On Fri May 1 00:50:51 2026 +0000, eric pouech wrote:
focusing first on flags=**CERT_PE_IMAGE_DIGEST_ALL_IMPORT_INFO** (which is IMO the "simplest" to implement) * native fails to compute digest under certain cases * when NumberOfRVaAndSizes is 0 * on some other images, for which I couldn't find a simple trigger; need to write some tooling for that; but IMO we can leave it for future work (or at least is orthogonal to next point, which is the most annoying) * on some images the basereloc directory is zeroed by the MR while native doesn't do it... (sometimes only when reporting in the callback the embedding section, or in some other cases for both the NT headers and the embeddeding section); from the output, it looks like the tests for zeroing the fields in basereloc miss some cases * uninformatively from cursory couldn't find a simple trigger pattern * if you want to dig further, here's a winedump -f of one of the offending image [p](/uploads/f232953407e5063c7583c1854db62d2d/p) Can I get the binary? Or where you got it, if I can go there to get it.
-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10295#note_138486