On Wed Nov 2 01:06:36 2022 +0000, Mohamad Al-Jaf wrote:
Yes, you're right. Those changes were made in order to conform to the behavior of Windows and pass the tests, specifically the last test which checks if ppidlLast is NULL after the function returns E_INVALIDARG. Though, this only applied to SHBindToFolderIDListParent. There was no test for SHBindToParent for this behavior. I've added one now. I'm still not really sure what Alexandre meant by merging the two code change patches. Merging the commits just made it unclear IMO. I've added a separate commit for reimplementing SHBindToParent as well as a separate test commit. Please let me know if this is okay. I can always just start a new merge request for reimplementing SHBindToParent if necessary.
The point is that merging the two changes makes it clear how SHBindToFolderIDListParent() differs from SHBindToParent(). I.e. in this case, the implementation is the same, except that you're using a custom shell folder instead of the desktop. Changing around the error handling muddies that, and if it's necessary then it should be its own patch.