2 files changed, 177 insertions, 0 deletions

diff --git a/include/cmmm/rs274ngc/lexer.h b/include/cmmm/rs274ngc/lexer.h
new file mode 100644
index 0000000..eca972e
--- /dev/null
+++ b/include/cmmm/rs274ngc/lexer.h
@@ -0,0 +1,35 @@
+// include/cmmm/lexer.h
+
+#ifndef INCLUDED__RS274NGC__LEXER__H
+#define INCLUDED__RS274NGC__LEXER__H
+
+// #define CMMM__RS274NGC__LEXER__STRIP_VENDOR
+
+#include <stddef.h>
+#include "cmmm/string-view.h"
+#include "cmmm/rs274ngc/lexeme.h"
+
+#define RS274NGC__LEXER__FIRST_MATCH
+#define RS274NGC__LEXER__TRIM
+// #define RS274NGC__LEXER__FILTER_COMMENTS
+
+struct cmmm__rs274ngc__lexer {
+  struct cmmm__string_view sv;
+  const char *cursor;
+  const char *filename;
+  struct cmmm__rs274ngc__lexeme lexeme;
+};
+
+struct cmmm__rs274ngc__lexer cmmm__rs274ngc__lexer__from_string_view (struct cmmm__string_view, const char *filename);
+void                         cmmm__rs274ngc__lexer__next_lexeme (struct cmmm__rs274ngc__lexer *);
+
+#ifdef CMMM__RS274NGC__LEXER__STRIP_VENDOR
+#define RS274NGC__LEXER__STRIP_VENDOR CMMM__RS274NGC__LEXER__STRIP_VENDOR
+#define rs274ngc__lexer cmmm__rs274ngc__lexer
+#define rs274ngc__lexer__from_string_view cmmm__rs274ngc__lexer__from_string_view
+#define rs274ngc__lexer__next_lexeme      cmmm__rs274ngc__lexer__next_lexeme
+#endif // CMMM__RS274NGC__LEXER__STRIP_VENDOR
+
+#endif // INCLUDED__RS274NGC__LEXER__H
+
+// include/cmmm/lexer.h ends here
diff --git a/src/lexer.c b/src/lexer.c
new file mode 100644
index 0000000..e645bee
--- /dev/null
+++ b/src/lexer.c
@@ -0,0 +1,142 @@
+// src/lexer.c
+
+#define CMMM__RS274NGC__LEXEME__STRIP_VENDOR
+#define CMMM__RS274NGC__LEXER__STRIP_VENDOR
+#include "cmmm/rs274ngc/lexer.h"
+#include <regex.h>
+#include <assert.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <ctype.h>
+
+#define X(SYM, REG, ...)                                                       \
+  [RS274NGC__LEXEME__KIND__##SYM - RS274NGC__LEXEME__REGEX_OFFSET] = "^(" REG ")",
+static const char *lexeme_regex_strings[RS274NGC__LEXEME__REGEX_COUNT] = {
+  RS274NGC__LEXEME__KINDS ()
+};
+#undef X
+
+static regex_t lexeme_regexes[RS274NGC__LEXEME__REGEX_COUNT];
+
+static void
+print_regex_error (const regex_t *comp, const char *regex, int errcode)
+{
+  size_t length = regerror (errcode, comp, NULL, 0);
+  char *buffer = malloc (length);
+  regerror (errcode, comp, buffer, length);
+  fprintf (stderr, "error: regcomp (\"%s\"): %s\n", regex, buffer);
+  free (buffer);
+}
+
+static void __attribute__((constructor))
+compile_regexes (void)
+{
+  bool errored = false;
+
+  for (int i = 0; i < RS274NGC__LEXEME__REGEX_COUNT; ++i) {
+    regex_t *compiled = &lexeme_regexes[i];
+    const char *pattern = lexeme_regex_strings[i];
+    int errcode = regcomp (compiled, pattern, REG_EXTENDED);
+    if (errcode) {
+      print_regex_error (compiled, pattern, errcode);
+      errored = true;
+    }
+  }
+
+  if (errored)
+    exit (EXIT_FAILURE);
+}
+
+struct rs274ngc__lexer
+rs274ngc__lexer__from_string_view (struct cmmm__string_view sv,
+                                   const char *filename)
+{
+  struct rs274ngc__lexer lexer = {
+    .sv = sv,
+    .cursor = sv.begin,
+    .filename = filename,
+  };
+
+  rs274ngc__lexer__next_lexeme (&lexer);
+  return lexer;
+}
+
+static void
+skip_whitespace (struct rs274ngc__lexer *lexer)
+{
+  while (lexer->cursor < lexer->sv.end && (*lexer->cursor == ' ' ||
+                                           *lexer->cursor == '\t'))
+    lexer->cursor += 1;
+}
+
+void
+rs274ngc__lexer__next_lexeme (struct rs274ngc__lexer *lexer)
+{
+  skip_whitespace (lexer);
+  lexer->lexeme = (struct rs274ngc__lexeme) { .sv.begin = lexer->cursor };
+
+  if (lexer->cursor >= lexer->sv.end) {
+    lexer->lexeme.kind = RS274NGC__LEXEME__KIND__END_OF_FILE;
+    lexer->lexeme.sv.end = lexer->sv.end;
+    return;
+  }
+
+  if (*lexer->cursor == '\r') {
+    if (*lexer->cursor == '\n')
+      lexer->cursor += 1;
+
+    lexer->lexeme.kind = RS274NGC__LEXEME__KIND__END_OF_LINE;
+    lexer->lexeme.sv.end = ++lexer->cursor;
+    return;
+  }
+
+  if (*lexer->cursor == '\n') {
+    lexer->lexeme.kind = RS274NGC__LEXEME__KIND__END_OF_LINE;
+    lexer->lexeme.sv.end = ++lexer->cursor;
+    return;
+  }
+
+  regmatch_t regmatch;
+  int longest_matching_regex = RS274NGC__LEXEME__REGEX_COUNT;
+  int longest_matching_regex_length = 0;
+  for (int i = 0; i < RS274NGC__LEXEME__REGEX_COUNT; ++i) {
+    if (regexec (&lexeme_regexes[i], lexer->cursor, 1, &regmatch, 0)) continue;
+    if (regmatch.rm_eo <= longest_matching_regex_length) continue;
+
+    longest_matching_regex_length = regmatch.rm_eo;
+    longest_matching_regex = i;
+
+    // HACK(cmmm): choose first match
+#ifdef RS274NGC__LEXER__FIRST_MATCH
+    break;
+#endif
+  }
+
+  if (longest_matching_regex != RS274NGC__LEXEME__REGEX_COUNT) {
+    lexer->lexeme.kind = longest_matching_regex + RS274NGC__LEXEME__REGEX_OFFSET;
+    lexer->lexeme.sv.end = lexer->cursor += longest_matching_regex_length;
+
+    // HACK(cmmm): right trim lexemes
+#ifdef RS274NGC__LEXER__TRIM
+    while (lexer->lexeme.sv.end > lexer->lexeme.sv.begin &&
+           (lexer->lexeme.sv.end[-1] == ' ' ||
+            lexer->lexeme.sv.end[-1] == '\t')) {
+      lexer->lexeme.sv.end -= 1;
+    }
+#endif
+
+    switch (lexer->lexeme.kind) {
+#ifdef RS274NGC__LEXER__FILTER_COMMENTS
+    case RS274NGC__LEXEME__KIND__COMMENT:
+      return rs274ngc__lexer__next_lexeme (lexer);
+#endif
+    default: return;
+    }
+  }
+
+  // INVALID
+  lexer->lexeme.sv.end = ++lexer->cursor;
+}
+
+// src/lexer.c ends here