replaced

diff --git a/src/apc/lexer_lex.rl b/src/apc/lexer_lex.rl
new file mode 100644 (file)
index 0000000..4c9aeea
--- /dev/null
+++ b/src/apc/lexer_lex.rl
@@ -0,0 +1,114 @@
+/* Ragel State Machine for tokenizing text */
+#include <stdio.h>
+#include <string.h>
+#include <apc/parser.tab.h>
+
+extern void lexer_pushtok(int, YYSTYPE);
+
+int lexer_lex(const char*);
+int ipow(int, int);
+int ttov(const char* str, int);
+uint64_t ttor(const char* str, int);
+char* ttos(const char* str, int);
+
+
+#define MAX_TOK_LEN 64
+#define MAX_TOKENS 16
+#define MAX_STR_SIZE (MAX_TOK_LEN * MAX_TOKENS)
+#define $($)#$
+#define PUSHTOK(TOK,LFUNC,UTYPE)                       \
+  do {                                                 \
+    printf("PUSHTOK(" $(TOK) $(LFUNC) $(UTYPE) ")\n"); \
+    tok_t = TOK;                                       \
+    yylval.UTYPE = LFUNC(ts, p-ts);                    \
+    lexer_pushtok(tok_t, yylval);                      \
+    ++ntok;                                            \
+  } while (0)
+
+%%{
+  machine token_matcher;
+
+  # set up yylval and tok_t to be pushed to stack
+  action set_ref  { PUSHTOK(REF, ttor, ref); }
+  action set_val  { PUSHTOK(NUM, ttov, val); }
+  action set_name { PUSHTOK(NAME, ttos, str); }
+  action set_ts   { ts = p; }
+
+  # instantiate machines for each possible token
+  ref = '0x' xdigit+ %set_ref;
+  val = digit+ %set_val;
+  name = alpha+ %set_name;
+  tok = ref | val | name;
+  segment = (tok . '_') %set_ts;
+
+  main := segment* . tok;
+}%%
+
+
+%%write data;
+
+/* Scan filename and push the its tokens
+   onto the stack */
+int lexer_lex (const char* str)
+{
+  const char *p, *pe, *ts, *eof;
+  int  cs, tok_t, ntok = 0;
+  printf ("Lexing: %s\n",str);
+  p = ts = str;
+  pe = p + strlen(str) + 1;
+  %%write init;
+  %%write exec;
+  printf ("Lexed %i tokens\n",ntok);
+  return ntok;
+}
+
+int ipow(int base, int exp)
+{
+  int result = 1;
+  while (exp)
+    {
+      if (exp & 1)
+        result = result * base;
+      exp = exp >> 1;
+      base *= base;
+    }
+
+  return result;
+}
+
+/*  Token to Value */
+int ttov(const char* str, int len)
+{
+  int i, val = 0;
+
+  for (i = 0; i < len; i++)
+    {
+      val += ((str[len - (i + 1)] - '0') * ipow(10,i));
+    }
+
+  return val;
+}
+
+uint64_t ttor(const char* str, int len)
+{
+  int i;
+  uint64_t num = 0;
+
+  for (i = 0; i < len; i++)
+    {
+      num += ((str[len - (i + 1)] - '0') * ipow(10,i));
+    }
+
+  return num;
+}
+
+char* ttos(const char* str, int len)
+{
+  int i;
+  char token_buf[MAX_TOK_LEN];
+
+  memmove(token_buf, str, len);
+  token_buf[len+1] = '\0';
+
+  return strdup(token_buf);
+}