From f6c588ee7e0694a0f630e3e362f751ac0a6d91ae Mon Sep 17 00:00:00 2001
From: xiaoyifang <yifang.xiao@gmail.com>
Date: Wed, 20 Apr 2022 20:53:57 +0800
Subject: [PATCH] fix: collapse function does not work expectedly.

the website dictionaries do not collapse as it will always treated as 0 .
the normal dictionaries html size is not correct calculated due to a qt bug https://bugreports.qt.io/browse/QTBUG-102757
---
 article_maker.cc | 20 +++++++++++++++++++-
 article_maker.hh |  2 +-
 2 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/article_maker.cc b/article_maker.cc
index db56ca2b..e18b63fb 100644
--- a/article_maker.cc
+++ b/article_maker.cc
@@ -596,7 +596,7 @@ void ArticleRequest::bodyFinished()
               }
             }
 
-            int size = QTextDocumentFragment::fromHtml( text ).toPlainText().length();
+            int size = htmlTextSize( text );
             if( size > articleSizeLimit )
               collapse = true;
           }
@@ -750,6 +750,24 @@ void ArticleRequest::bodyFinished()
   }
 }
 
+int ArticleRequest::htmlTextSize( QString html )
+{
+  // website dictionary.
+  if( html.contains( QRegularExpression( "<iframe\\s*[^>]*>", QRegularExpression::CaseInsensitiveOption ) ) )
+  {
+    //arbitary number;
+    return 1000;
+  }
+
+  //https://bugreports.qt.io/browse/QTBUG-102757
+  QString stripStyleSheet =
+    html.remove( QRegularExpression( "<link\\s*[^>]*>", QRegularExpression::CaseInsensitiveOption ) )
+      .remove( QRegularExpression( "<script[\\s\\S]*?>[\\s\\S]*?<\\/script>", QRegularExpression::CaseInsensitiveOption|QRegularExpression::MultilineOption ) );
+  int size = QTextDocumentFragment::fromHtml( stripStyleSheet ).toPlainText().length();
+
+  return size;
+}
+
 void ArticleRequest::stemmedSearchFinished()
 {
   // Got stemmed matching results
diff --git a/article_maker.hh b/article_maker.hh
index c3a017d4..665515f3 100644
--- a/article_maker.hh
+++ b/article_maker.hh
@@ -148,7 +148,7 @@ private slots:
   void individualWordFinished();
 
 private:
-
+  int htmlTextSize( QString html );
   /// Appends the given string to 'data', with locking its mutex.
   void appendToData( std::string const & );