On Tue, Sep 19, 2017 at 10:49 PM, userwithuid via dev-security-policy < [email protected]> wrote:
> Either way, in the specific case, StartCom, this criticism seems to be > inapplicable, as the revoked one was never deployed in the first place. I don't think that's a fair conclusion. As noted in the StartCom discussions, the mere existence of such certificate is enough to make it possible for clients (hostile or misinformed) to serve this chain and trigger the behaviour, and thus, as Gerv noted: "While this is probably not a policy violation, it's not good practice." - which is absolutely true. This is not the first time StartCom has done something 'like' this - that is, not a policy violation, but not good practice. During the SHA-1/SHA-2 migration, the (old management) of StartCom signed an existing SHA-1 intermediate with the same name, key, and other attributes with SHA-2. That these were otherwise equivalent meant clients were non-deterministic in their selection of SHA-1 and SHA-2, and may block such certificates, warn on such certificates, or fail to work. The "good practice" in this case is to issue new certificates off a new (SHA-2) intermediate with a new name and new key - not trying to continue to use 'both' intermediates. Many other CAs did this successfully. Things like this are relative in the sense of execution on best practice / avoidance of issues, both those explicitly prohibited and those being 'bad ideas'. _______________________________________________ dev-security-policy mailing list [email protected] https://lists.mozilla.org/listinfo/dev-security-policy

