https://bugs.winehq.org/show_bug.cgi?id=45116 Hanno Meyer-Thurow <h.mth(a)web.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #61313|0 |1 is obsolete| | --- Comment #10 from Hanno Meyer-Thurow <h.mth(a)web.de> --- Created attachment 61331 --> https://bugs.winehq.org/attachment.cgi?id=61331 Implementation of NotifyAddrChange: build fixes for 3.7 (v3) Added 'struct ip_async_io' to handle callback with user data. Taken from fileio code. Minor fixes like releasing non-acquired fd in IPHLPAPI_initmonitorhandle and properly releasing ip_async_io instance in IPHLPAPI_apc_NotifyAddrChange. With this variant I see call to blocking NotifyAddrChange(PHANDLE handle=nil, LPOVERLAPPED overlapped=nil) misbehaving after IPHLPAPI_apc_NotifyAddrChange processed once successfully (with releasing of ip_async_io instance). IPHLPAPI_monitorifchange returns success always for whatever reason. In return user application is "hammering" NotifyAddrChange. Once out-of-fds it crashes between fd=255 fd=256 with null pointer exception. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.