You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Timestamp parsing does not fully support RFC 3339 #709
@duglin any opinions here? My read on the note in the spec is that this is an optional thing that applications may opt into, and in other lanugage ecosystems (and the default here), the space formatting is not considered valid.
IMO we should clarify this at the spec level and then make it consistent between SDKs
NOTE: ISO 8601 defines date and time separated by "T".
Applications using this syntax may choose, for the sake of
readability, to specify a full-date and full-time separated by
(say) a space character.
wow - talk about hurting interop. I think we should clarify that "T" MUST be used - I'll bring this up with the group.
@marian-varga I opened cloudevents/spec#1410 to clarify the spec. Thanks so much for noticing. Small favor... in the future if you notice more spec-related issues please open an issue in the spec repo so it doesn't get lost in the SDK workload. thanks again!
Activity
Cali0707 commentedon Nov 12, 2026
@duglin any opinions here? My read on the note in the spec is that this is an optional thing that applications may opt into, and in other lanugage ecosystems (and the default here), the space formatting is not considered valid.
IMO we should clarify this at the spec level and then make it consistent between SDKs
duglin commentedon Nov 12, 2026
From RFC3339:
wow - talk about hurting interop. I think we should clarify that "T" MUST be used - I'll bring this up with the group.
duglin commentedon Nov 12, 2026
I love the "say" word in there too... leaves it open for any char to be used. Amazing
duglin commentedon Nov 12, 2026
@marian-varga I opened cloudevents/spec#1410 to clarify the spec. Thanks so much for noticing. Small favor... in the future if you notice more spec-related issues please open an issue in the
specrepo so it doesn't get lost in the SDK workload. thanks again!