ASN.1 Integer minimal encoding is required per X.690-0207 Section 8.3.2

[brimworks]

Validation of ASN1.Integer in Go code was failing due to the encoding not being minimal, and the spec indicates that it is required to be minimal.

[CBenoit]

Thank you! I approved CI run. I’ll take more time for reviewing next Monday 😃

[CBenoit]

We can’t take owership of bytes anymore in this case though. Is start == 0 not common enough? Also, do you have code in mind requiring keeping the bytes at call site? In my experience the input object is never used again once the IntegerAsn1 struct is built.

[brimworks]

I can remove this comment if you would like... but in our usage, we are trying to translate an EccParameter into an IntegerAsn1. The EccParameter only has a value() method that allows one to obtain a &[u8] which represents the parameter:

https://docs.rs/tss-esapi/latest/tss_esapi/structures/struct.EccParameter.html

As it stands, I believe our only option is to convert .to_vec() the &[u8] in order to construct the IntegerAsn1... and in the scenario where a resize needs to take place, we slice the Vec and create yet another Vec, but if the parameter was a &[u8] we could avoid one of the Vec conversions.

Alternatively, if the API took a mut Vec then we could drain(0..start) and avoid the need for the extra Vec allocation.

...although, perhaps all of this is premature optimization.

[brimworks]

@CBenoit how would you like me to proceed? Would you like me to remove the comment in regard to taking a &[u8]?

[CBenoit]

@CBenoit how would you like me to proceed? Would you like me to remove the comment in regard to taking a &[u8]?

I’m open to changing the API to a &[u8] if you want to.

[brimworks]

@CBenoit how would you like me to proceed? Would you like me to remove the comment in regard to taking a &[u8]?

I’m open to changing the API to a &[u8] if you want to.

How about if I just remove this comment for now, that idea can always be addressed in a separate PR.

[CBenoit]

@CBenoit how would you like me to proceed? Would you like me to remove the comment in regard to taking a &[u8]?

I’m open to changing the API to a &[u8] if you want to.

How about if I just remove this comment for now, that idea can always be addressed in a separate PR.

Sounds good!

[CBenoit]

Looks good to me! Thank you 😄
I’ll publish next candidate version next Monday!

[brimworks]

Thanks!