<html> <head> <base href="https://bugzilla.gnome.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - memory leak" href="https://bugzilla.gnome.org/show_bug.cgi?id=761312#c9">Comment # 9</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - memory leak" href="https://bugzilla.gnome.org/show_bug.cgi?id=761312">bug 761312</a> from <a href="page.cgi?id=describeuser.html&login=rstrode%40redhat.com" title="Ray Strode [halfline] <rstrode@redhat.com>"> Ray Strode [halfline]</a> <pre>Review of <a href="attachment.cgi?id=320153&action=diff" name="attach_320153" title="wayland: stage uncommited changes to dedicated buffer">attachment 320153</a> <a href="attachment.cgi?id=320153&action=edit" title="wayland: stage uncommited changes to dedicated buffer">[details]</a> <a href='review?bug=761312&attachment=320153'>[review]</a>: ::: gdk/wayland/gdkwindow-wayland.c @@ +685,3 @@ + cairo_set_operator (cr, CAIRO_OPERATOR_SOURCE); + cairo_paint (cr); + cairo_destroy (cr); So this is costly. It won't happen much in practice now that mutter releases the buffer before the frame callback, but we could optimize this by avoiding the copy up front, only doing incremental changes into the staging buffer, and then compositing the read back buffer on top of the incremental changes using the DEST_OUT operator. Another idea would be to use memcpy instead of cairo to do the copy. We know the buffers are linear with width == rowstride, so we can get by with one memcpy call. It could be that cairo/pixman already fall back to a memcpy in this situation though.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are on the CC list for the bug.</li> </ul> </body> </html>