Hi, > Sorry, but the patch is wrong. The Wine list implementation makes sure > that next and prev are always valid pointers, such that the list is > circular. The only way to detect the end of the list is if entry->next == > start. If one of the entries is NULL then this generally indicates either > a use after free or some other kind of corruption of the memory used by > the list. > If the memory is in the heap then HeapValidate may help find the > corruption.
If you say this, I will willingly believe it. I did some debugging searching for an exception which was caused by one of my applications, and at a certain point there are four calls of HeapAlloc(). Three of them worked great, but the forth failed cause elem->next was 0. So my idea was that its cause of it's the last entry. The other pointers looked really well. So do you think the is a heap corruption caused by my application or by some other part of Wine? > One further note: on projects as big as Wine patching the ChangeLog is not > practical as it would result in a lot of conflicts. Therefore, Alexandre > Julliard automatically generates ChangeLog entries from CVS just before > each release. Oh, I only did it cause its requested on http://www.winehq.com/site/sending_patches: "When submitting patches, please: Include a ChangeLog entry with your name, email address and a description of what your patch does." Did I understand something wrong here??? Cheers Markus Gömmel [EMAIL PROTECTED]
