 @@ void codegen(tokenizer& tokenizer, std::ostream& out)
         switch (token.type)
         {
             case bf_token::INC_DP:
-                out       << "\taddi  $t0, $t0, 1\n";
+                out       << "\taddi  $t0, $t0, " << token.offset << "\n";
                 break;
             case bf_token::DEC_DP:
-                out       << "\tsubi  $t0, $t0, 1\n";
+                out       << "\tsubi  $t0, $t0, " << token.offset << "\n";
                 break;
             case bf_token::INC_DV:
                 out       << "\tlb    $t1, ($t0)\n"
-                          << "\taddi  $t1, $t1, 1\n"
+                          << "\taddi  $t1, $t1, " << static_cast<uint8_t>(token.offset) << "\n"
                           << "\tsb    $t1, ($t0)\n";
                 break;
             case bf_token::DEC_DV:
                 out << "\tlb    $t1, ($t0)\n"
-                          << "\tsubi  $t1, $t1, 1\n"
+                          << "\tsubi  $t1, $t1, " << static_cast<uint8_t>(token.offset) << "\n"
                           << "\tsb    $t1, ($t0)\n";
                 break;
             case bf_token::PRINT:
@@ -77,10 +74,8 @@ void codegen(tokenizer& tokenizer, std::ostream& out)
                 out       << "\tlb    $t1, ($t0)\n"
                           << "\tbeqz  $t1, BF_CLOSE_" << name << '\n'
                           << "BF_OPEN_" << name << ":\n";
-                sp++;
                 break;
             case bf_token::CLOSE:
-                sp--;
                 out       << "\tlb    $t1, ($t0)\n"
                           << "\tbnez  $t1, BF_OPEN_" << name << '\n'
                           << "BF_CLOSE_" << name << ":\n";
diff --git a/src/bf_tokenizer.cpp b/src/bf_tokenizer.cpp
index d160c1c..c016ecf 100644
--- a/src/bf_tokenizer.cpp
+++ b/src/bf_tokenizer.cpp
@@ -8,64 +8,90 @@
 #include <random>
 #include <iostream>
 
-class characterizer {
+struct char_pair
+{
+    char type;
+    size_t count;
+};
+
+class characterizer
+{
     private:
         std::string program;
         size_t currentIndex = 0;
     public:
-        inline bool hasNext(){
+        explicit characterizer(std::string program): program(std::move(program))
+        {}
+        
+        inline bool hasNext()
+        {
             return currentIndex < program.size();
         }
         
-        inline size_t advance(){
+        inline size_t advance()
+        {
             return currentIndex++;
         }
         
-        inline char next(){
+        inline char next()
+        {
             return program[advance()];
         }
         
-        inline char peek(){
+        inline char_pair fetch()
+        {
+            size_t start = currentIndex;
+            char type = next();
+            
+            while (peek() == type)
+                advance();
+            
+            return {type, currentIndex - start};
+        }
+        
+        inline char peek()
+        {
             return program[currentIndex];
         }
 };
 
 void tokenizer::tokenize(const std::string& program)
 {
-    size_t index = 0;
-    while (index < program.size())
+    characterizer tk{program};
+    while (tk.hasNext())
     {
-        auto c = program[index];
-        switch (c)
+        auto dv = tk.fetch();
+        bf_token type = bf_token::PRINT;
+        switch (dv.type)
         {
             case '>':
-                tokens.emplace_back(bf_token::INC_DP);
+                type = (bf_token::INC_DP);
                 break;
             case '<':
-                tokens.emplace_back(bf_token::DEC_DP);
+                type = (bf_token::DEC_DP);
                 break;
             case '+':
-                tokens.emplace_back(bf_token::INC_DV);
+                type = (bf_token::INC_DV);
                 break;
             case '-':
-                tokens.emplace_back(bf_token::DEC_DV);
+                type = (bf_token::DEC_DV);
                 break;
             case '.':
-                tokens.emplace_back(bf_token::PRINT);
+                type = (bf_token::PRINT);
                 break;
             case ',':
-                tokens.emplace_back(bf_token::READ);
+                type = (bf_token::READ);
                 break;
             case '[':
-                tokens.emplace_back(bf_token::OPEN);
+                type = (bf_token::OPEN);
                 break;
             case ']':
-                tokens.emplace_back(bf_token::CLOSE);
+                type = (bf_token::CLOSE);
                 break;
             default:
                 break;
         }
-        index++;
+        tokens.emplace_back(type, dv.count);
     }
 }
 
@@ -111,19 +137,19 @@ void tokenizer::print(size_t index)
         {
             case bf_token::INC_DP:
                 tabulate(sp);
-                std::cout << "Increase DP\n";
+                std::cout << "Increase DP " << token.offset << "\n";
                 break;
             case bf_token::DEC_DP:
                 tabulate(sp);
-                std::cout << "Decrease DP\n";
+                std::cout << "Decrease DP " << token.offset << "\n";
                 break;
             case bf_token::INC_DV:
                 tabulate(sp);
-                std::cout << "Increase DV\n";
+                std::cout << "Increase DV " << token.offset << "\n";
                 break;
             case bf_token::DEC_DV:
                 tabulate(sp);
-                std::cout << "Decrease DV\n";
+                std::cout << "Decrease DV " << token.offset << "\n";
                 break;
             case bf_token::PRINT:
                 tabulate(sp);