kayg/infraternal
Watch
1
0
Fork
You've already forked infraternal
 0
Code Issues 356 Pull requests Projects Releases Packages Wiki Activity Actions

J70: run the full fast gate on arm64 build hosts #296

New issue
Open
opened 2026-11-01 16:38:14 +00:00 by kayg · 24 comments
kayg commented 2026-11-01 16:38:14 +00:00
Owner
Copy link

J70: make the full fast gate portable to the arm64 build VM.

Plan (estimate: 3 hours including VM queue/build time):

  1. Reproduce the five restart failures on dev-02 at the origin/dev base; trace the failing operation and compare host state with compute01.
  2. Enable both static targets from either host; execute runtime checks on the native target and explicitly skip foreign execution where necessary.
  3. Compare same-commit x86 ELF outputs and toolchain/build metadata to explain or eliminate the reported 8,704-byte difference (D213).
  4. Commit logical changes atomically, push gate-on-arm, run full scripts/verify fast on dev-02 and compute01, retain verbatim logs and remove job target directories.

Scope: no new engines or general infrastructure, no deployment or merge.

J70: make the full fast gate portable to the arm64 build VM. Plan (estimate: 3 hours including VM queue/build time): 1. Reproduce the five restart failures on dev-02 at the origin/dev base; trace the failing operation and compare host state with compute01. 2. Enable both static targets from either host; execute runtime checks on the native target and explicitly skip foreign execution where necessary. 3. Compare same-commit x86 ELF outputs and toolchain/build metadata to explain or eliminate the reported 8,704-byte difference (D213). 4. Commit logical changes atomically, push gate-on-arm, run full scripts/verify fast on dev-02 and compute01, retain verbatim logs and remove job target directories. Scope: no new engines or general infrastructure, no deployment or merge.
kayg commented 2026-11-01 16:38:30 +00:00
Author
Owner
Copy link

started J70 on branch gate-on-arm, base origin/dev a26615c50e. isolated control checkout: /home/kayg/Developer/Worktrees/infraternal-j70. plan and 3-hour estimate are in the issue body.

started J70 on branch gate-on-arm, base origin/dev a26615c50e76ed86f36d25aa094283cf840d5320. isolated control checkout: /home/kayg/Developer/Worktrees/infraternal-j70. plan and 3-hour estimate are in the issue body.
kayg commented 2026-11-01 16:41:19 +00:00
Author
Owner
Copy link

Finding: the five restart failures are consistent with trusted-path rejection, not arm instructions or a short timeout. dev-02: /home/claude=0750, /home/claude/worktrees=0775. compute01: /home/claude=0700, /home/claude/worktrees=0775. Opened::open_with walks ancestors root-first and refuses group-write before an owner-only 0700 ancestor. Existing dev-02 ~/worktrees/j64.restart.log records outcome="refused" for the kernel-resolved current test image, followed by exactly the five reported 60s failures. Reproducing on this job's unmodified origin/dev base before correcting host state; no timeout changes planned.

Static dispatch regression was observed failing on aarch64 before the script fix, and passing afterward. Native CLI/key-memory selection and explicit foreign-runtime skips are being checked too. The historical x86-cross target has already been removed, so comparing freshly built identical a26615c50 sources on both machines instead. Current job target: ~/worktrees/j70/target on each VM; detached logs ~/worktrees/j70-{baseline-tests,static}.log.

Finding: the five restart failures are consistent with trusted-path rejection, not arm instructions or a short timeout. dev-02: /home/claude=0750, /home/claude/worktrees=0775. compute01: /home/claude=0700, /home/claude/worktrees=0775. `Opened::open_with` walks ancestors root-first and refuses group-write before an owner-only 0700 ancestor. Existing dev-02 `~/worktrees/j64.restart.log` records outcome="refused" for the kernel-resolved current test image, followed by exactly the five reported 60s failures. Reproducing on this job's unmodified origin/dev base before correcting host state; no timeout changes planned. Static dispatch regression was observed failing on aarch64 before the script fix, and passing afterward. Native CLI/key-memory selection and explicit foreign-runtime skips are being checked too. The historical x86-cross target has already been removed, so comparing freshly built identical a26615c50 sources on both machines instead. Current job target: ~/worktrees/j70/target on each VM; detached logs ~/worktrees/j70-{baseline-tests,static}.log.
kayg commented 2026-11-01 16:45:22 +00:00
Author
Owner
Copy link

Confirmed root cause and correction: host-state permissions, not architecture or timeouts.

At unmodified origin/dev Rust sources a26615c50, cargo test --locked -p infraternal --bin infraternal restart::tests -- --nocapture on dev-02 finished with exactly the five reported failures: 13 passed, 5 failed, 61.45s, EXIT=101. The log records outcome="refused" for /home/claude/worktrees/j70/target/debug/deps/infraternal-6ba337e4243386dd.

Changed only /home/claude from 0750 to 0700, matching compute01; /home/claude/worktrees remains 0775. Re-ran the exact already-built test executable, without recompilation. All five formerly failing tests now pass. The production ancestry rule correctly requires a private ancestor before allowing group-writable build paths beneath it. This is a host hardening fix, with no weakening of admission or timeout changes. Logs: ~/worktrees/j70-baseline-tests.log and ~/worktrees/j70-restart-fixed.log.

Static gate portability committed and pushed: 3798e7e6f. Both same-source x86 comparison builds continue detached in ~/worktrees/j70-static.log.

Confirmed root cause and correction: host-state permissions, not architecture or timeouts. At unmodified origin/dev Rust sources a26615c50, `cargo test --locked -p infraternal --bin infraternal restart::tests -- --nocapture` on dev-02 finished with exactly the five reported failures: 13 passed, 5 failed, 61.45s, EXIT=101. The log records `outcome="refused"` for `/home/claude/worktrees/j70/target/debug/deps/infraternal-6ba337e4243386dd`. Changed only `/home/claude` from 0750 to 0700, matching compute01; `/home/claude/worktrees` remains 0775. Re-ran the exact already-built test executable, without recompilation. All five formerly failing tests now pass. The production ancestry rule correctly requires a private ancestor before allowing group-writable build paths beneath it. This is a host hardening fix, with no weakening of admission or timeout changes. Logs: ~/worktrees/j70-baseline-tests.log and ~/worktrees/j70-restart-fixed.log. Static gate portability committed and pushed: 3798e7e6f. Both same-source x86 comparison builds continue detached in ~/worktrees/j70-static.log.
kayg commented 2026-11-01 16:54:57 +00:00
Author
Owner
Copy link

Current-source same-recipe comparison (a26615c50 Rust sources, same checkout path, nix dev shell, Rust 1.99.0 b940084d7, LLVM 23.1.1, Zig 0.15.2): native x86 image 21,642,320 B; arm-host x86 cross image 21,649,296 B (+6,976 B). No occurrences of /home/claude, /nix/store/, /build/, /rustup/ or /.cargo/ in either shipping binary. Historical 37bbda455 rebuild is in progress to recover the exact reported +8,704 B case.

Native archives ruled out as runtime differences: all 982 differing OpenSSL object files differ only in .debug_str/.debug_info. All VT archive objects match except libhighway_zcu.o, whose only differing section is .debug_line. Those sections are stripped from shipping images. Cargo host metadata differs: identical crossbeam-utils build-script output SHA cb6814e09e64fb079e8defed607f9407cb5ef09b64fd8a31d4c581164f39bac5, but native target crate identity 615b1674294a9fde versus cross c11c495add14c870, with different host build-script dependency identities. Target, features, profile, source path, rustflags and compile_kind match. Continuing to trace the Rust identity/codegen distinction; no compiler-flag or release-profile changes.

Current-source same-recipe comparison (a26615c50 Rust sources, same checkout path, nix dev shell, Rust 1.99.0 b940084d7, LLVM 23.1.1, Zig 0.15.2): native x86 image 21,642,320 B; arm-host x86 cross image 21,649,296 B (+6,976 B). No occurrences of /home/claude, /nix/store/, /build/, /rustup/ or /.cargo/ in either shipping binary. Historical 37bbda455 rebuild is in progress to recover the exact reported +8,704 B case. Native archives ruled out as runtime differences: all 982 differing OpenSSL object files differ only in .debug_str/.debug_info. All VT archive objects match except libhighway_zcu.o, whose only differing section is .debug_line. Those sections are stripped from shipping images. Cargo host metadata differs: identical crossbeam-utils build-script output SHA cb6814e09e64fb079e8defed607f9407cb5ef09b64fd8a31d4c581164f39bac5, but native target crate identity 615b1674294a9fde versus cross c11c495add14c870, with different host build-script dependency identities. Target, features, profile, source path, rustflags and compile_kind match. Continuing to trace the Rust identity/codegen distinction; no compiler-flag or release-profile changes.
kayg commented 2026-11-01 16:58:54 +00:00
Author
Owner
Copy link

progress: pushed b99ed63cb (portable static gate plus documented host trust/compiler identity findings). historical 37bbda455 cross rebuild reproduced exactly 21,456,976 bytes. full verify fast runs are detached in ~/worktrees/j70-verify.log on both VMs; dev-02 is queued behind two jobs and its free disk reached 24 GB, below build-slot floor. removed only J70 target/debug/incremental (1.1 GB), which is disposable compiler cache. restart before/after logs retained locally at ~/.cache/codex-scratch/j70/evidence and on dev-02. no pending run will be left at handoff.

progress: pushed b99ed63cb (portable static gate plus documented host trust/compiler identity findings). historical 37bbda455 cross rebuild reproduced exactly 21,456,976 bytes. full verify fast runs are detached in ~/worktrees/j70-verify.log on both VMs; dev-02 is queued behind two jobs and its free disk reached 24 GB, below build-slot floor. removed only J70 target/debug/incremental (1.1 GB), which is disposable compiler cache. restart before/after logs retained locally at ~/.cache/codex-scratch/j70/evidence and on dev-02. no pending run will be left at handoff.
kayg commented 2026-11-01 17:01:40 +00:00
Author
Owner
Copy link

blocked on dev-02 shared disk: now 22 GB free versus mandatory 25 GB build-slot floor. J70 is first in queue, with one approximately 5 GB target after deleting its 1.1 GB incremental cache. asked orchestrator to clear finished-job targets; no other job data touched. compute01 full gate has passed CLI compile-out and is running clippy.

blocked on dev-02 shared disk: now 22 GB free versus mandatory 25 GB build-slot floor. J70 is first in queue, with one approximately 5 GB target after deleting its 1.1 GB incremental cache. asked orchestrator to clear finished-job targets; no other job data touched. compute01 full gate has passed CLI compile-out and is running clippy.
kayg commented 2026-11-01 17:10:05 +00:00
Author
Owner
Copy link

disk blocker resolved by orchestrator cleanup: dev-02 now has 57 GB free. arm full gate admitted at head b99ed63cba; it is running CLI compile-out. compute01 is building workspace tests after passing clippy and real-binary fixture checks. keeping the one J70 target on each VM until validation, then deleting both as requested.

disk blocker resolved by orchestrator cleanup: dev-02 now has 57 GB free. arm full gate admitted at head b99ed63cba0a0f2e05617cbd28d9e812dddab2d2; it is running CLI compile-out. compute01 is building workspace tests after passing clippy and real-binary fixture checks. keeping the one J70 target on each VM until validation, then deleting both as requested.
kayg commented 2026-11-01 17:36:30 +00:00
Author
Owner
Copy link

First full fast runs finished on both VMs with static builds PASS and overall VERIFY_EXIT=1 because the workspace test row failed. All pre-existing Rust sources were byte-identical to a26615c50 (git diff a26615c50 HEAD -- crates empty before the following test repair).

Additional dev-02 host-state assumption: cli_envelope::bare_infraternal_is_the_overview_and_points_to_help assumes exit 0 on the real machine. dev-02 has ~/.config/systemd/user/infraternal-preview.service, which is stopped. The production overview correctly returns exit 5, error=component_down, role=server and an infraternal --help next step; compute01 returns exit 0, role=none. Running the unchanged test executable alone on dev-02 reproduces the failure. Repairing only that assertion to accept the two documented overview outcomes, while verifying the matching ok/error result and help link. No real service will be removed to appease a test.

Additional compute01 failure: install::session_restaging_uses_one_compound_command returned SetupRunning/exit 73 from a private fixture. The exact unchanged test passes five isolated runs (0.41–0.72s each). A bounded full-suite concurrency reproduction is running through build-slot before any proposed change to lock handling. No timeout changes.

Verbatim first-run logs retained at ~/worktrees/j70-verify-first.log and ~/.cache/codex-scratch/j70/evidence/verify-first.log on each VM. They will be attached along with final results.

First full fast runs finished on both VMs with static builds PASS and overall VERIFY_EXIT=1 because the workspace test row failed. All pre-existing Rust sources were byte-identical to a26615c50 (`git diff a26615c50 HEAD -- crates` empty before the following test repair). Additional dev-02 host-state assumption: cli_envelope::bare_infraternal_is_the_overview_and_points_to_help assumes exit 0 on the real machine. dev-02 has ~/.config/systemd/user/infraternal-preview.service, which is stopped. The production overview correctly returns exit 5, error=component_down, role=server and an infraternal --help next step; compute01 returns exit 0, role=none. Running the unchanged test executable alone on dev-02 reproduces the failure. Repairing only that assertion to accept the two documented overview outcomes, while verifying the matching ok/error result and help link. No real service will be removed to appease a test. Additional compute01 failure: install::session_restaging_uses_one_compound_command returned SetupRunning/exit 73 from a private fixture. The exact unchanged test passes five isolated runs (0.41–0.72s each). A bounded full-suite concurrency reproduction is running through build-slot before any proposed change to lock handling. No timeout changes. Verbatim first-run logs retained at ~/worktrees/j70-verify-first.log and ~/.cache/codex-scratch/j70/evidence/verify-first.log on each VM. They will be attached along with final results.
kayg commented 2026-11-01 17:42:42 +00:00
Author
Owner
Copy link

CLI host-state test repair committed and pushed as fecb98e9c. The unchanged test failed on dev-02 with exit 5; all 8 cli_envelope tests pass after checking the documented component_down envelope rather than assuming an unconfigured real host. Dedicated fixture tests retain exact exit-status coverage.

Final full fast gates at fecb98e9c are queued on both machines. The bounded unchanged SSH suite sweep is ahead of the compute01 gate. Only one target per VM is retained at ~/worktrees/j70/target; both will be deleted after the runs finish.

Attached are the complete first failed gate logs, restart before/after receipts, historical cross-build receipt, and compiler comparison. No restart timeout or production admission changes were made. Cross-host byte equality is not achieved: host-built Cargo dependency identities differ and affect LTO. docs/RUST.md records this limitation and the measurements.

CLI host-state test repair committed and pushed as fecb98e9c. The unchanged test failed on dev-02 with exit 5; all 8 cli_envelope tests pass after checking the documented component_down envelope rather than assuming an unconfigured real host. Dedicated fixture tests retain exact exit-status coverage. Final full fast gates at fecb98e9c are queued on both machines. The bounded unchanged SSH suite sweep is ahead of the compute01 gate. Only one target per VM is retained at ~/worktrees/j70/target; both will be deleted after the runs finish. Attached are the complete first failed gate logs, restart before/after receipts, historical cross-build receipt, and compiler comparison. No restart timeout or production admission changes were made. Cross-host byte equality is not achieved: host-built Cargo dependency identities differ and affect LTO. docs/RUST.md records this limitation and the measurements.
compute01-verify-first.log
320 KiB
dev02-verify-first.log
220 KiB
dev02-restart-baseline.log
14 KiB
dev02-restart-private-home.log
1.8 KiB
compiler-comparison.txt
9.8 KiB
dev02-historical-static.log
2.1 KiB
kayg commented 2026-11-01 17:50:51 +00:00
Author
Owner
Copy link

The unchanged SSH install suite passed all five concurrent sweeps on compute01: each round reports 94 passed; 0 failed; 2 ignored and ROUND_EXIT=0 (11.78–12.82s). Together with five isolated passes, the first-run restaging exit 73 remains intermittent and its cause is unconfirmed. No speculative lease/process-group change is being made. Complete sweep and isolated receipts attached. The final full gate will rerun this suite as well.

The unchanged SSH install suite passed all five concurrent sweeps on compute01: each round reports `94 passed; 0 failed; 2 ignored` and `ROUND_EXIT=0` (11.78–12.82s). Together with five isolated passes, the first-run restaging exit 73 remains intermittent and its cause is unconfirmed. No speculative lease/process-group change is being made. Complete sweep and isolated receipts attached. The final full gate will rerun this suite as well.
compute01-restaging-sweep.log
36 KiB
compute01-restaging-alone.log
845 B
kayg commented 2026-11-01 17:54:34 +00:00
Author
Owner
Copy link

dev-02 final full scripts/verify fast passed at fecb98e9c:

| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| SSH runtime test helper   | PASS         |
| real-binary fixture paths | PASS         |
| test                      | PASS         |
| SSH inode reuse fixture   | PASS         |
| shim process-per-user     | PASS         |
| cargo-deny                | PASS         |
| gate architecture         | PASS         |
| static builds             | PASS         |

The TypeScript/frontend checks also finished; the full log ends:

 39 pass
 0 fail
 59970 expect() calls
Ran 39 tests across 10 files. [2.39s]

VERIFY_EXIT=0

Full verbatim output attached. Removed ~/worktrees/j70/target (37 GB) on dev-02 and verified absence. compute01 final gate remains queued; no completion claim until that run ends.

dev-02 final full `scripts/verify fast` passed at fecb98e9c: ``` | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | SSH runtime test helper | PASS | | real-binary fixture paths | PASS | | test | PASS | | SSH inode reuse fixture | PASS | | shim process-per-user | PASS | | cargo-deny | PASS | | gate architecture | PASS | | static builds | PASS | ``` The TypeScript/frontend checks also finished; the full log ends: ``` 39 pass 0 fail 59970 expect() calls Ran 39 tests across 10 files. [2.39s] VERIFY_EXIT=0 ``` Full verbatim output attached. Removed ~/worktrees/j70/target (37 GB) on dev-02 and verified absence. compute01 final gate remains queued; no completion claim until that run ends.
dev02-verify-final.log
482 KiB
kayg commented 2026-11-01 18:12:02 +00:00
Author
Owner
Copy link

compute01 final fast gate admitted after approximately 29 minutes in the build-slot queue. HEAD=fecb98e9c1d2309a8c5ed34dfe3d906bac828e2e, log ~/worktrees/j70-verify.log. Main binary suite has passed 103 tests with 1 ignored; workspace integration tests continue. dev-02's full gate is already PASS, its full log attached above, and its target has been deleted. Only remaining J70 target is compute01 ~/worktrees/j70/target; it will be removed after validation. No uncommitted changes.

compute01 final fast gate admitted after approximately 29 minutes in the build-slot queue. HEAD=fecb98e9c1d2309a8c5ed34dfe3d906bac828e2e, log ~/worktrees/j70-verify.log. Main binary suite has passed 103 tests with 1 ignored; workspace integration tests continue. dev-02's full gate is already PASS, its full log attached above, and its target has been deleted. Only remaining J70 target is compute01 ~/worktrees/j70/target; it will be removed after validation. No uncommitted changes.
kayg commented 2026-11-01 18:26:53 +00:00
Author
Owner
Copy link

J70 finished. Branch gate-on-arm, head fecb98e9c1d2309a8c5ed34dfe3d906bac828e2e, pushed. Base a26615c50e76ed86f36d25aa094283cf840d5320.

Changes: portable static gate runs both target builds on either host, executes native CLI/key-memory probes, and explicitly skips foreign runtime probes with reasons. Architecture regressions cover dispatch, native selection and failed-build propagation. Documented restart trust requirements and compiler identity measurements. Fixed the overview integration test's assumption that a real build host has no stopped configured component.

Restart root cause: dev-02 HOME=0750 above worktrees=0775 caused production trusted-path admission to refuse its own test image. Changing only HOME to 0700 made the identical executable go from 13 passed / 5 failed to 18 passed / 0 failed. No timeout or production admission changes.

Size finding: historical 37bbda455 arm cross build reproduced 21,456,976 B, +8,704 B against the orchestrator's native 21,448,272 B receipt. Current-source x86 is 21,642,320 B on compute01 versus 21,649,296 B on dev-02 (+6,976 B). Host-built Cargo dependency identities differ and propagate through LTO. Differing native archive sections were debug-only; no embedded checkout/Cargo/Rustup/Nix paths found. Cross-host byte equality is not achieved; this compiler/dependency identity limitation is documented, with no release-profile or toolchain changes. Production size delta versus the measured base on the same host: 0 B. Current arm images: 19,794,072 B built on compute01; 19,792,184 B built on dev-02.

First full runs exposed an additional intermittent SSH fixture exit 73 on compute01. Cause remains unconfirmed: five isolated reruns, five concurrent full SSH install suites, and the final full gate all passed unchanged. No speculative lock weakening was applied. Failed first logs and rerun receipts remain attached above.

Both final full gates at the head completed with exit 0. Complete verbatim logs attached. Extracts below are copied verbatim from those logs.

dev02:

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| SSH runtime test helper   | PASS         |
| real-binary fixture paths | PASS         |
| test                      | PASS         |
| SSH inode reuse fixture   | PASS         |
| shim process-per-user     | PASS         |
| cargo-deny                | PASS         |
| gate architecture         | PASS         |
| static builds             | PASS         |
+---------------------------+--------------+

 39 pass
 0 fail
 59970 expect() calls
Ran 39 tests across 10 files. [2.39s]

VERIFY_EXIT=0

compute01:

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| SSH runtime test helper   | PASS         |
| real-binary fixture paths | PASS         |
| test                      | PASS         |
| SSH inode reuse fixture   | PASS         |
| shim process-per-user     | PASS         |
| cargo-deny                | PASS         |
| gate architecture         | PASS         |
| static builds             | PASS         |
+---------------------------+--------------+

 39 pass
 0 fail
 59970 expect() calls
Ran 39 tests across 10 files. [6.13s]

VERIFY_EXIT=0

Cleanup: removed J70 target directories on both VMs (37 GB arm, 35 GB compute01), plus temporary arm comparison ELF/archive copies. Text receipts retained. No J70 build/test run remains pending. Worktrees and branch retained for review; issue left open.

J70 finished. Branch `gate-on-arm`, head `fecb98e9c1d2309a8c5ed34dfe3d906bac828e2e`, pushed. Base `a26615c50e76ed86f36d25aa094283cf840d5320`. Changes: portable static gate runs both target builds on either host, executes native CLI/key-memory probes, and explicitly skips foreign runtime probes with reasons. Architecture regressions cover dispatch, native selection and failed-build propagation. Documented restart trust requirements and compiler identity measurements. Fixed the overview integration test's assumption that a real build host has no stopped configured component. Restart root cause: dev-02 HOME=0750 above worktrees=0775 caused production trusted-path admission to refuse its own test image. Changing only HOME to 0700 made the identical executable go from 13 passed / 5 failed to 18 passed / 0 failed. No timeout or production admission changes. Size finding: historical 37bbda455 arm cross build reproduced 21,456,976 B, +8,704 B against the orchestrator's native 21,448,272 B receipt. Current-source x86 is 21,642,320 B on compute01 versus 21,649,296 B on dev-02 (+6,976 B). Host-built Cargo dependency identities differ and propagate through LTO. Differing native archive sections were debug-only; no embedded checkout/Cargo/Rustup/Nix paths found. Cross-host byte equality is not achieved; this compiler/dependency identity limitation is documented, with no release-profile or toolchain changes. Production size delta versus the measured base on the same host: 0 B. Current arm images: 19,794,072 B built on compute01; 19,792,184 B built on dev-02. First full runs exposed an additional intermittent SSH fixture exit 73 on compute01. Cause remains unconfirmed: five isolated reruns, five concurrent full SSH install suites, and the final full gate all passed unchanged. No speculative lock weakening was applied. Failed first logs and rerun receipts remain attached above. Both final full gates at the head completed with exit 0. Complete verbatim logs attached. Extracts below are copied verbatim from those logs. dev02: ```text +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | SSH runtime test helper | PASS | | real-binary fixture paths | PASS | | test | PASS | | SSH inode reuse fixture | PASS | | shim process-per-user | PASS | | cargo-deny | PASS | | gate architecture | PASS | | static builds | PASS | +---------------------------+--------------+ 39 pass 0 fail 59970 expect() calls Ran 39 tests across 10 files. [2.39s] VERIFY_EXIT=0 ``` compute01: ```text +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | SSH runtime test helper | PASS | | real-binary fixture paths | PASS | | test | PASS | | SSH inode reuse fixture | PASS | | shim process-per-user | PASS | | cargo-deny | PASS | | gate architecture | PASS | | static builds | PASS | +---------------------------+--------------+ 39 pass 0 fail 59970 expect() calls Ran 39 tests across 10 files. [6.13s] VERIFY_EXIT=0 ``` Cleanup: removed J70 target directories on both VMs (37 GB arm, 35 GB compute01), plus temporary arm comparison ELF/archive copies. Text receipts retained. No J70 build/test run remains pending. Worktrees and branch retained for review; issue left open.
dev02-verify-final.log
482 KiB
compute01-verify-final.log
484 KiB
kayg commented 2026-11-01 22:55:04 +00:00
Author
Owner
Copy link

J81 start: branch repro-296, base origin/dev ad2b2c0466. Plan (estimate 3 hours including queued builds): (1) compare J70 section/symbol evidence, Cargo dependency identities and pinned toolchains on compute01 and dev-02; (2) reproduce both static architectures and isolate the earliest differing input; (3) apply the smallest correction or document a precise demonstrated limitation; (4) build one identical commit on both hosts, report SHA-256 and sizes, run scripts/verify fast, push this branch and delete both job target directories. No engine changes or unrelated tooling.

J81 start: branch repro-296, base origin/dev ad2b2c04666ca466107e80288ed47c5980c9a7ce. Plan (estimate 3 hours including queued builds): (1) compare J70 section/symbol evidence, Cargo dependency identities and pinned toolchains on compute01 and dev-02; (2) reproduce both static architectures and isolate the earliest differing input; (3) apply the smallest correction or document a precise demonstrated limitation; (4) build one identical commit on both hosts, report SHA-256 and sizes, run scripts/verify fast, push this branch and delete both job target directories. No engine changes or unrelated tooling.
kayg commented 2026-11-01 22:58:33 +00:00
Author
Owner
Copy link

J81 finding: inspected the exact Cargo source shipped by both VMs: cargo 1.99.0 commit 5f94df4789f005f9a352888e8355ffc645b7ed0e. In src/compiler/build_runner/compilation_files.rs, hash_rustc_version includes rustc verbose-version host for every CompileKind::Host unit, while compute_metadata recursively mixes ALL dependency c_metadata hashes (including host build scripts) into target c_metadata. Thus omitting host from target units alone is insufficient, and path remaps/SOURCE_DATE_EPOCH cannot affect this input. Investigating the smallest safe control without replacing Cargo identities with a collision-prone package-name/version salt. Both baseline builds (both targets, same ad2b2c046 sources) are queued through build-slot, evidence in ~/.cache/codex-scratch/j81/evidence on each VM; the sole target is ~/worktrees/j81/target on each VM.

J81 finding: inspected the exact Cargo source shipped by both VMs: cargo 1.99.0 commit 5f94df4789f005f9a352888e8355ffc645b7ed0e. In src/compiler/build_runner/compilation_files.rs, hash_rustc_version includes rustc verbose-version host for every CompileKind::Host unit, while compute_metadata recursively mixes ALL dependency c_metadata hashes (including host build scripts) into target c_metadata. Thus omitting host from target units alone is insufficient, and path remaps/SOURCE_DATE_EPOCH cannot affect this input. Investigating the smallest safe control without replacing Cargo identities with a collision-prone package-name/version salt. Both baseline builds (both targets, same ad2b2c046 sources) are queued through build-slot, evidence in ~/.cache/codex-scratch/j81/evidence on each VM; the sole target is ~/worktrees/j81/target on each VM.
kayg commented 2026-11-01 23:06:22 +00:00
Author
Owner
Copy link

J81 finding: Rust 1.99.0 options.rs declares metadata as Vec using parse_list; repeated -Cmetadata values are appended (slot.extend), not replaced. Therefore supplying a fixed salt via target rustflags/CARGO_ENCODED_RUSTFLAGS leaves Cargo host metadata present. Exact Cargo-source analysis and this distinction are drafted in docs/RUST.md. Current x86 crossbeam-utils c_metadata is b35f2907267e2169 on compute01 and b790444f03dc4227 on dev-02, reproducing J70 identities 615b1674294a9fde / c11c495add14c870. dev-02 baseline builds passed: x86 21,732,208 B, SHA256 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091; arm 19,856,104 B, SHA256 e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398. compute01 comparison still running; full fast gate is running on dev-02. No release profile, Zig flags or identity policy changed.

J81 finding: Rust 1.99.0 options.rs declares metadata as Vec<String> using parse_list; repeated -Cmetadata values are appended (slot.extend), not replaced. Therefore supplying a fixed salt via target rustflags/CARGO_ENCODED_RUSTFLAGS leaves Cargo host metadata present. Exact Cargo-source analysis and this distinction are drafted in docs/RUST.md. Current x86 crossbeam-utils c_metadata is b35f2907267e2169 on compute01 and b790444f03dc4227 on dev-02, reproducing J70 identities 615b1674294a9fde / c11c495add14c870. dev-02 baseline builds passed: x86 21,732,208 B, SHA256 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091; arm 19,856,104 B, SHA256 e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398. compute01 comparison still running; full fast gate is running on dev-02. No release profile, Zig flags or identity policy changed.
kayg commented 2026-11-01 23:09:45 +00:00
Author
Owner
Copy link

J81 crash-safety checkpoint: pushed WIP bdab1efc5 on repro-296, documenting exact Cargo/Rust source and unsupported-flag boundary. Both targets built successfully on dev-02; compute01 baseline/control comparison and dev-02 scripts/verify fast remain in progress. Logs and control source: ~/.cache/codex-scratch/j81/evidence/{baseline.log,baseline-*.log,probe.log,probe/,verify.log}. Sole target per VM: ~/worktrees/j81/target. Final measurement receipts will be added and WIP amended before handoff; no pending run will be left.

J81 crash-safety checkpoint: pushed WIP bdab1efc5 on repro-296, documenting exact Cargo/Rust source and unsupported-flag boundary. Both targets built successfully on dev-02; compute01 baseline/control comparison and dev-02 scripts/verify fast remain in progress. Logs and control source: ~/.cache/codex-scratch/j81/evidence/{baseline.log,baseline-*.log,probe.log,probe/,verify.log}. Sole target per VM: ~/worktrees/j81/target. Final measurement receipts will be added and WIP amended before handoff; no pending run will be left.
kayg commented 2026-11-01 23:18:25 +00:00
Author
Owner
Copy link

J81 capacity warning: dev-02 free disk has fallen from 41 GB to 14 GB during the admitted fast gate (workspace test compilation). No J81 command is queued behind the 25 GB floor; this gate is already running. I will delete J81 targets after validation, and will not touch other jobs targets. Logs remain ~/.cache/codex-scratch/j81/evidence/verify.log. compute01 has completed x86 and is linking aarch64. New x86 comparison: native 21,729,392 B / 94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1; arm-host cross 21,732,208 B / 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091 (+2,816 B). Plain no_std control IR matches byte-for-byte; adding a no-op build script changes its symbol/module identity across hosts.

J81 capacity warning: dev-02 free disk has fallen from 41 GB to 14 GB during the admitted fast gate (workspace test compilation). No J81 command is queued behind the 25 GB floor; this gate is already running. I will delete J81 targets after validation, and will not touch other jobs targets. Logs remain ~/.cache/codex-scratch/j81/evidence/verify.log. compute01 has completed x86 and is linking aarch64. New x86 comparison: native 21,729,392 B / 94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1; arm-host cross 21,732,208 B / 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091 (+2,816 B). Plain no_std control IR matches byte-for-byte; adding a no-op build script changes its symbol/module identity across hosts.
kayg commented 2026-11-01 23:20:58 +00:00
Author
Owner
Copy link

J81 gate capacity failure: stopped the admitted dev-02 fast gate when free disk reached 4.4 GB (99% used), before ENOSPC could affect shared jobs. Captured the gate process tree, sent TERM then KILL only to that J81 subtree, and am deleting only ~/worktrees/j81/target/debug. Log retained as ~/.cache/codex-scratch/j81/evidence/verify-arm-disk-aborted.log with explicit cancellation reason and exit 143. Release artifacts remain for final same-commit hashes. Moving final gate to compute01 (over 150 GB free), rather than risking another arm disk exhaustion. No test failure is being represented as a pass.

J81 gate capacity failure: stopped the admitted dev-02 fast gate when free disk reached 4.4 GB (99% used), before ENOSPC could affect shared jobs. Captured the gate process tree, sent TERM then KILL only to that J81 subtree, and am deleting only ~/worktrees/j81/target/debug. Log retained as ~/.cache/codex-scratch/j81/evidence/verify-arm-disk-aborted.log with explicit cancellation reason and exit 143. Release artifacts remain for final same-commit hashes. Moving final gate to compute01 (over 150 GB free), rather than risking another arm disk exhaustion. No test failure is being represented as a pass.
kayg commented 2026-11-01 23:23:40 +00:00
Author
Owner
Copy link

J81 final documentation commit pushed: 8c07420607 on repro-296. docs/RUST.md records the exact pinned Cargo hash path, additive Rust metadata semantics, six controlled IR comparisons, four ELF hashes and section deltas. Both targets reproduce the no-op-build-script identity distinction; plain IR matches on both architectures. This documents the precise limitation of the pinned supported build recipe, not a universal impossibility claim. Byte equality is not achieved; changing Cargo or replacing its identities requires a separately reviewed disambiguation recipe. Production/build configuration delta is zero. Both VMs now build/check the identical final commit; compute01 runs the final full fast gate with -j4 and CARGO_INCREMENTAL=0. Final logs: ~/.cache/codex-scratch/j81/evidence/final.log on each VM. Both job target directories will be deleted after validation.

J81 final documentation commit pushed: 8c07420607ca005d9c1450134e0aa2aca333f37d on repro-296. docs/RUST.md records the exact pinned Cargo hash path, additive Rust metadata semantics, six controlled IR comparisons, four ELF hashes and section deltas. Both targets reproduce the no-op-build-script identity distinction; plain IR matches on both architectures. This documents the precise limitation of the pinned supported build recipe, not a universal impossibility claim. Byte equality is not achieved; changing Cargo or replacing its identities requires a separately reviewed disambiguation recipe. Production/build configuration delta is zero. Both VMs now build/check the identical final commit; compute01 runs the final full fast gate with -j4 and CARGO_INCREMENTAL=0. Final logs: ~/.cache/codex-scratch/j81/evidence/final.log on each VM. Both job target directories will be deleted after validation.
kayg commented 2026-11-01 23:24:53 +00:00
Author
Owner
Copy link

J81 same-commit static comparison completed at 8c07420607ca005d9c1450134e0aa2aca333f37d on both VMs. Both build commands exited 0; each final binary is byte-identical to that host's measured origin/dev source baseline (documentation-only slice delta: 0 B).

Host Target Bytes SHA-256
compute01 x86_64 21729392 94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1
dev-02 x86_64 21732208 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091
compute01 aarch64 19855816 69411ccaab083267b68e7d6d160abc682044fdd52611a6dea321c5b048934a75
dev-02 aarch64 19856104 e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398

Cross-host equality is false: dev-02 minus compute01 is +2816 B x86 and +288 B aarch64. Both images on both targets contain zero occurrences of the scanned checkout, Cargo, Rustup, target and Nix-store path prefixes. Section-map differences are in allocated code/constants/unwind tables; the attached comparison lists them.

The isolated IR experiment is decisive for the pinned Cargo recipe: the dependency-free no_std control matches across hosts for both architectures, adding only a no-op build script changes the target symbol/module identity, and adding -Cmetadata=j81-fixed leaves a mismatch on both. Exact source, commands, symbols and hashes are retained in the attached text evidence (including the exact pinned Cargo source and Rust flag parser).

Deleted dev-02 ~/worktrees/j81/target, verified absence; free disk recovered to 39 GB. The earlier arm fast-gate log is explicitly marked disk-aborted/143, not a test pass. Final full fast gate is running at the same head on compute01, log ~/.cache/codex-scratch/j81/evidence/final.log, with the sole remaining target ~/worktrees/j81/target. No completion claim until it ends.

J81 same-commit static comparison completed at `8c07420607ca005d9c1450134e0aa2aca333f37d` on both VMs. Both build commands exited 0; each final binary is byte-identical to that host's measured origin/dev source baseline (documentation-only slice delta: 0 B). | Host | Target | Bytes | SHA-256 | | --- | --- | ---: | --- | | compute01 | x86_64 | 21729392 | `94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1` | | dev-02 | x86_64 | 21732208 | `19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091` | | compute01 | aarch64 | 19855816 | `69411ccaab083267b68e7d6d160abc682044fdd52611a6dea321c5b048934a75` | | dev-02 | aarch64 | 19856104 | `e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398` | Cross-host equality is false: dev-02 minus compute01 is +2816 B x86 and +288 B aarch64. Both images on both targets contain zero occurrences of the scanned checkout, Cargo, Rustup, target and Nix-store path prefixes. Section-map differences are in allocated code/constants/unwind tables; the attached comparison lists them. The isolated IR experiment is decisive for the pinned Cargo recipe: the dependency-free no_std control matches across hosts for both architectures, adding only a no-op build script changes the target symbol/module identity, and adding `-Cmetadata=j81-fixed` leaves a mismatch on both. Exact source, commands, symbols and hashes are retained in the attached text evidence (including the exact pinned Cargo source and Rust flag parser). Deleted dev-02 `~/worktrees/j81/target`, verified absence; free disk recovered to 39 GB. The earlier arm fast-gate log is explicitly marked disk-aborted/143, not a test pass. Final full fast gate is running at the same head on compute01, log `~/.cache/codex-scratch/j81/evidence/final.log`, with the sole remaining target `~/worktrees/j81/target`. No completion claim until it ends.
j81-comparison.txt
3.7 KiB
j81-evidence-text.tgz
543 KiB
kayg commented 2026-11-01 23:51:18 +00:00
Author
Owner
Copy link

J81 progress checkpoint: branch/head unchanged and fully pushed at 8c07420607; no uncommitted implementation work. Final same-commit static builds passed on both hosts and both targets with the hashes recorded above. Final compute01 full fast gate has passed CLI compile-out, clippy, SSH runtime helper and real-binary fixture checks; main binary tests passed 119/0 failed/1 ignored, and workspace server authentication/socket tests are still running. No completion claim. Log remains ~/.cache/codex-scratch/j81/evidence/final.log; only remaining target is compute01 ~/worktrees/j81/target. dev-02 target is already deleted and evidence retained/attached. I remain blocked on the running gate in this turn and will clean the last target after it exits.

J81 progress checkpoint: branch/head unchanged and fully pushed at 8c07420607ca005d9c1450134e0aa2aca333f37d; no uncommitted implementation work. Final same-commit static builds passed on both hosts and both targets with the hashes recorded above. Final compute01 full fast gate has passed CLI compile-out, clippy, SSH runtime helper and real-binary fixture checks; main binary tests passed 119/0 failed/1 ignored, and workspace server authentication/socket tests are still running. No completion claim. Log remains ~/.cache/codex-scratch/j81/evidence/final.log; only remaining target is compute01 ~/worktrees/j81/target. dev-02 target is already deleted and evidence retained/attached. I remain blocked on the running gate in this turn and will clean the last target after it exits.
kayg commented 2026-11-02 00:03:59 +00:00
Author
Owner
Copy link

J81 final report. Branch repro-296, head 8c07420607ca005d9c1450134e0aa2aca333f37d, pushed. Base ad2b2c04666ca466107e80288ed47c5980c9a7ce. One atomic documentation-only commit; worktree clean.

Byte equality is not achieved. docs/RUST.md documents the precise limitation of the pinned supported Cargo recipe: Cargo 1.99.0 hashes the compiler host into host build-tool identities and recursively incorporates those identities into target c_metadata; Rust appends repeated -Cmetadata values, so a fixed additional flag cannot erase Cargo's salt. Both target controls match without a build script, differ with only a no-op build script, and still differ with a fixed metadata flag. This is not a universal impossibility claim: achieving equality needs a Cargo change or a reviewed replacement identity recipe that preserves disambiguation. No toolchain, release-profile or production-code change was made.

Both VMs built the identical final commit and confirmed each artifact is byte-identical to that host's measured origin/dev source baseline. Per-slice static delta: 0 B for both architectures on both hosts.

Host Target Bytes SHA-256
compute01 x86_64 21729392 94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1
dev-02 x86_64 21732208 19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091
compute01 aarch64 19855816 69411ccaab083267b68e7d6d160abc682044fdd52611a6dea321c5b048934a75
dev-02 aarch64 19856104 e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398

Cross-host differences: +2816 B x86, +288 B aarch64 (dev-02 minus compute01). ELF section/symbol maps and all controlled IR hashes are retained in the evidence attached above and the final bundles attached here.

Final full scripts/verify fast passed on compute01 at the head, with CARGO_BUILD_JOBS=4 and CARGO_INCREMENTAL=0. Verbatim summary and final output:

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| SSH runtime test helper   | PASS         |
| real-binary fixture paths | PASS         |
| test                      | PASS         |
| SSH inode reuse fixture   | PASS         |
| shim process-per-user     | PASS         |
| cargo-deny                | PASS         |
| gate architecture         | PASS         |
| static builds             | PASS         |
+---------------------------+--------------+


 39 pass
 0 fail
 59970 expect() calls
Ran 39 tests across 10 files. [8.01s]

VERIFY_EXIT=0

FINAL_RUN_EXIT=0

Complete verbatim gate output is attached as j81-compute01-final.log, with the full text receipts and control sources in both bundles. The earlier dev-02 gate was explicitly stopped at 4.4 GB free disk; its retained log is marked disk-aborted/143 and is not represented as a pass. Upstream source license notices accompany the source evidence in the final dev-02 bundle.

Cleanup verified: deleted ~/worktrees/j81/target on both VMs and the temporary ELF comparison copies. Both orphan sweeps returned J81_ORPHAN_SWEEP=none; no J81 build/test/server run remains pending. Worktrees, branch and small text evidence retained for review. Issue left open. No merge or deployment.

J81 final report. Branch `repro-296`, head `8c07420607ca005d9c1450134e0aa2aca333f37d`, pushed. Base `ad2b2c04666ca466107e80288ed47c5980c9a7ce`. One atomic documentation-only commit; worktree clean. Byte equality is **not achieved**. `docs/RUST.md` documents the precise limitation of the pinned supported Cargo recipe: Cargo 1.99.0 hashes the compiler host into host build-tool identities and recursively incorporates those identities into target `c_metadata`; Rust appends repeated `-Cmetadata` values, so a fixed additional flag cannot erase Cargo's salt. Both target controls match without a build script, differ with only a no-op build script, and still differ with a fixed metadata flag. This is not a universal impossibility claim: achieving equality needs a Cargo change or a reviewed replacement identity recipe that preserves disambiguation. No toolchain, release-profile or production-code change was made. Both VMs built the identical final commit and confirmed each artifact is byte-identical to that host's measured origin/dev source baseline. Per-slice static delta: **0 B for both architectures on both hosts**. | Host | Target | Bytes | SHA-256 | | --- | --- | ---: | --- | | compute01 | x86_64 | 21729392 | `94b5b4d210213a990ca43b146c9d65fec983181c75ec2d6b6a969fd8108956e1` | | dev-02 | x86_64 | 21732208 | `19d80905d835af9d8ca386bccb33a158b8c4ee59d8479c827a451f39a0887091` | | compute01 | aarch64 | 19855816 | `69411ccaab083267b68e7d6d160abc682044fdd52611a6dea321c5b048934a75` | | dev-02 | aarch64 | 19856104 | `e4259d9284f6e54ff8c9586f5d8c3e36a297c0aeed88a8650b4f1b6daf410398` | Cross-host differences: +2816 B x86, +288 B aarch64 (dev-02 minus compute01). ELF section/symbol maps and all controlled IR hashes are retained in the evidence attached above and the final bundles attached here. Final full `scripts/verify fast` passed on compute01 at the head, with CARGO_BUILD_JOBS=4 and CARGO_INCREMENTAL=0. Verbatim summary and final output: ```text +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | SSH runtime test helper | PASS | | real-binary fixture paths | PASS | | test | PASS | | SSH inode reuse fixture | PASS | | shim process-per-user | PASS | | cargo-deny | PASS | | gate architecture | PASS | | static builds | PASS | +---------------------------+--------------+ 39 pass 0 fail 59970 expect() calls Ran 39 tests across 10 files. [8.01s] VERIFY_EXIT=0 FINAL_RUN_EXIT=0 ``` Complete verbatim gate output is attached as `j81-compute01-final.log`, with the full text receipts and control sources in both bundles. The earlier dev-02 gate was explicitly stopped at 4.4 GB free disk; its retained log is marked disk-aborted/143 and is not represented as a pass. Upstream source license notices accompany the source evidence in the final dev-02 bundle. Cleanup verified: deleted `~/worktrees/j81/target` on **both** VMs and the temporary ELF comparison copies. Both orphan sweeps returned `J81_ORPHAN_SWEEP=none`; no J81 build/test/server run remains pending. Worktrees, branch and small text evidence retained for review. Issue left open. No merge or deployment.
j81-compute01-final.log
515 KiB
j81-compute01-text.tgz
580 KiB
j81-dev02-text-final.tgz
548 KiB
kayg commented 2026-11-02 11:26:08 +00:00
Author
Owner
Copy link

Owner decision (kayg, 2026-11-02): all release builds, both architectures, run on x86 build hosts (compute01 cross-builds aarch64, as today). Anyone can then reproduce a release byte-for-byte on any x86_64 Linux machine; arm build hosts keep running day-to-day gates only. Record in docs/RUST.md and the release docs.

**Owner decision (kayg, 2026-11-02): all release builds, both architectures, run on x86 build hosts** (compute01 cross-builds aarch64, as today). Anyone can then reproduce a release byte-for-byte on any x86_64 Linux machine; arm build hosts keep running day-to-day gates only. Record in docs/RUST.md and the release docs.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
j470b-server-mesh-rejoin
mesh-compat-382
j444b-d381-key-wins-flood
j461b-macos-utun-reader-thread
j475b-remove-host-motion
fix-reconnect-new-link
j481-d380
dev
j479-mesh-soak
fix/j476-host-facts
j475-motion-pill-split
j470-mac-node-join
root-upgrade-d376
j474-review-merge
latency-targets-120hz
nightly-suite-shards
gotatun-pin-j469
macos-utun-reader-thread
fix-flood-echo-stall
wip/j416
motion-fast-ease-out
mesh-addhost-j259
takeover-close-terminals-d368
kbd-instant-followups
fix-trust-key-double-click
nightly-suite-partial-report
appwin-icons-j207
rusqlite-slice1
swipe-content-first-d374
mesh-mtu-381
addhost-instant-open
glass-contrast-d365-mock
glass-contrast-d365
signout-followups-442
fix-startup-toast-red
research-gotatun-matrix-j317
themes-sync-j304
fix-load-flakes
fix-nix-zig-patch
fix-compute01-browser-baseline
ratchet-size-baseline-419
predict-auto-d130
revoke-coverage-d367
nightly-suite-queue-timeout
nightly-browser-suite
fix-rustsec-2026-0459
fonts-ioskeley-obel
log-retention-d375
fix-kbd-no-anim-427
fix-dev-red-specs-2
takeover-proof-305-j205
wip/j410-server-pump
ui-consistency-d372
fix-osc52-clipboard
libghostty-vt-1.0
forget-retired-host-d369
j404-merge-426
fix-setup-lock-race
fix-touch-select-426
revoke-access-report-ui
fix-predictive-echo-specs
fix-selinux-agent-exec
wip/j395-checkpoint2
wip/j395
j391-review-scratch
ui-consistency-j195-deferred-rebased
fix-rustc-paths-421
decisions-wording-341
fix-install-suite-red
fix-dev-rtt-specs
fix-pending-removal-align
fix-privileged-skip
fix-install-suite-404
review-j368
mesh-dns-j254
revoke-access-report-d356b
revoke-privilege-round5
fix-removal-confirm
review-j360
gotatun-rotation-367
deps-refresh-340
review-j354
review-J351
review-j350
addhost-takeover-d359
revoke-report-items-347
revoke-end-sessions-d362
review-j348
size-mesh-ssh-346
size-trim-337
revoke-access-diff-d356
j259-pre-rebase-b06c5a6
j305-ended-at-start-spec
addhost-sudo-consent
fix-layout-real-specs-411
fix-shell-open-411
fix-console-hosts-373
fix-shell-exit
fuzz-host-name-353
fix-host-details-410
macos-filter-j197
research-gotatun-tricks
revoke-p3-j216l
stale-state-pill
revoke-followups-j216k
revoke-progress-j216e
backup/macos01-20261106/lume/j197
backup/macos01-20261106/lume/appwin-icons-j207-stash0
backup/macos01-20261106/claude/stash-0
backup/macos01-20261106/claude/stash-1
backup/macos01-20261106/claude/J80-macos-s1
backup/macos01-20261106/claude/J80
backup/macos01-20261106/claude/J80-slice1-20261102
backup/macos01-20261106/claude/J16b-appwin-on-dev
backup/macos01-20261106/claude/appwin-launch-idle-146
backup/macos01-20261106/claude/J80-s1-run
wip/j259
trust-285-p3c
runtime-lock-j216f
mesh-ssh-p3-j175
tun-proof-p3b
trust-285-p3
terminal-polish-j271
tun-proof-p3
ssh-install-flakes-j216i
fuzz-fixes-j264
trust-285-a
trust-285-b
restart-oversized-fixture-j198
backup/macos01-20261106/lume/appwin-icons-j207
restart-bound-j216h
up-preflight-p3-j198
backup/macos01-20261106/lume/j251
up-preflight-j198
wip/up-preflight-j198-j245-p2
wip/up-preflight-j198-j245-p1
agent-fsync-offloop-j196
wip/up-preflight-j198-j245
settings-offscreen-j231
wip/up-preflight-j198-j226-inspection
wip/j103f
themes-j206
wip/up-preflight-j198-j226
flakes-j216c
backup/macos01-20261106/claude/j197
flakes-j216b
backup/macos01-20261106/claude/j197-red
wip/up-preflight-j198-review
flakes-j216
backup/macos01-20261106/claude/appwin-icons-j207
engineering-j208b
appwin-mock-j209
toast-stack-j165
engineering-j208
decisions-d351
decisions-d350
decisions-d349
appearance-presets-mock-j129
decisions-d348
docs-audit-spec
doc-coverage-baseline-raise
decisions-d347
decisions-d346b
decisions-d346
j88h-profile
ui-consistency-j195-deferred
ui-consistency-j195
engineering-practices-j194
preferred-login-ui-j166
signin-pools-263-j174
macos-node-265-s1
decisions-d345
decisions-d344
proxy-cgroup-j164
verify-quick-fuzz-329
proxy-start-io-j168
timing-flakes-j163
up-263-s1
mcp-events-j147
verify-quick-fuzz-j158
proxy-p3-j159
agent-resource-j157-detect
wip/j103d-final
remote-cancel-hang-j155
proxy-traefik-281-j62d
j149-verify-quick
proxy-traefik-281-j62b
j150-drop-unnamed
wip/j79d-red-r4
wip/j103d-d332
rootless-toast-j142
wip/j103d-head
wip/j103d
j125c-mcp
macos-build-j136
backup/macos01-20261106/claude/j136
wip/j79d-red-cli
wip/j79d-red-server
j121-link-liveness
j118c-tracer-failures
wip/j79c-head
release-desktop-j131
j124c-signed-macos
wip/j79c-red-p3
wip/j79c-red-p23
wip/j79c-red-p22
merge-j128-check
perf-deferred-j128
sidebar-toggle-260
j127-shortcut-changed-dot
j125-mcp
j118-tracer-failures
j126-account-summary
settings-sections-mock-j92
j120-pinch-lock
sidebar-292
backup/macos01-20261106/claude/j120
taste-2026-11-03
row-actions-242-215
backup/macos01-20261106/claude/stash-0-untracked
backup/macos01-20261106/claude/J80-265-s1-run
backup/macos01-20261106/claude/J80-macos-s1-20261102
j112-canvas-commands
ui-no-overshoot-scroll
fix-recovery-restart
backup/macos01-20261106/claude/stash-1-untracked
taste-2026-11-02
size-baseline-ratchet
machine-cards-structure-j108
j109-boot-copy-mockups
fix-dev-red
wip/up-263-s1
j95b-settings-rows
proxy-traefik-281
j95-option-rows
docs-human-j91
docs-keys-palette
fix-flood-1013-270-j38d
mesh-ssh-s1
mesh-ssh-s0
addhost-d291-tunnel
test-stability-load
mesh-relay-220
repro-296
shim-thp-off-24
mesh-pins-258-v2
mesh-dns
signin-d311-merged
docs-pages
ui-size-283
fix-ui-batch-specs
wip/j12b-edge-update
gate-on-arm
fix-rename-sync-293
trust-285-j60b
fix-focus-ring-clip
integ-ui-1101
edge-update-v2
signin-d311
taste-2026-11-01
layout-per-tab-250
ui-approved-splitopt-224
ui-approved-splitopt-224-on-dev
icon-text-centre-282
ui-approved-189-212-216
research-reverse-proxy
transfer-budget-320
review-mesh-pins-258-r2
sidebar-pill-259
review-mesh-relay-220
wip/j12-edge-update-v2
nix-package
status-crash-loop-237
fix-dev-test-failures
fix-fixture-socket-path
wip/mesh-relay-220-pre-dev-advance-rebase
fix-canvas-side-clip
research-head-to-head
frosted-real-256
review-mesh-pins-258
wip/mesh-relay-220-pre-final-rebase
review-layout-per-tab-250-r2
broad-suite-triage
nix-devshell
review-status-crash-loop-237-r3
wip/nix-devshell-pre-p3
taste-frosted-real
wip/code-serve-demo-2026-10-30
wip/code-serve-glass-2026-10-30
wip/code-serve-clusters-2026-10-30
wip/code-merge-release94-dev-2026-10-30
wip/rproxy-delta-scratch-2026-10-30
wip/stale-note-257-staged-2026-10-30
wip/d241-size-staged-2026-10-30
review-broad-suite-207
wip/mesh-pins-258-pre-split
wip/mesh-relay-220-pre-rebase
mesh-measure-220
plan-211-update
review-edge-autoupdate-r3
perf24-reconnect
review-stale-note-r2
release-prod-live
d241-shell-marker
review-mesh-pins-server-258
review-revocation-r3
review-setup-token-r2
review-sync-journal-r3
wip/review-mesh-pins-258-prune-review-1995fee46
frosted-scroll-256
sync-journal-r3
revocation-per-account-wip
review-tokio-asyncfd-0386
review-sync-p3s
wip/stash-8by64-stash0-2026-10-30
mesh-drag-r2
setup-token-data-wip
release-9-4
auth-busy-1013
mesh-view-overlap
sync-p3s
review-auth-busy-wake
review-readd-r8-wip
mesh-drag-bounds
audit-auth-93
app-windows-stream-start
review-authority-cap-r3
app-windows-r7-measure
readd-minimal-r6
readd-minimal-r5
app-windows-r7-soak
readd-minimal-r4
app-windows-r7
app-windows-r6
readd-minimal-r3
readd-minimal-r2
readd-minimal
merge-release8-dev
review-busy-page-liveness-r2
review-busy-page-liveness
mesh-view-trim
review-authority-read-retry-r2
review-remove-uninstall-r2
release-retry-reattach
release-retry
review-remove-uninstall
review-authority-read-retry
app-windows-ship
mesh-autoswitch-p2
mesh-ssh
mesh-relay
mesh-macos-node
screen-and-control
session-labels
review-real-logins-r5
review-real-logins-r4
review-real-logins-r3
shim-memory-r2
review-real-logins-r2
shim-memory
shadow-addhost-instant
shadow-real-logins
shadow-same-session-flood
shadow-prod-perf-fix
hosts-from-server
app-windows-real
app-windows-transport
release-prod-6w
mesh-policy-verify
quiet-close-toast
dev-merge-r4
review-mesh-agent-transport
review-app-windows-transport
settings-roundtrip
review-release-pipeline-r4
mesh-agent-transport
ended-strip-fix
review-audit-ca-update-d244
settings-rows
phone-width
signout-clear
fix-restart-tests
dev-merge-rp
signer-hardening
review-audit-ca-update
terminal-fidelity
review-audit-ca-update-wip
mesh-integrate
review-mesh-agent-transport-515d3712
review-mesh-dataplane-feec1988
mesh-dataplane
signout
research-canvas-node
wip/review-renderer-cost
wip/review-terminal-fidelity
padlock-glyph
review-release-pipeline-2e95b689
phone-polish
sidebar-polish
inbox-generic
review-mesh-dataplane-378d4055
prod-batch
ui-inbox
ui-ca-unlock
ui-machine-sections
review-mesh-dataplane-a5f744fb
tls-session-prod
session-bounds-prod
fix-dead-session
session-bounds
tls-session
palette-polish
shortcuts-d222
research-appmeta
phone-gutter
ui-centre-button
wip/ui-centre-button
wip/tls-session
review-addhost-breakit
ui-platform
renderer-cost
remote-agents
mesh-api
ssh-bootstrap
prod-hotfix2
prod-fixes
review-remote-agents
review-mesh-dataplane
session-life
research-browser
review-mesh-api
research-capture
review-mesh-api-round3
review-mesh-api-round1
gate-stability
review-gate-stability
review-relay-direct
review-mesh-api-findings
proto-sessions
font-picker
proto-addhost
proto-windows
mesh-core
latency-gate
research-bsd
research-discovery
keytray-mods
mock-core
ui-tweaks
fix-motion-sel
fix-motion
integrate2
fix-robust
t-web
t-server
vt-ghostty
fix-snap-echo
t-web-auth
review-vt2
t-agent
lane-protocol
lane-shim
lane-auth
s1-base
s1-contracts
s1-workspace
No results found.
Labels
Clear labels   
launch-qa
Findings from launch QA.

No labels
launch-qa
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
kayg/infraternal#296
No description provided.