Roland Bracewell Shoemaker
Roland Bracewell Shoemaker
Yeah Tag would probably make the most sense, although we've somewhat backed ourselves into a corner by making it just a typed uint8, so there isn't really an obvious place...
Coming back to this, the API design of this package is clearly not ideal, and isn't particularly conducive to making these kinds of changes cleanly. There are a number of...
For just the Boolean type, the new APIs would be: ``` func (b *Builder) AddASN1BooleanWithTag(v bool, tag asn1.Tag) func (s *String) ReadASN1BooleanWithTag(out *bool, tag asn1.Tag) bool func (s *String) ReadOptionalASN1BooleanWithTag(out...
I think the easiest way to implement this would be to hold an in-memory map of authorization IDs that are currently being validated.
Possible solution: https://github.com/letsencrypt/boulder/compare/prevent-parallel-validation?expand=1
Ah yup, good point. I was trying to avoid having to actually touch the database but that looks like it's the only real solution that supports a multi-RA deployment 😕.
This is related to the [`Present: xxx, Valid for issuance: xxx` logging code in the VA](https://github.com/letsencrypt/boulder/blob/b82c244e65e1eeae8b93989ebbf2a40de2189937/va/va.go#L611).
Re-opening for proposal acceptance.
Uh, this is not the right issue.
@andrewgdotcom you're running into https://github.com/golang/go/issues/25098 which was fixed by https://github.com/golang/crypto/commit/ae8bce0, likely combined with usage of an older version of Go. What is your `go version`?