Hi, would anyone disagree about killing exception support? The exception support we have in gmem is very limited and we actually we do not support it correctly since in the poppler core code we have some gmalloc_checkoverlow that expect NULL on failure, not an exception. Comments? Albert