qlog
qlog copied to clipboard
quicwg
The IETF I-D documents for the qlog format
Alternative to #284. This follows Lennox's suggestion to take influence from RFC 8285 and reference extension schema according to a URI. IETF documents can use URNs under a formal namespace,...
Fixes #379. Also closes https://github.com/quicwg/qlog/issues/261, https://github.com/quicwg/qlog/issues/176, https://github.com/quicwg/qlog/issues/170, https://github.com/quicwg/qlog/issues/124, https://github.com/quicwg/qlog/issues/192, https://github.com/quicwg/qlog/issues/297. Adds extensibility for all events through the `* $$category-name-ext` pattern first used in #400, now extended to ALL events as...
As noticed by @LPardue [here](https://github.com/quicwg/qlog/pull/417#discussion_r1529706045), we don't actually say anywhere that CDDL field ordering does not matter (it's implied by CDDL and JSON themselves, but probably good to reinforce explicitly)....
With #417 and the `additional-schema` proposal, we're moving to a good state for qlog's extensibility. However, we feel additional prose is needed to make people make good use of this...
With https://github.com/quicwg/qlog/pull/415 and the upcoming PR describing general extensibility, we won't yet have discussed guidelines for extending other stuff (e.g., $QuicFrame or $H3Parameters or $H3FrameType), so don't forget to add...
Fixes #411 Same approach as the one we look for ALPNInformation. Allows endpoints to log either, because sometimes it is more useful to print and consume a string like "It's...
While normally HTTP header names and values are expected to be safe strings (possibly with even further restrictions on the chartacter set that is allowed), it's possible that an endpoint...
Closes #398. PTAL @hlandau. The suggestion of using `null` as a JSON value in #398 makes conceptual sense, but not really since we're working with CDDL to define the events...