Failing on chain loops

[Razumain]

Section 10.1 states:

Federation participants MUST NOT attempt to fetch Entity Statements they already have obtained during this process to prevent loops. An invalid_trust_chain error SHOULD be returned if a loop is encountered.

I believe this statement is incorrect.

I think the first part is correct, but as long as you are able to build a valid path, it should be OK. I fail to see why you "SHOULD" fail.

[selfissued]

It's not clear to me that interoperability would be improved by not rejecting an invalid Trust Chain. That said, returning the error is a SHOULD - not a MUST - so if you want to proceed and find a way to do so, the spec doesn't prohibit that.

[Razumain]

I think you miss my point. The fact that you "encounter" a loop, does not mean that you incorporate it in your validated chain.

Do you see the difference?

Say there is a valid chain:
A -> B -> C -> D -> E

Say that this chain also has a potential loop that you could be tricked into if you are not careful:
A -> B -> C -> D -> B -> C -> D -> B -> C -> D -> B -> C -> D -> B -> C -> D ....

Just because you "encounter" = observe this loop, does not make the valid chain invalid. (A -> B -> C -> D -> E)

Maybe this is just a wording thing. As long as you find a valid chain that has no repetition (loops) in it. You are good and should not fail because it is possible to build a loop.

[selfissued]

Addressed by #253

[Razumain]

@selfissued I made a comment to the PR that you didn't respond to. Was that because you didn't see it, or because you dismissed it?