Am 04.06.2007 um 23:35 schrieb Jean-Marc Lasgouttes:
"Stefan" == Stefan Schimanski <[EMAIL PROTECTED]> writes:Stefan> Hmm, it's a not very complex patch for a critical bug about a Stefan> segfault and nobody has a comment? Actually I am wondering why suddenly we have these crashes in very simple operations (like breakParagraph this one). Isn't there something else hidden?
This one has nothing to do with breakParagraph. It is IMO just non- defensive programming which strikes back. The InsetMathNest just assumes that all its cells are drawn (and hence are in the coordinate cache). But this assumption shouldn't be made. If it is made, we have to document it and check the derived classes. I am pretty sure that someof them warm up the cache of all cells as a consequence of patchworking against crashes in other InsetMathMethods method (e.g. editXY).
Stefan
PGP.sig
Description: Signierter Teil der Nachricht
