From 2d33640c037c9f3bcbddc200eec39bb713ca0c9f Mon Sep 17 00:00:00 2001
From: Aryadev Chavali <aryadev@aryadevchavali.com>
Date: Sat, 2 Sep 2023 15:21:56 +0100
Subject: INIT

Simple parser, took like an hour.

Not very smart but at least it's like O(n).
---
 .gitignore |   2 +
 Makefile   |  20 +++++++
 main.c     | 193 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 215 insertions(+)
 create mode 100644 .gitignore
 create mode 100644 Makefile
 create mode 100644 main.c

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..386f581
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+*.out
+*.o
\ No newline at end of file
diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000..cfd0ecb
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,20 @@
+CC=gcc
+CFLAGS=-Wall -Wextra -Wpedantic -ggdb -fsanitize=address -std=c11
+LIBS=
+OBJECTS=main.o
+OUT=obf.out
+ARGS=
+
+%.o: %.c
+	$(CC) $(CFLAGS) -c $^ -o $@ $(LIBS)
+
+$(OUT): $(OBJECTS)
+	$(CC) $(CFLAGS) $^ -o $@ $(LIBS)
+
+.PHONY:
+clean:
+	rm -rfv $(OUT) $(OBJECTS)
+
+.PHONY: run
+run: $(OUT)
+	./$^ $(ARGS)
diff --git a/main.c b/main.c
new file mode 100644
index 0000000..5174e3d
--- /dev/null
+++ b/main.c
@@ -0,0 +1,193 @@
+/* main.c
+ * Created: 2023-09-02
+ * Author: Aryadev Chavali
+ * Description: Entrypoint of compiler
+ */
+
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define MEMORY_DEFAULT 30000
+
+typedef struct
+{
+  size_t dp;
+  uint8_t memory[MEMORY_DEFAULT];
+} machine_t;
+
+bool usable_character(char c)
+{
+  return c == '>' || c == '<' || c == '+' || c == '-' || c == '-' || c == '.' ||
+         c == ',' || c == '[' || c == ']';
+}
+
+typedef struct AST
+{
+  size_t col, row;
+  enum
+  {
+    NEXT = 0,
+    PREV,
+    INC,
+    DEC,
+    OUT,
+    READ,
+    LIN,
+    LOUT
+  } type;
+  int loop_ref;
+} node_t;
+
+struct PResult
+{
+  node_t *nodes;
+  size_t size;
+} parse_input(const char *input, size_t input_size)
+{
+  node_t *nodes = NULL;
+  size_t usable = 0, loops = 0;
+  for (size_t i = 0; i < input_size; ++i)
+    if (usable_character(input[i]))
+    {
+      ++usable;
+      if (input[i] == '[')
+        ++loops;
+    }
+  nodes = calloc(usable, sizeof(*nodes));
+
+  // First pass: Get my info
+  for (size_t i = 0, col = 0, row = 1, nptr = 0; i < input_size; ++i)
+  {
+    ++col;
+    if (input[i] == '\n')
+    {
+      ++row;
+      col = 0;
+    }
+    else if (usable_character(input[i]))
+    {
+      nodes[nptr].col = col;
+      nodes[nptr].row = row;
+      switch (input[i])
+      {
+      case '>':
+        nodes[nptr].type = NEXT;
+        break;
+      case '<':
+        nodes[nptr].type = PREV;
+        break;
+      case '+':
+        nodes[nptr].type = INC;
+        break;
+      case '-':
+        nodes[nptr].type = DEC;
+        break;
+      case '.':
+        nodes[nptr].type = OUT;
+        break;
+      case ',':
+        nodes[nptr].type = READ;
+        break;
+      case '[':
+        nodes[nptr].type     = LIN;
+        nodes[nptr].loop_ref = -1;
+        break;
+      case ']':
+        nodes[nptr].type     = LOUT;
+        nodes[nptr].loop_ref = -1;
+        break;
+      }
+      ++nptr;
+    }
+  }
+
+  // Second pass: setup any loop references
+  node_t *stack[loops];
+  memset(stack, 0, loops);
+  size_t stackptr = 0;
+  for (size_t i = 0; i < usable; ++i)
+  {
+    node_t *node = nodes + i;
+    if (node->type == LIN)
+      stack[stackptr++] = node;
+    else if (node->type == LOUT)
+    {
+      if (stackptr == 0)
+      {
+        fputs("ERROR: Unbalanced square brackets!\n", stderr);
+        goto error;
+      }
+      // Access last IN loop
+      --stackptr;
+      node->loop_ref            = stack[stackptr] - nodes;
+      stack[stackptr]->loop_ref = i;
+    }
+  }
+
+  if (stackptr > 0)
+  {
+    fputs("ERROR: Unbalanced square brackets!\n", stderr);
+    goto error;
+  }
+
+  return (struct PResult){nodes, usable};
+error:
+  if (nodes)
+    free(nodes);
+  return (struct PResult){0};
+}
+
+char *ast_to_str(node_t *ast, size_t size)
+{
+  char *out = calloc(size + 1, 1);
+  for (size_t i = 0; i < size; ++i)
+  {
+    switch (ast[i].type)
+    {
+    case NEXT:
+      out[i] = '>';
+      break;
+    case PREV:
+      out[i] = '<';
+      break;
+    case INC:
+      out[i] = '+';
+      break;
+    case DEC:
+      out[i] = '-';
+      break;
+    case OUT:
+      out[i] = '.';
+      break;
+    case READ:
+      out[i] = ',';
+      break;
+    case LIN:
+      out[i] = '[';
+      break;
+    case LOUT:
+      out[i] = ']';
+      break;
+    }
+  }
+  out[size] = '\0';
+  return out;
+}
+
+int main(void)
+{
+  const char input[] = "[[->+<]";
+  struct PResult res = parse_input(input, sizeof(input) / sizeof(input[0]));
+  if (res.nodes == NULL)
+  {
+    fputs("Exiting early...\n", stderr);
+    return 1;
+  }
+  char *str = ast_to_str(res.nodes, res.size);
+  free(str);
+  free(res.nodes);
+  return 0;
+}
-- 
cgit v1.2.3-13-gbd6f