On Sat Mar 28 18:12:47 2026 +0000, Nikolay Sivov wrote:
Again, even if the tests work, you release capture only to capture same window again. The fix as it is now does not make sense to me. You asked:
_The real question is whether WM_CAPTURECHANGED should happen at all in this case of repeated WM_LBUTTONDOWN?_ Yes, test case shows that on repeated WM_LBUTTONDOWN happens WM_CAPTURECHANGED. Since it is happening, we have to do ReleaseCapture to be able to then process WM_LBUTTONUP properly. According to the tests, windows has same behavior, releases capture (and sending WM_CAPTURECHANGED) only to capture same window again. Are these tests not enough? Even if it acts identically as windows? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10104#note_134233