On Tue Feb 21 14:32:37 2023 +0000, Hans Leidekker wrote:
I didn't base my patches on Mounir's work, I started from scratch. I see that my implementation is a bit more complete. It has configure checks and wow64 support for example. It also does a number of things differently. One is that I'm not loading libpcsclite dynamically. Because this dll is not useful without libpcsclite support we can just let it fail to load. I also think it's cleaner to use UINT64 instead of doing those preprocessor tricks with DWORD. My MR is also already split into smaller commits. I see that you have support for Apple which is not included in my own patches. You also have some good tests it seems.
I left dynamic loading just because of Apple support which was in Mounir's original version. I also added __wine_unix_call_wow64_funcs implementation, but found that actual calling of these functions does not happen. I also made sure that the current code works correctly in wow64, with the i386 version of libpcsclite1 on the system.
Since your implementation is of better quality, probably my version should be discarded. However, I suggest to take the above mentioned setup for tests based on virt_cacard. As you know, three events happen during testing at which further testing makes no sense: 1) pcscd is not running 2) no card readers are available, you can solve this by installing vsmartcard-vpcd 3) there is no card in any of the readers, solve with virt_cacard