The basic loop correctly finds the first element, however
the second element repeats the first node, since it was
incremented and the ns->next was skipped over.
--
v2: msxml3: Correct looping of Document Element node map.
https://gitlab.winehq.org/wine/wine/-/merge_requests/6993
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=56940
Attempt to fix the regression caused by 0bad544aab9e2c9ee93bbabac0386e02c58a39c0.
Unfortunately the author of that patch doesn't seem to be active anymore, so here's an attempt, with the remark that I don't know why it worked before that patch, and I'm not sure sure if this is the correct way to fix this.
The only thing I know for sure is that on windows one can start an app via ShellExecuteW with a trailing space, and in wine this doesn't work anymore, and it did work before that commit.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/6997
The load configuration is mandatory for ARM64EC binaries, and the MSVC linker emits a warning when it is missing. It makes sense, so I plan to add a similar warning to lld-link. Since the configure scripts treat linker warnings as errors, defining the load configuration ensures future compatibility.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/6992
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=57285
This actually fixes the underlying issue, which is also present with undecorated windows / virtual desktop mode.
The window manager problem is a different issue, more likely a window manager bug that adds to the soup, and the other fix also makes this work probably because it ends up reset the X window shape somehow.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/6990
* winebth.sys: Use the correct byte-ordering for setting the radio's address property.
* winebth.sys: Set the Information field in the IRP's STATUS_BLOCK after handling IOCTL_BTH_GET_LOCAL_INFO.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/6988
Instead of checking the winstation name and registry on every lock_display_devices call, which is now called on every mouse input to get monitor dpi. This would perhaps be better implemented with a shared memory winstation object instead of putting the flag on the desktop objects, but it would be a larger and riskier change.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=57504
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/6987