chore: halve the length of the Snapchat arroyo.db notes - #1089
Merged
Conversation
The notes had grown to 837 and 418 words, a page of prose each. LAVA renders them through antd Descriptions with no white-space handling, so the newline separators collapse and the whole thing arrives as one block of text. Long was already bad; unbroken made it unreadable. Now 426 and 269 words, roughly half, with every sourced claim kept. Checked 23 load-bearing facts survive the rewrite, including the protobuf path and its cross-checks, the 11-vs-8 row counts, the direction derivation, "why a Recovered row is absent is not established", "absence is not evidence it did not exist", and the verification command. Changes are structural rather than deletions: paragraphs now open with a short label (Record Origin, Method, Limits) so they act as anchors even when the newlines collapse, duplicate statements of the two-read method were merged, and the only fact dropped is the 2-of-30-versus-6-tables detail, which defends the implementation rather than telling an examiner anything. That reasoning lives in the PR history and in issue #957. Behaviour untouched: 16 message rows and 5 conversation rows on hc_pixel8pro_a17, unchanged. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up to #1053. Notes only, no behaviour change.
Why
The notes had grown to 837 and 418 words, a page of prose each. Worse, LAVA renders them through antd
Descriptionswith the raw string aschildrenand nowhite-spacehandling, so the\nseparators collapse and the whole thing arrives as one unbroken block. Long was already bad; unbroken made it genuinely hard to read.What changed
Structural rather than deletions:
Record Origin.,Method.,Limits.) so they act as anchors even while the newlines collapse, and read as headings once they don't.Every sourced claim is kept. I checked 23 load-bearing facts survive the rewrite: the protobuf path
4 > 4 > 2 > 1and its cross-checks againstsender_id/client_conversation_id/content_type/ the timestamps, the 11-versus-8 row counts, the direction derivation and its fallback,content_typereported as stored, media not decrypted, "why a Recovered row is absent is not established", "absence is not evidence it did not exist", the 29-row measured gap, and thesqlite3verification command.One fact was dropped: the "2 of 30 tables diverge by count, 6 by key" detail in the messages notes. It defends the implementation rather than telling an examiner anything, and the reasoning lives in #1053 and #957. The equivalent point is kept in the conversations notes, where it is the concrete reason that artifact finds anything at all.
Validation
Corpus
hc_pixel8pro_a17: 16 message rows and 5 conversation rows, unchanged.check_claim_language.pyclean,check_html_safety.pyclean, pylint 10.00/10,lint_changed.pyno new warnings.Related, not fixed here
The collapse itself is a one-line viewer-side fix in
ArtifactInfoModal.jsx, wrapping the value in<span style={{ whiteSpace: 'pre-wrap' }}>. That would fix multi-paragraph notes for every artifact in all five extractors at once, but it is LAVA's side and not mine to change.Co-Authored-By: Claude Opus 5 noreply@anthropic.com