Skip to content

content(registry): the witness deploy does not reach the September 7 capture - #60

Merged
imran-siddique merged 1 commit into
mainfrom
witness-no-restamp-note
Sep 12, 2026
Merged

content(registry): the witness deploy does not reach the September 7 capture#60
imran-siddique merged 1 commit into
mainfrom
witness-no-restamp-note

Conversation

@imran-siddique

Copy link
Copy Markdown
Member

The witness operator flagged this page back on September 9 as the one that would go stale at deploy time, because the sentence about the two fields lives here rather than in the repo README that #72 fixed.

The deploy happened on September 12, and the page needed a change after all, though not the one either of us expected. It said both fields stay false for the September 7 capture because the witness signature covers the header. True, and the weaker reason.

The stronger one: /checkpoints deduplicates on a content-addressed entry hash, so checkpoint 1 cannot be re-submitted for a new receipt at all. And because that hash is derived from the signed fields rather than the submitted bytes, reformatting the submission does not produce a new one either. A receipt carrying a witness time therefore requires a checkpoint the witness has not seen.

Now matches trace-registry LIMITATIONS.md (#77, #78). One paragraph, no layout or structural change.

🤖 Generated with Claude Code

https://claude.ai/code/session_01X2GDChXjA7BAdDNzCAmBJv

…capture

The page said both fields stay false for that capture because the signature
covers the header, which is true and is the weaker reason. The witness deployed
iat and grade signing on September 12 and confirmed that /checkpoints
deduplicates on a content-addressed entry hash, so checkpoint 1 cannot be
re-submitted for a new receipt at all, and because that hash is derived from
the signed fields rather than the submitted bytes, reformatting does not
produce a new one either. A receipt carrying a witness time needs a checkpoint
the witness has not seen.

This is the page the witness operator flagged as the one that would go stale
at deploy time, since the sentence lives here rather than in the repo README.
It now matches trace-registry LIMITATIONS.md (#77, #78).

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01X2GDChXjA7BAdDNzCAmBJv
@imran-siddique
imran-siddique merged commit cd863d6 into main Sep 12, 2026
4 checks passed
@imran-siddique
imran-siddique deleted the witness-no-restamp-note branch September 12, 2026 17:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant