Interfaces for SCTs, feedback wanted

[alex]

Refs #3461

[mention-bot]

@alex, thanks for your PR! By analyzing the history of the files in this pull request, we identified @reaperhulk, @intgr and @queenp to be potential reviewers.

[alex]

@reaperhulk important question, should this be a bare-class, or an ABC+builder like RevokedCertificate.

[reaperhulk]

What does it look like to build an SCT with just the bare class?

[alex]

Unclear! SCTs need to signed (it's in the name), so I think that means they need a builder on one side.

…

On Mon, Mar 20, 2017 at 7:39 PM, Paul Kehrer ***@***.***> wrote: What does it look like to build an SCT with just the bare class? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3467 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAADBAaFJEaM7sV4VHlz65DbVBJVKR5zks5rnw5GgaJpZM4MjC0K> .

-- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero GPG Key fingerprint: D1B3 ADC0 E023 8CA6

[reaperhulk]

I think anything with signatures should have a builder 👍

[alex]

Ok, which means this gets to be an ABC. :writes code:

…

On Mon, Mar 20, 2017 at 7:44 PM, Paul Kehrer ***@***.***> wrote: I think anything with signatures should have a builder 👍 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3467 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAADBALXnSrBN8U-DYW6v-vLNzfmfZNiks5rnw9GgaJpZM4MjC0K> .

-- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero GPG Key fingerprint: D1B3 ADC0 E023 8CA6

[eranmes]

Nit: Mention this is the hash of the log's public key? Or could refer to the RFC.

[alex]

Good call.

[eranmes]

The timestamps from logs are always in UTC. Not sure how it affects this field, if at all.

[alex]

@reaperhulk what do you think? The not_before and not_after on x.509 certs are naive datetimes, do you remember why we did that? Is it because python is bad and didn't have a cheap way to give us a UTC tzinfo?

[reaperhulk]

That is exactly why.

[reaperhulk]

We document that as stating they're naïve datetimes representing UTC in the x509 docs.

[alex]

I'll go copy that prose.

…

On Tue, Mar 21, 2017 at 11:29 AM, Paul Kehrer ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In docs/x509/certificate-transparency.rst <#3467 (comment)>: > + :type: :class:`~cryptography.x509.certificate_transparency.Version` + + The SCT version as an enumeration. Currently only one version has been + specified. + + .. attribute:: log_id + + :type: bytes + + An opaque identifier, indicating which log this SCT is from. + + .. attribute:: timestamp + + :type: :class:`datetime.datetime` + + A naïve datetime representing the timestamp at which the log asserts We document that as stating they're naïve datetimes representing UTC in the x509 docs. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3467 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAADBM6Vc-_bDVyBTI3CIJc70Y-fGWIiks5rn-zvgaJpZM4MjC0K> .

-- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero GPG Key fingerprint: D1B3 ADC0 E023 8CA6

[reaperhulk]

Is the acceptable time period something each log decides?

[alex]

Yeah, that's the MMD, max merge delay. Want me to include that bit of terminology in the docs?

…

On Tue, Mar 21, 2017 at 1:59 PM, Paul Kehrer ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In docs/x509/certificate-transparency.rst <#3467 (comment)>: > +.. currentmodule:: cryptography.x509.certificate_transparency + +Certificate Transparency is a set of protocols specified in :rfc:`6962` which +allow X.509 certificates to be sent to append-only logs and have small +cryptographic proofs that a certificate has been publicly logged. This allows +for external auditing of the certificates that a certificate authority has +issued. + +.. class:: SignedCertificateTimestamp + + .. versionadded:: 1.9 + + SignedCertificateTimestamps (SCTs) are small cryptographically signed + assertions that the specified certificate has been submitted to a + Certificate Transparency Log, and that it will be part of the public log + within some time period. Is the acceptable time period something each log decides? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3467 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAADBET96wMvnwdb5xTz7qhOupF4eyvSks5roBAYgaJpZM4MjC0K> .

-- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero GPG Key fingerprint: D1B3 ADC0 E023 8CA6

[reaperhulk]

Yes please

[alex]

Jenkins, retest this please.