Use Docxodus 12.6.2 and reduce editor input latency - #46
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Styled typing inside the NVCA footnote previously reached 816 ms because its leading tab forced a complete native transaction. This change uses the released Docxodus 12.6.2 operation and removes additional React-side blocking found in browser traces. The final five production NVCA runs pass all nine interaction phases against the strict 150 ms gate; their maxima are 104, 120, 136, 96, and 144 ms.
Pins
docxodusand optional@docxodus/exportto 12.6.2 with unmodified, digest-verified runtime assets. Incoming document preparation waits for idle input, yields during sanitization, and observes layout readiness instead of forcing layout during setup. Canvas anchor resolution reuses native formatting reads; canonical aliases share results and invalidate together while retaining story identity, native version, and transaction guards. Public APIs, native pagination decisions, and atomic fallbacks for excluded structures are preserved.The isolated native footnote insertion improves from 600–622 ms to 29.5–32.7 ms; its final production interior-typing phase takes 32 ms. Traced incoming setup tasks fall from 55.5–80.9 ms to 15.5–20.5 ms. Browser regressions cover leading tabs, formatting alias invalidation, Enter without a full anchor scan before its first frame, and cancellation of incoming layouts during real IME composition.
Measurement summaries and reproduction commands. Generated benchmark JSON and trace artifacts are kept outside version control.
The earlier failed measurements remain in the documented summaries. Two original 184 ms delays did not reproduce under tracing, so their exact causes remain unproven. The final result covers the measured phases on one machine; background layout, individual native calls, and unmeasured workloads can still exceed 150 ms.
Validation at application revision
af634f9:npm run check: lint, 82 unit tests, types, library/demo builds, complete API identity audit, and generated pagination check pass.Upstream fix: Docxodus #802, PR #803.