XPost: alt.comp.os.windows-11, alt.comp.microsoft.windows   
   From: mariasophia@comprehension.com   
      
   Maria Sophia wrote:   
   > Since I covered the topic at the level of a published paper, there's really   
   > not much more left to discuss, as far as I can tell. It's all fixed now.   
      
   All that's left after fixing the issue was understanding what actually   
   went wrong in the first place (which killed the control+A in Notepad++).   
      
   It turns out that Windows does not convert the HTML Format entry into text   
   until an application explicitly asks for a text format.   
      
   So the corruption happens at the moment Notepad++ requests CF_UNICODETEXT.   
      
   The sequence (as far as I can re-construct it) is...   
      
   1. With Ctrl+C, Chromium places several formats on the clipboard,   
      including HTML Format, CF_UNICODETEXT, and its internal metadata.   
      
   2. With Ctrl+V, Notepad++ asks Windows:   
         "Give me CF_UNICODETEXT."   
      
   3. Windows sees that HTML Format is available and may choose to generate   
      the CF_UNICODETEXT stream by converting the HTML fragment.   
      
      Kaboom!   
      
   4. That conversion step can produce a corrupted CF_UNICODETEXT stream.   
      The corruption is not visible text. Which is why I couldn't "see" it.   
      It is a bad length field or a hidden control character (apparently).   
      
      Is that a bug?   
      I don't know.   
      
   5. Scintilla loads that corrupted stream into its internal buffer.   
      But the buffer boundaries are now wrong, so Ctrl+A fails because   
      Scintilla thinks the document ends earlier than it actually does.   
      
   So why didn't I see it in the Notepad++ hex editor?   
      
   The HTML is never pasted into the file, so it can't be seen.   
   But it affects the text Windows hands to Notepad++ at paste time.   
      
   Well then, why does adding and deleting a character fix it?   
      
   Because the corruption lives only in Scintilla's internal buffer   
   structures, not in the visible text. When the macro inserts a space,   
   Scintilla is forced to rebuild its entire buffer. That rebuild wipes out   
   the corrupted boundary. Removing the space forces a second rebuild,   
   which simply restores the original content. The second rebuild is not   
   needed for the fix; it is only needed to undo the temporary change.   
      
   After that, the macro selects all and cuts the text. Cutting forces   
   Windows to create a brand new clipboard entry. This new clipboard entry   
   contains only plain text formats, because Scintilla does not generate   
   HTML Format or any Chromium internal formats.   
      
   I don't know if this is a bug or not, as all I know, in the end, is...   
   1. The corrupted CF_UNICODETEXT stream from the original paste is gone.   
   2. The clipboard now contains only clean plain text.   
   3. Scintilla now has a clean buffer with correct boundaries.   
   4. Ctrl+A works again.   
      
   Woo hoo!   
      
   So the fix works because Windows created the problem when converting the   
   HTML fragment into text, which corrupted Scintilla's internal buffer.   
   Adding and removing a character forces Scintilla to rebuild its buffer,   
   and cutting the text forces Windows to rebuild the clipboard without   
   HTML Format. The corruption cannot survive those two rebuilds.   
      
   I think we explained it as simply as we could, but not simpler.   
   --   
         How wonderful that we have met with a paradox.   
           Now we have some hope of making progress.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)