Fix turn-input ingress perf assertion (attachment survival, not position)#128
Merged
Conversation
The turn-input ingress perf measurement asserted the inline attachment was the last materialized item. A claim aggregates many enqueued inputs and only the first carries an attachment, so the attachment sits mid-list and the check could never pass. It has failed every nightly since it was introduced, leaving the runtime performance budgets ungated. Assert that the attachment survives the claim/materialize round trip rather than where it lands. Release-Notes: Internal: Fixed a turn-input ingress performance assertion that checked attachment position instead of survival, which had left nightly runtime performance budgets ungated.
SamGalanakis
force-pushed
the
samuel-perf-attachment-assert
branch
from
July 25, 2026 10:54
37d1334 to
a5aad9d
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 nightly Performance lane has failed since 2026-07-24 with
turn-input ingress active claim lost attachment bytes.Diagnosed: a stale harness assertion, not a product defect — there is no attachment loss. The attachment is present and byte-identical; the assertion checked the wrong position.
FIG-564 rewrote both checks from an order-independent lookup to a positional
items.last(), copying the shape from a conformance test where the claim holds exactly one input. The perf harness claims 16–24 inputs and only the first carries an attachment, so it materializes at index 1 of 17 (active) and 17 of 25 (next) —items.last()is aText. It could never have passed. Verified by running the harness with non-fatal probes:Attachment(Inline { media_type: "image/png", bytes: [1,2,3,0] }).Both assertions now check that the attachment survives the claim/materialize round trip rather than where it lands.
Impact: the nightly perf lane is the sole enforcement point for
scripts/perf_guard_budgets.json, so runtime performance budgets have been ungated since 2026-07-22 (the lane is nightly-only, so it went red two days after the change shipped).Verified:
cargo build -p lash-perf --release, fmt, clippy-D warningsclean.