On Wed Apr 15 09:05:13 2026 +0000, Jan Robert Gerigk wrote:
changed this line in [version 7 of the diff](/wine/wine/-/merge_requests/10604/diffs?diff_id=260404&start_sha=dcf011df1c4cec708e4c48b204f1e26bdc0e5298#e7910e8bc80d92de144443f14a9ddaae8fea4de1_1820_1821) Rebased onto upstream/master.
Dropped 36c156fe ("cfgmgr32: Read device properties stored as registry subkeys.") 3501d052 ("cfgmgr32: Fix querying unnamed properties.") already addresses the same root cause via a cleaner direct subkey lookup. Conflicts in cfgmgr32.spec / setupapi.spec resolved by adopting the upstream stub --\> cfgmgr32-forward migration and adding CM_Get_Parent_Ex on top. setupapi/stubs.c picked up upstreams removal of the moved CM\_\* stubs. Build clean on --enable-archs=i386,x86_64; cfgmgr32, setupapi/devinst (4366/0) and ntoskrnl (26430/0) tests pass. Also testet again with my USB-License-Dongle, works fine. Pre-existing test failure noted (unrelated, also fails on master): cfgmgr32.c:767 fails for real HID devices --\> wcsupr() in dlls/setupapi/devinst.c:828 uppercases the entire InstanceId including the instance-specific suffix that the test expects to remain in original case. This only fails with real HID Devices. Update for Rémi: Done. Inverted CM_Get_Parent / CM_Get_Parent_Ex and dropped the redundant TRACE. Not testet everything again after that. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10604#note_136248