New module for parser functions

Same principle as main->lib (previous commit).

1 files changed, 144 insertions, 0 deletions

diff --git a/parser.c b/parser.c
new file mode 100644
index 0000000..d1bc9c9
--- /dev/null
+++ b/parser.c
@@ -0,0 +1,144 @@
+/* parser.c
+ * Created: 2023-09-02
+ * Author: Aryadev Chavali
+ * Description: Functions for (de)serialising
+ */
+
+#include <string.h>
+
+#include "./parser.h"
+
+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;
+}
+
+struct PResult parse_buffer(buffer_t *buffer)
+{
+  node_t *nodes = NULL;
+  size_t usable = 0, loops = 0;
+  for (size_t i = 0; i < buffer->size; ++i)
+    if (usable_character(buffer->data[i]))
+    {
+      ++usable;
+      if (buffer->data[i] == '[')
+        ++loops;
+    }
+  nodes = calloc(usable, sizeof(*nodes));
+
+  // First pass: Get my info
+  for (size_t i = 0, col = 0, row = 1, nptr = 0; i < buffer->size; ++i)
+  {
+    ++col;
+    if (buffer->data[i] == '\n')
+    {
+      ++row;
+      col = 0;
+    }
+    else if (usable_character(buffer->data[i]))
+    {
+      nodes[nptr].col = col;
+      nodes[nptr].row = row;
+      switch (buffer->data[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)
+      {
+        print_error(buffer->name, node->row, node->col,
+                    "ERROR: Unbalanced square brackets!");
+        goto error;
+      }
+      // Access last IN loop
+      --stackptr;
+      node->loop_ref            = stack[stackptr] - nodes;
+      stack[stackptr]->loop_ref = i;
+    }
+  }
+
+  if (stackptr > 0)
+  {
+    node_t *node = nodes + usable - 1;
+    print_error(buffer->name, node->row, node->col,
+                "ERROR: Unbalanced square brackets!");
+    goto error;
+  }
+
+  return (struct PResult){nodes, usable};
+error:
+  if (nodes)
+    free(nodes);
+  return (struct PResult){0};
+}