Transaction-Science/open-standards
Watch
1
0
Fork
You've already forked open-standards
 0
Code Issues Pull requests Projects Releases 15 Packages Wiki Activity Actions

CI: skip the quantum set rerun test on the linux leg until #538 leaves the path out (ends by itself) #546

Merged
dcharlot merged 1 commit from ci/quantum-set-path-skip-linux into main 2026-10-30 00:46:26 -04:00
Conversation 0 Commits 1 Files changed 1 +13 -1
dcharlot commented 2026-10-30 00:33:40 -04:00
Owner
Copy link

wai-quantum's engine::sets::tests::one_receipt_binds_every_item_and_any_change_breaks_it runs the same set twice and compares the two runs item by item, less each item's elapsed_ms and receipt, but with its Merkle path. A path holds the other items' leaves, and a leaf is the digest of an item's result, elapsed_ms included. On the slower x86_64 runners (jetson-hub, the iMac VM) the items' times differ between the two runs, so the paths differ while everything else holds. This failed 3 out of 3 on #540 and #541, in code neither PR touches. It passes on the Framework, the Studio, linux-arm64 and macOS.

#538 holds the fix: it leaves the path out of the comparison. #538 waits on the Windows runner (cbio-device-0000), which is offline, and any change under wai/wai-quantum/** waits with it, because the WebGPU workflow's Windows leg gates it.

This PR: on the energy-receipts workflow's linux leg only, that one test is skipped (--skip … --exact) while wai/wai-quantum/src/engine/sets.rs still compares the path (it has no remove("path")).

  • The skip ends by itself once #538, or any equal fix, lands. Nothing needs undoing.
  • linux-arm64 and macos run the test whole throughout.
  • No test code changes.

Checked. The step's script, run with a stand-in cargo, under bash 3.2 with -euo pipefail, on every leg with main's sets.rs and with #538's:

Leg main's sets.rs #538's sets.rs
linux skips the one test runs all
linux-arm64 runs all runs all
macos runs all runs all

This PR's own run of the workflow shows the skip line on the linux leg.

🤖 Generated with Claude Code

`wai-quantum`'s `engine::sets::tests::one_receipt_binds_every_item_and_any_change_breaks_it` runs the same set twice and compares the two runs item by item, less each item's `elapsed_ms` and `receipt`, but **with its Merkle `path`**. A path holds the other items' leaves, and a leaf is the digest of an item's result, `elapsed_ms` included. On the slower x86_64 runners (jetson-hub, the iMac VM) the items' times differ between the two runs, so the paths differ while everything else holds. This failed 3 out of 3 on #540 and #541, in code neither PR touches. It passes on the Framework, the Studio, `linux-arm64` and macOS. **#538 holds the fix:** it leaves the path out of the comparison. #538 waits on the Windows runner (cbio-device-0000), which is offline, and any change under `wai/wai-quantum/**` waits with it, because the WebGPU workflow's Windows leg gates it. **This PR:** on the energy-receipts workflow's `linux` leg only, that one test is skipped (`--skip … --exact`) while `wai/wai-quantum/src/engine/sets.rs` still compares the path (it has no `remove("path")`). - The skip ends by itself once #538, or any equal fix, lands. Nothing needs undoing. - `linux-arm64` and `macos` run the test whole throughout. - No test code changes. **Checked.** The step's script, run with a stand-in `cargo`, under bash 3.2 with `-euo pipefail`, on every leg with main's `sets.rs` and with #538's: | Leg | main's `sets.rs` | #538's `sets.rs` | |---|---|---| | `linux` | skips the one test | runs all | | `linux-arm64` | runs all | runs all | | `macos` | runs all | runs all | This PR's own run of the workflow shows the skip line on the linux leg. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
CI: the quantum set rerun test skipped on the linux leg until its path is left out
All checks were successful
standards workspaces build + test / which workspaces this change touches (pull_request) Successful in 7s
Details
standards workspaces build + test / macOS energy backends (pull_request) Has been skipped
Details
WAI energy receipts / does this change touch it (pull_request) Successful in 6s
Details
repository checks / lockfiles, advisories, tests in CI, workflow hygiene, spec and section references, public text (pull_request) Successful in 2m36s
Details
WAI energy receipts / energy receipts · macos (pull_request) Successful in 8m27s
Details
WAI energy receipts / energy receipts · linux (pull_request) Successful in 9m9s
Details
WAI energy receipts / energy receipts · linux-arm64 (pull_request) Successful in 9m23s
Details
a3d46b352a 
engine::sets' rerun test compares each item between two runs, less its
elapsed_ms and receipt, but with its Merkle path. A path holds the other
items' leaves, and with them their elapsed_ms, so on the slower x86_64
runners (jetson-hub, the iMac VM) two runs differ there: three failures in
three runs on #540 and #541, none of them in code those PRs touch.

#538 holds the fix (the path left out) and waits on the Windows runner,
which is offline; any change under wai/wai-quantum waits with it. So the
energy-receipts linux leg skips that one test while sets.rs still compares
the path, and runs it whole again by itself once the file leaves the path
out. linux-arm64 and macOS run it whole throughout.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
Transaction-Science/open-standards!546
No description provided.