Changing Net User things to use ffi instead of win32-api #3344
Conversation
| else | ||
| val.to_wstring | ||
| end | ||
| val = FFI::MemoryPointer.from_string(encoded) |
There was a problem hiding this comment.
If you're setting val = directly here, you don't need the val = on line 88 or the else block below.
|
👍 Loading a library to fetch error messages is too insane for this pull-request. Leaving that for another time. The rest of this looks awesome! |
| LOGON32_PROVIDER_DEFAULT = 0 | ||
| LOGON32_LOGON_NETWORK = 3 | ||
| LOGON32_PROVIDER_DEFAULT = Security::LOGON32_PROVIDER_DEFAULT | ||
| LOGON32_LOGON_NETWORK = Security::LOGON32_LOGON_NETWORK | ||
| #XXX for an extra painful alternative, see: http://support.microsoft.com/kb/180548 | ||
| def validate_credentials(passwd) |
There was a problem hiding this comment.
in retrospect, this should have been something like valid_credentials? but wutevs. #breakfastwhenevs
|
👍 nice work man. it's a big change, so let's hope we don't have any regressions where we failed to have tests before. the functional tests pass on my 8.1 workstation too 🚀. |
|
👍 Looks good and I can make tests pass. Squash before merge? |
|
I think I'm ok with the commits (except the last one)... they tell the story I want them to and each one should be working point in time |
|
cool beans |
Changing Net User things to use ffi instead of win32-api
This was a mess of memory corruption, so I've rewritten using the saner ffi approach. It now works with 64-bit rubies.
cc @chef/client-windows @adamedx @ksubrama