http://bugs.winehq.org/show_bug.cgi?id=8275
------- Additional Comments From juan_lang@yahoo.com 2007-04-05 15:49 ------- It looks as though the second log still uses builtin crypt32, else we wouldn't see all the trace:crypt:CryptHashData calls. So the bug must be in rsaenh somewhere.
It looks superficially as though CryptAcquireContext might be failing, since when using builtin rsaenh, the HCRYPTPROV argument to CryptCreateHash et al is NULL. But the builtin rsaenh returns an index into a handle table rather than a pointer, so the first returned handle is always 0. So the bug appears to be elsewhere.