Add sensible default arguments to CredentialPassword::createPassword? about cutelyst

[dantti]

I thought of doing this before v2, but gave up on the idea that the user should know what he is doing, but I think you might be right about misuse.
The problem with this now is that I believe such change might break ABI, but not sure.

[buschmann23]

As I had the same “issue“ to find the most appropriate values for PBKDF2 - especially the iterations count, I wrote a small test application to try to find the best fitting value for the iterations that would take up to 500ms for the pure algorithm.

See https://github.com/Huessenbergnetz/pbkdf2test

[louwers]

##C++ said that adding default arguments won't break ABI on any of the platforms they are aware of since it simply is not part of the ABI. The check for default argument is made at the call site. Furthermore, default arguments are added to the header, so by just swapping out headers without recompiling you could add default arguments.

[dantti]

Hmm when I searched for this I found:
https://stackoverflow.com/questions/41763930/does-adding-an-arguments-with-default-values-to-functions-break-abi/41764100

[dantti]

An alternative would be a method that hides all arguments but the password, iirc there's a library that does this.
If the default argument is stored by the compilation of the client an update of the values would not reflect on an application unless it's rebuilt, so this option might be a good idea I think.

[dantti]

I've talked to some people and they told me it does breaks the ABI, so to avoid problems I added a simple method.

[louwers]

@dantti Adding arguments with default values changes the ABI, this is what the StackOverflow thread is about. But what I proposed is that we add default values for already existing arguments. I don't see any logical way how this can change the ABI. Where you clear on this difference when you asked those people?

Sorry for replying late.