DescriptionI have some books which use an on-the-fly markup procedure that refers to a context in a closure. I now question the wisdom of this, but that's what I have. In certain circumstances (when processing a certain book with hundreds of scores, but not other books with tens of scores), one or more of the ancestors of the important context are destroyed. When the time finally comes to evaluate the markup, the first thing the on-the-fly procedure does is get a context property, which walks up the now-invalid hierarchy until a segfault occurs.
Nullifying the child's pointer in the parent's destructor was the obvious way to fix this. Is there a better way?
Patch Set 1 #
MessagesTotal messages: 1
|