Repository navigation
Stream immutable hashes and Orca speech evidence windows - #705
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The original Orca run38173254808/1 exceeded the unchanged 2 GiB RSS cap. Its observer hashes a 293 MB Chromium executable with a whole-file allocation and repeatedly copies the speech log prefix. Hash every byte in fixed 64 KiB chunks and read only the selected original speech byte window during live observation and final validation. Keep all phrases, 15 recovery actions, generation fences, deadlines, first causes and the cap.
Validation: focused original qualification passed 29 navigation controls, the diagnostic self-test and all 15 new hash/window controls. Tests cover empty/boundary/multiple/short reads, real executable hash correspondence, original failures, offsets, split UTF-8/partial lines, phrases, append/truncation and evidence limits. The original native browser reports passed 8 cases on each engine with zero unexpected/skipped/flaky cases; the native journey remains failed. The RSS census identifies actor allocation, but copy timing and RSS cause attribution remain unknown. Hosted exact-head checks and fresh downstream native qualification remain pending.