On Wed Aug 2 06:22:03 2023 +0000, Jactry Zeng wrote:
Is "file://foo/%E4%AD%E6%96%87/bar" the case: https://gitlab.winehq.org/wine/wine/-/merge_requests/585/diffs?commit_id=d48b42bbb0750497dbddc972e7287c30feaccd4d#39c88e69f9ddfbd37ae3565e18bb4723cebb11cb_422_444? MultiByteToWideChar() will convert the whole corrupt UTF-8 character to 0xfffd.
Ah, I think I see: "file://foo/%E4%AD%E6%96%87/bar" is the case, but lstrcmpW() didn't do a good job in this comparison, using wcscmp() instead will expose the issue.