<html>
    <head>
      <base href="https://bugzilla.gnome.org/" />
    </head>
    <body><span class="vcard"><a href="page.cgi?id=describeuser.html&login=jadahl%40gmail.com" title="Jonas Ådahl <jadahl@gmail.com>"> <span class="fn">Jonas Ådahl</span></a>
</span> changed
              <a class="bz_bug_link 
          bz_status_NEW "
   title="NEW - wayland: Don't destroy cursor backing storage when the buffer is released"
   href="https://bugzilla.gnome.org/show_bug.cgi?id=770952">bug 770952</a>
          <br>
             <table border="1" cellspacing="0" cellpadding="8">
          <tr>
            <th>What</th>
            <th>Removed</th>
            <th>Added</th>
          </tr>

         <tr>
           <td style="text-align:right;">Attachment #334911 status</td>
           <td>none
           </td>
           <td>reviewed
           </td>
         </tr></table>
      <p>
        <div>
            <b><a class="bz_bug_link 
          bz_status_NEW "
   title="NEW - wayland: Don't destroy cursor backing storage when the buffer is released"
   href="https://bugzilla.gnome.org/show_bug.cgi?id=770952#c2">Comment # 2</a>
              on <a class="bz_bug_link 
          bz_status_NEW "
   title="NEW - wayland: Don't destroy cursor backing storage when the buffer is released"
   href="https://bugzilla.gnome.org/show_bug.cgi?id=770952">bug 770952</a>
              from <span class="vcard"><a href="page.cgi?id=describeuser.html&login=jadahl%40gmail.com" title="Jonas Ådahl <jadahl@gmail.com>"> <span class="fn">Jonas Ådahl</span></a>
</span></b>
        <pre>Review of <span class=""><a href="attachment.cgi?id=334911&action=diff" name="attach_334911" title="wayland: Don't destroy cursor backing storage when the buffer is released">attachment 334911</a> <a href="attachment.cgi?id=334911&action=edit" title="wayland: Don't destroy cursor backing storage when the buffer is released">[details]</a></span> <a href='review?bug=770952&attachment=334911'>[review]</a>:

I don't think this is the correct solution. Before this patch you could set the
cursor and get rid of your cairo_surface_t reference without any unexpected
side effects, since the wl_surface implicitly had a reference to the cairo
surface via the buffer release callback. With this applied, would we destroy
the cairo surface would the wl_buffer also be destroyed (see
gdk_wayland_cairo_surface_destroy()).

You write that the issue is a refcount underrun, did you investigate how the
ref count went wrong?

For reference, each time one wl_surface.attach(buffer); wl_surface.commit(), we
should receive one wl_buffer.release on the attached buffer. Do we
cairo_surface_reference() for each time we attach and commit?</pre>
        </div>
      </p>
      <hr>
      <span>You are receiving this mail because:</span>
      
      <ul>
          <li>You are on the CC list for the bug.</li>
      </ul>
    </body>
</html>