feat(lexer): header, line breaks, paragraph breaks

include/mrk/lexer.h

@@ -14,7 +14,11 @@ typedef enum mrk_lexer_err {
 } mrk_lexer_err;
 
 typedef enum mrk_token_type {
-  mrk_token_type_pound = 0,
+  mrk_token_type_header = 0,
+  mrk_token_type_blank_line,
+  mrk_token_type_star_star,
+  mrk_token_type_space_space,
+  mrk_token_type_line_break,
 } mrk_token_type;
 
 typedef struct mrk_token {
@@ -28,6 +32,11 @@ typedef struct mrk_token {
  */
 mrk_err mrk_lexer_init(mrk_lexer **out);
 
+/**
+ * Deallocate the lexer object.
+ */
+void mrk_lexer_free(mrk_lexer *lexer);
+
 /**
  * Open the buffer with the given lexer struct. `buf` is expected to live for
  * the duration of the lexing.
@@ -41,7 +50,7 @@ void mrk_lexer_open(mrk_lexer *lexer, const char *buf, size_t len);
 /**
  * Returns whether the lexer is done.
  */
-bool mrk_lexer_at_end(const mrk_lexer *lexer);
+bool mrk_lexer_done(const mrk_lexer *lexer);
 
 /**
  * Output the next lexed token for the given input.

src/_include/mrk/lexer_internal.h

@@ -16,6 +16,7 @@ struct mrk_lexer {
   struct {
     size_t start;
     size_t end;
+    bool emitted;
   } token;
 };
 
@@ -25,12 +26,29 @@ struct mrk_lexer {
  */
 char mrk_lexer_peek(mrk_lexer *lexer);
 
+/**
+ * Returns true if the nul-terminated string s is equal to the next characters
+ * in the token stream.
+ */
+bool mrk_lexer_match(mrk_lexer *lexer, const char *s);
+
 /**
  * Advance the current position by one character, adding the new character to
  * the curent token's context and returning it.
  */
 char mrk_lexer_advance(mrk_lexer *lexer);
 
+/**
+ * Advance `n` positions; equivalent to running advance `n` times and returning
+ * the last call's result.
+ */
+char mrk_lexer_advance_n(mrk_lexer *lexer, size_t n);
+
+/**
+ * Reset the lexer's current token context.
+ */
+void mrk_lexer_reset(mrk_lexer *lexer);
+
 /**
  * Output the currently matched token to the token struct with the given type,
  * and reset the lexer's tracked token.

src/lexer/lexer.c

@@ -1,3 +1,4 @@
+#include "mrk/lexer.h"
 #include "mrk/lexer_internal.h"
 
 mrk_err mrk_lexer_init(mrk_lexer **out) {
@@ -6,6 +7,14 @@ mrk_err mrk_lexer_init(mrk_lexer **out) {
   return mrk_err_ok;
 }
 
+void mrk_lexer_free(mrk_lexer *lexer) {
+  if (lexer == NULL) {
+    return;
+  }
+
+  free(lexer);
+}
+
 void mrk_lexer_open(mrk_lexer *lexer, const char *buf, size_t len) {
   lexer->buf.s = buf;
   lexer->buf.len = len;
@@ -13,15 +22,16 @@ void mrk_lexer_open(mrk_lexer *lexer, const char *buf, size_t len) {
   lexer->pos.buf_index = 0;
   lexer->token.start = 0;
   lexer->token.end = 0;
+  lexer->token.emitted = false;
 }
 
-bool mrk_lexer_at_end(const mrk_lexer *lexer) {
+bool mrk_lexer_done(const mrk_lexer *lexer) {
   return (lexer->buf.len > 0 && lexer->pos.buf_index == lexer->buf.len) ||
          (lexer->buf.s[lexer->pos.buf_index] == '\0');
 }
 
 char mrk_lexer_advance(mrk_lexer *lexer) {
-  if (mrk_lexer_at_end(lexer)) {
+  if (mrk_lexer_done(lexer)) {
     return '\0';
   }
 
@@ -42,32 +52,90 @@ char mrk_lexer_advance(mrk_lexer *lexer) {
   return c;
 }
 
+char mrk_lexer_advance_n(mrk_lexer *lexer, size_t n) {
+  while (n > 1) {
+    mrk_lexer_advance(lexer);
+  }
+
+  return mrk_lexer_advance(lexer);
+}
+
 char mrk_lexer_peek(mrk_lexer *lexer) {
-  if (mrk_lexer_at_end(lexer)) {
+  if (mrk_lexer_done(lexer)) {
     return '\0';
   }
 
   return lexer->buf.s[lexer->pos.buf_index];
 }
 
+void mrk_lexer_reset(mrk_lexer *lexer) {
+  lexer->token.start = lexer->pos.buf_index;
+  lexer->token.end = lexer->pos.buf_index;
+  lexer->token.emitted = false;
+}
+
 void mrk_lexer_emit(mrk_token *out, mrk_lexer *lexer, mrk_token_type type) {
   out->type = type;
   out->start = lexer->token.start;
-  out->start = lexer->token.end;
+  out->end = lexer->token.end;
 
-  lexer->token.start = lexer->token.end;
+  lexer->token.emitted = true;
 }
 
 mrk_lexer_err mrk_lexer_next(mrk_token *out, mrk_lexer *lexer) {
-  if (mrk_lexer_at_end(lexer)) {
+  if (mrk_lexer_done(lexer)) {
     return mrk_lexer_err_done;
   }
 
-  switch (mrk_lexer_advance(lexer)) {
-  case '#':
-    mrk_lexer_emit(out, lexer, mrk_token_type_pound);
-    break;
+  mrk_lexer_reset(lexer);
+
+  while (!lexer->token.emitted && !mrk_lexer_done(lexer)) {
+    switch (mrk_lexer_advance(lexer)) {
+    // Match one or more hashtags as a single header definition
+    case '#':
+      while (mrk_lexer_peek(lexer) == '#') {
+        mrk_lexer_advance(lexer);
+      }
+
+      mrk_lexer_emit(out, lexer, mrk_token_type_header);
+      break;
+      // Two consecutive newlines constitute a blank line, otherwise they're
+      // ignored as whitespace
+    case '\n':
+      if (mrk_lexer_peek(lexer) == '\n') {
+        mrk_lexer_advance(lexer);
+        mrk_lexer_emit(out, lexer, mrk_token_type_blank_line);
+      } else {
+        mrk_lexer_reset(lexer);
+      }
+      break;
+    case ' ': {
+      // Either a double space or a line break
+      if (mrk_lexer_peek(lexer) == ' ') {
+        mrk_lexer_advance(lexer);
+
+        if (mrk_lexer_peek(lexer) == '\n') {
+          mrk_lexer_advance(lexer);
+          mrk_lexer_emit(out, lexer, mrk_token_type_line_break);
+        } else {
+          mrk_lexer_emit(out, lexer, mrk_token_type_space_space);
+        }
+      } else {
+        mrk_lexer_reset(lexer);
+      }
+    } break;
+      /* case '*': */
+      /*   if (mrk_lexer_peek(lexer) == '*') { */
+      /*     mrk_lexer_advance(lexer); */
+      /*     mrk_lexer_emit(out, lexer, mrk_token_type_star_star); */
+      /*   } else { */
+      /*     // TODO match word */
+      /*   } */
+      /* default: */
+      /*   return mrk_lexer_err_unexpected_char; */
+    }
   }
 
-  return mrk_lexer_err_ok;
+  return lexer->token.emitted ? mrk_lexer_err_ok : mrk_lexer_err_done;
+  ;
 }

test/lexer/lexer.c

@@ -0,0 +1,27 @@
+#include "test.h"
+
+#include "mrk/lexer.h"
+
+#define LEXER_INIT() \
+  mrk_lexer *lxr; \
+  TEST_CHECK(mrk_lexer_init(&lxr) == mrk_err_ok)
+
+void test_lexer_header() {
+  LEXER_INIT();
+
+  const char *buf = "#### hallo";
+  mrk_lexer_open(lxr, buf, 0);
+
+  mrk_token t;
+  TEST_CHECK(mrk_lexer_next(&t, lxr) == mrk_lexer_err_ok);
+  TEST_CHECK(t.type == mrk_token_type_header);
+  TEST_CHECK_(t.start == 0, "t.start == %lu", t.start);
+  TEST_CHECK(t.end == 4);
+
+  mrk_lexer_free(lxr);
+}
+
+TEST_LIST = {
+  { "lexer header", test_lexer_header },
+  { NULL, NULL }
+};