From 8e8a7e0f1886275a2ac5f2b315fec048ec608e7f Mon Sep 17 00:00:00 2001
From: levlam <levlam@telegram.org>
Date: Mon, 6 Nov 2023 00:45:24 +0300
Subject: [PATCH] Support language code specification for PreCode entities in
 parseMarkdown.

---
 td/telegram/MessageEntity.cpp | 35 ++++++++++++++++++++++++++++-------
 test/message_entities.cpp     | 13 +++++++++++++
 2 files changed, 41 insertions(+), 7 deletions(-)

diff --git a/td/telegram/MessageEntity.cpp b/td/telegram/MessageEntity.cpp
index 99bd4f9e1..0e4e84a1a 100644
--- a/td/telegram/MessageEntity.cpp
+++ b/td/telegram/MessageEntity.cpp
@@ -2711,11 +2711,35 @@ static FormattedText parse_pre_entities_v3(Slice text) {
           if (end_tag_end - end_tag_begin == j - i) {
             // end tag found
             CHECK(entity_length > 0);
-            entities.emplace_back(j - i == 3 ? MessageEntity::Type::Pre : MessageEntity::Type::Code, utf16_offset,
-                                  entity_length);
-            result.append(text.begin() + j, end_tag_begin - j);
+            auto entity_begin = j;
+            string language_code;
+            if (j - i == 3) {
+              size_t language_code_end = j;
+              while (language_code_end < end_tag_begin - 1 && 33 <= text[language_code_end] &&
+                     text[language_code_end] <= 126) {
+                language_code_end++;
+              }
+              if (language_code_end < end_tag_begin - 1 && text[language_code_end] == '\n' &&
+                  (language_code_end != entity_begin || i == 0 || text[i - 1] == '\n')) {
+                language_code = text.substr(entity_begin, language_code_end - entity_begin).str();
+                entity_begin = language_code_end + 1;
+                entity_length -= entity_begin - j;
+                CHECK(entity_length > 0);
+              }
+            }
+            if (!language_code.empty()) {
+              entities.emplace_back(MessageEntity::Type::PreCode, utf16_offset, entity_length,
+                                    std::move(language_code));
+            } else {
+              entities.emplace_back(j - i == 3 ? MessageEntity::Type::Pre : MessageEntity::Type::Code, utf16_offset,
+                                    entity_length);
+            }
+            result.append(text.begin() + entity_begin, end_tag_begin - entity_begin);
             utf16_offset += entity_length;
             i = end_tag_end - 1;
+            if (end_tag_end < size && text[end_tag_end] == '\n') {
+              i++;
+            }
             is_found = true;
             break;
           } else {
@@ -2774,10 +2798,7 @@ static FormattedText parse_pre_entities_v3(Slice text, vector<MessageEntity> ent
         result_text_utf16_length += part_end - max_end;
       } else {
         FormattedText parsed_text = parse_pre_entities_v3(parsed_part_text);
-        int32 new_skipped_length = 0;
-        for (auto &entity : parsed_text.entities) {
-          new_skipped_length += (entity.type == MessageEntity::Type::Pre ? 6 : 2);
-        }
+        int32 new_skipped_length = parsed_part_text.size() - parsed_text.text.size();
         CHECK(new_skipped_length < part_end - max_end);
         result.text += parsed_text.text;
         for (auto &entity : parsed_text.entities) {
diff --git a/test/message_entities.cpp b/test/message_entities.cpp
index 0cc7bd8f9..b8267e111 100644
--- a/test/message_entities.cpp
+++ b/test/message_entities.cpp
@@ -1835,6 +1835,19 @@ TEST(MessageEntities, parse_markdown_v3) {
        {td::MessageEntity::Type::Italic, 123, 17},
        {td::MessageEntity::Type::Bold, 129, 15},
        {td::MessageEntity::Type::Spoiler, 145, 7}});
+  check_parse_markdown_v3("```\nsome code\n```", "some code\n", {{td::MessageEntity::Type::Pre, 0, 10}});
+  check_parse_markdown_v3("asd\n```\nsome code\n```cabab", "asd\nsome code\ncabab",
+                          {{td::MessageEntity::Type::Pre, 4, 10}});
+  check_parse_markdown_v3("asd\naba```\nsome code\n```cabab", "asd\naba\nsome code\ncabab",
+                          {{td::MessageEntity::Type::Pre, 7, 11}});
+  check_parse_markdown_v3("asd\naba```\nsome code\n```\ncabab", "asd\naba\nsome code\ncabab",
+                          {{td::MessageEntity::Type::Pre, 7, 11}});
+  check_parse_markdown_v3("asd\naba```a b\nsome code\n```\ncabab", "asd\nabaa b\nsome code\ncabab",
+                          {{td::MessageEntity::Type::Pre, 7, 14}});
+  check_parse_markdown_v3("asd\naba```a!@#$%^&*(b\nsome code\n```\ncabab", "asd\nabasome code\ncabab",
+                          {{td::MessageEntity::Type::PreCode, 7, 10, "a!@#$%^&*(b"}});
+  check_parse_markdown_v3("```aba\n```", "aba\n", {{td::MessageEntity::Type::Pre, 0, 4}});
+  check_parse_markdown_v3("```\n```", "\n", {{td::MessageEntity::Type::Pre, 0, 1}});
 
   td::vector<td::string> parts{"a", " #test__a", "__", "**", "~~", "||", "[", "](t.me)", "`"};
   td::vector<td::MessageEntity::Type> types{