From cd4d2e98c34ed413ca164dbad61ba19636377f77 Mon Sep 17 00:00:00 2001
From: dec05eba <dec05eba@protonmail.com>
Date: Mon, 19 Jul 2021 19:28:45 +0200
Subject: Add hotexamples

---
 src/NetUtils.cpp | 39 +++++++++++++++++++++++++++++++++++----
 1 file changed, 35 insertions(+), 4 deletions(-)

(limited to 'src/NetUtils.cpp')

diff --git a/src/NetUtils.cpp b/src/NetUtils.cpp
index cc19094..28256cb 100644
--- a/src/NetUtils.cpp
+++ b/src/NetUtils.cpp
@@ -34,12 +34,43 @@ namespace QuickMedia {
         std::string unescaped_str;
     };
 
+    static bool to_num(const char *str, size_t size, int &num) {
+        num = 0;
+        for(size_t i = 0; i < size; ++i) {
+            const char num_c = str[i] - '0';
+            if(num_c < 0 || num_c > 9)
+                return false;
+            num = (num * 10) + num_c;
+        }
+        return true;
+    }
+
+    static void html_unescape_sequence_numbers(std::string &str) {
+        size_t index = 0;
+        while(true) {
+            index = str.find("&#", index);
+            if(index == std::string::npos)
+                break;
+            
+            index += 2;
+            size_t end_index = str.find(';', index);
+            if(end_index != std::string::npos && end_index - index <= 3) {
+                const size_t num_length = end_index - index;
+                int num;
+                if(to_num(str.c_str() + index, num_length, num)) {
+                    const char num_c = (char)num;
+                    str.replace(index - 2, 2 + num_length + 1, &num_c, 1);
+                    index += (-2 + 1);
+                }
+            }
+        }
+    }
+
     void html_unescape_sequences(std::string &str) {
-        const std::array<HtmlUnescapeSequence, 7> unescape_sequences = {
+        html_unescape_sequence_numbers(str);
+
+        const std::array<HtmlUnescapeSequence, 4> unescape_sequences = {
             HtmlUnescapeSequence { "&quot;", "\"" },
-            HtmlUnescapeSequence { "&#039;", "'" },
-            HtmlUnescapeSequence { "&#39;", "'" },
-            HtmlUnescapeSequence { "&#10;", "\n" },
             HtmlUnescapeSequence { "&lt;", "<" },
             HtmlUnescapeSequence { "&gt;", ">" },
             HtmlUnescapeSequence { "&amp;", "&" } // This should be last, to not accidentally replace a new sequence caused by replacing this
-- 
cgit v1.2.3