On Thu Jun 25 07:12:53 2026 +0000, Thomas Portal wrote:
changed this line in [version 3 of the diff](/wine/wine/-/merge_requests/11204/diffs?diff_id=277821&start_sha=149057d6ab2d16e318162a77435e3cd1fbc3accb#1d6f44518922f013f44583f62c641b5e4316857e_1430_1430) Woops, you're right thanks. Reworked :
hidclass now reports the output report length on the WRITE_REPORT completion rather than forwarding the minidriver's count. The test keeps an unrelated .ret_length = 5 and still asserts length == report_len, so it checks the fix-up directly instead of cancelling a subtraction. Verified locally, hid passes with 0 failures, and reverting the hidclass change with .ret_length = 5 in place turns test_write_file red ("WriteFile wrote 5") for both the numbered and unnumbered reports. Thanks for catching that ! -- https://gitlab.winehq.org/wine/wine/-/merge_requests/11204#note_144196