Allowed credential types should be a bitfield #1903
Conversation
|
I didn't realize that the @carlosmn are you happy with the language here? |
|
/cc @jamill |
|
Doh! Looks like something changes since #1853 was introduced. I'll dig in. |
include/git2/transport.h
Outdated
There was a problem hiding this comment.
I think this is the data/payload argument that lets the callback retrieve its context, rather than the data to sign. The callback gets passed two buffers for that already. One is the key data, the other the challenge data (though why this data (opaque) it's a void** is still a mystery).
|
@carlosmn I'm not sure that I really captured what you were looking to convey here? |
|
(Also I beefed up the test a little bit to complain about missing env vars, since that was not obvious to me.) |
|
So, I think that there are probably a number of issues here that we should look at in the longer-term, but I hope that this at least satisfies the immediate problems. |
|
My point was that |
|
This is not a bitfield at all, but yeah, the change looks nice. ^^ |
Allowed credential types should be a bitfield
Allowed credential types should be a bitfield
Obsoletes #1853.