Skip to content

x509: accept non-string types in Issuer - #552

Open
jsha wants to merge 1 commit into
zmap:masterfrom
jsha:non-string-pkix-name-2
Open

jsha wants to merge 1 commit into
zmap:masterfrom
jsha:non-string-pkix-name-2

Conversation

@jsha

@jsha jsha commented Nov 19, 2026

Copy link
Copy Markdown

Specifically for ParseRevocationList.

An example of a non-string type in Issuer is the id-rdna-trustAnchorID (a RELATIVE-OID) used in MTCs:

https://ietf-plants-wg.github.io/merkle-tree-certs/draft-ietf-plants-merkle-tree-certs.html#name-representing-certification-
https://datatracker.ietf.org/doc/html/draft-ietf-tls-trust-anchor-ids-07#section-4.1

Upstream fixed this a different way, by patching parseName to be more liberal: https://go-review.googlesource.com/c/go/+/778681.

However, in this repository, the cert-parsing paths don't use parseName, they just use asn1.Unmarshal, which accepts these attribute types just fine. To me it seems more consistent, and a smaller change, to use asn1.Unmarshal for the ParseRevocation path.

An alternative would be to backport the upstream parseName change, and also update the cert-parsing paths to use parseName (as they do in upstream). However, that appears to break tests with regards to some of the permissiveness properties that are the point of this package.

I came across this while adding MTC support to Boulder, and it's on the blocking path for us.

Specifically for ParseRevocationList.

Upstream fixed this a different way, by patching `parseName` to be more
liberal: https://go-review.googlesource.com/c/go/+/778681.

However, in this repository, the cert-parsing paths don't use `parseName`, they
just use `asn1.Unmarshal`. To me it seems more consistent, and a smaller change,
to use `asn1.Unmarshal` for the ParseRevocation path.

An alternative would be to backport the upstream `parseName` change, and also
update the cert-parsing paths to use `parseName` (as they do in upstream).
However, that appears to break tests with regards to some of the permissiveness
properties that are the point of this package.

I came across this while adding MTC support to Boulder, and it's on the blocking
path for us.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants