From 29e03ef74a814cb31a0ae53192e25cc75b638256 Mon Sep 17 00:00:00 2001
From: mattkae <mattkae@protonmail.com>
Date: Thu, 22 Jun 2023 13:48:11 -0400
Subject: Handling less than signs in plain text

---
 src/tokenizer.cpp | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

(limited to 'src/tokenizer.cpp')

diff --git a/src/tokenizer.cpp b/src/tokenizer.cpp
index 9931d59..dc0b8d7 100644
--- a/src/tokenizer.cpp
+++ b/src/tokenizer.cpp
@@ -28,6 +28,12 @@ inline void unconsume_previous(Tokenizer* tokenizer) {
     tokenizer->ptr--;
 }
 
+inline void emit_character(Tokenizer* tokenizer, code_point_t c) {
+    tokenizer->last.type = HtmlTokenType_Character;
+    tokenizer->last.character_token = c;
+    tokenizer->flag |= TokenizerFlag_Emit;
+}
+
 
 /// https://dev.w3.org/html5/spec-LC/tokenization.html#data-state
 inline void data_state(Tokenizer* tokenizer) {
@@ -45,9 +51,7 @@ inline void data_state(Tokenizer* tokenizer) {
         break;
     default:
         // TODO: @Error If null, throw an error
-        tokenizer->last.type = HtmlTokenType_Character;
-        tokenizer->last.character_token = *tokenizer->ptr;
-        tokenizer->flag |= TokenizerFlag_Emit;
+        emit_character(tokenizer, *tokenizer->ptr);
         break;
     }
 }
@@ -80,8 +84,11 @@ inline void tag_open_state(Tokenizer* tokenizer) {
             tokenizer->state = TokenizerState_BogusComment; // TODO:
         }
         else {
+            emit_character(tokenizer, CodePoints::LESS_THAN_SIGN);
             tokenizer->state = TokenizerState_Data;
+            tokenizer->flag = tokenizer->flag | TokenizerFlag_NoIncrement;
         }
+        break;
     }
 }
 
@@ -114,7 +121,7 @@ inline void tag_name_state(Tokenizer* tokenizer) {
         else if (c == EOF) {
             // TODO: @Error
             tokenizer->state = TokenizerState_Data;
-            tokenizer->flag = tokenizer->flag & TokenizerFlag_DecrementPtr;
+            tokenizer->flag = tokenizer->flag | TokenizerFlag_NoIncrement;
         }
         else {
             tokenizer->last.append_to_tag_name(c);
@@ -151,7 +158,7 @@ inline void end_tag_open_state(Tokenizer* tokenizer) {
         tokenizer->last.type = HtmlTokenType_Character;
         tokenizer->last.character_token = CodePoints::SOLIDUS;
         
-        tokenizer->flag = tokenizer->flag & TokenizerFlag_DecrementPtr;
+        tokenizer->flag = tokenizer->flag | TokenizerFlag_NoIncrement;
     }
     else if (c == CodePoints::NULL_CHAR) {
         // TODO: @Error
@@ -286,7 +293,7 @@ HtmlToken read_next(Tokenizer* tokenizer) {
             exit(1);
         }
         
-        if ((tokenizer->flag & TokenizerFlag_DecrementPtr) == 0) {
+        if ((tokenizer->flag & TokenizerFlag_NoIncrement) == 0) {
             tokenizer->ptr++;
         }
 
-- 
cgit v1.2.1