cli: a config file ends with end, so a paste closes itself #1768

Merged
jaylouisw merged 1 commit from rev/1117-0130 into main 2026-11-17 02:22:31 +00:00
Owner

show running-config now ends with end, the line a Cisco config file closes with, so a pasted
backup finishes at privileged EXEC instead of leaving the operator at (config)#. end typed at
# (where a paste usually starts) is accepted and silent rather than answering
% Unrecognized command: end.

Evidence: two vendor samples (IOS 11.2 command reference, ISR 4400 configuration guide) both end
with end and differ only in how many bare ! lines sit above it, so only end was adopted.
research/running-config-end.md records what is inferred (accepting end at EXEC) and what stays
unresolved (the ! count, and the Current configuration : N bytes byte count).

test/running-config-end.test.js - 14 assertions. Full suite: 299 suites, 10535 assertions, 0 failed.

`show running-config` now ends with `end`, the line a Cisco config file closes with, so a pasted backup finishes at privileged EXEC instead of leaving the operator at `(config)#`. `end` typed at `#` (where a paste usually starts) is accepted and silent rather than answering `% Unrecognized command: end`. Evidence: two vendor samples (IOS 11.2 command reference, ISR 4400 configuration guide) both end with `end` and differ only in how many bare `!` lines sit above it, so only `end` was adopted. `research/running-config-end.md` records what is inferred (accepting `end` at EXEC) and what stays unresolved (the `!` count, and the `Current configuration : N bytes` byte count). `test/running-config-end.test.js` - 14 assertions. Full suite: 299 suites, 10535 assertions, 0 failed.
cli: a config file ends with end, so a paste closes itself
All checks were successful
CI / Classify the change (pull_request) Successful in 7s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 17s
CI / Server smoke test (pull_request) Successful in 7s
CI / Docker build (pull_request) Successful in 22s
CI / Test (Node 20) (pull_request) Successful in 48s
CI / Test (Node 18) (pull_request) Successful in 56s
CI / Test (Node 22) (pull_request) Successful in 53s
CI / Arm auto-merge (pull_request) Successful in 5s
098dd60e1d
`show running-config` stopped at the last configured value. The output replayed, but a pasted
backup left the session at `(config)#` - the paste had finished and nothing said so - and `end`
typed at privileged EXEC, where a paste usually starts, answered `% Unrecognized command: end`.

Two Cisco samples end the file with `end` and differ only in how many bare `!` lines sit above it
(one in the IOS 11.2 command reference, two in the ISR 4400 guide), so only `end` is copied.
`end` at `#` is now accepted in silence; that half is inference from the printer's needs, and the
research note says so.
jaylouisw force-pushed rev/1117-0130 from 098dd60e1d
All checks were successful
CI / Classify the change (pull_request) Successful in 7s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 17s
CI / Server smoke test (pull_request) Successful in 7s
CI / Docker build (pull_request) Successful in 22s
CI / Test (Node 20) (pull_request) Successful in 48s
CI / Test (Node 18) (pull_request) Successful in 56s
CI / Test (Node 22) (pull_request) Successful in 53s
CI / Arm auto-merge (pull_request) Successful in 5s
to 25a5806189
Some checks failed
CI / Classify the change (pull_request) Successful in 8s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 16s
CI / Server smoke test (pull_request) Successful in 7s
CI / Docker build (pull_request) Successful in 22s
CI / Test (Node 20) (pull_request) Failing after 56s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Test (Node 22) (pull_request) Successful in 1m3s
CI / Arm auto-merge (pull_request) Successful in 3s
2026-11-17 02:01:39 +00:00
Compare
Author
Owner

Red on one leg only — and the CI logs are readable now, so here are the two assertions by name.

Full log is at /jaylouisw/packetforge/actions/runs/4505/jobs/2/logs fetched with Accept: text/plain (36564 bytes). The REST /actions/... API still 404s on this build, so logs are web-only.

Head 25a58061, CI / Test (Node 20):

✗ link-roundtrip        0 passed, 1 failed   1910 ms (no summary line)
✗ state-payload-shape  20 passed, 1 failed   4015 ms
    ✗ a state payload for 200 cables costs 66.1 ms (budget 60)  [66.12]
299 suite(s) in 46.8 s — 10498 assertions passed, 2 failed

Node 18 and Node 22 are green on the same commit.

Against main (82e33b70, run 4498, 02:01:47 — same host, same 10-way concurrency): link-roundtrip 36 passed 1594 ms, state-payload-shape 21 passed 2636 ms, 298 suites, 0 failed. Main is green on both suites two minutes before this run, so this is not repo-wide breakage.

Locally on this head (v24.11.0, three runs each): state-payload-shape 21/21 at 5.5 / 5.2 / 5.5 ms; link-roundtrip 36/36 three times. A 60 ms budget that measures 5 ms unloaded and 66 ms inside a 10-way, three-container CI burst is load-shaped, not code-shaped — and neither suite touches anything this PR changes (docs/CLI.md, engine/cli.js, research/running-config-end.md, test/running-config-end.test.js). link-roundtrip's 0 passed, 1 failed (no summary line) at 1910 ms means the suite process died before printing its summary; the runner's ceiling is 300 s, so it is not a timeout.

Caveat: this is the second consecutive head failing only Node 20, so it may be reproducible rather than luck.

Why it is still open: Jay's standing rule is a PR never merges on a red leg, and mergeable is true. The rerun endpoint is 404 on this Forgejo build and the auto-rebase workflow rebases without producing a new head, so only a push re-triggers CI. Owner's call: push an empty commit for a fresh run, or raise the two wall-clock budgets to CI-realistic values.

Same three-PR shape, for the record: #1739 is deterministic (device-aliases — B1LDCD finds both shelves [3] against a hardcoded 2, plus palette-search-keys same count), #1749 is its own new suite's 2,000-frame budget measured at 4051 ms in CI. — surgery watchdog

**Red on one leg only — and the CI logs are readable now, so here are the two assertions by name.** Full log is at `/jaylouisw/packetforge/actions/runs/4505/jobs/2/logs` fetched with `Accept: text/plain` (36564 bytes). The REST `/actions/...` API still 404s on this build, so logs are web-only. Head `25a58061`, `CI / Test (Node 20)`: ``` ✗ link-roundtrip 0 passed, 1 failed 1910 ms (no summary line) ✗ state-payload-shape 20 passed, 1 failed 4015 ms ✗ a state payload for 200 cables costs 66.1 ms (budget 60) [66.12] 299 suite(s) in 46.8 s — 10498 assertions passed, 2 failed ``` Node 18 and Node 22 are green on the same commit. **Against main (`82e33b70`, run 4498, 02:01:47 — same host, same 10-way concurrency):** `link-roundtrip 36 passed 1594 ms`, `state-payload-shape 21 passed 2636 ms`, 298 suites, 0 failed. Main is green on both suites two minutes before this run, so this is not repo-wide breakage. **Locally on this head (v24.11.0, three runs each):** `state-payload-shape` 21/21 at 5.5 / 5.2 / 5.5 ms; `link-roundtrip` 36/36 three times. A 60 ms budget that measures 5 ms unloaded and 66 ms inside a 10-way, three-container CI burst is load-shaped, not code-shaped — and neither suite touches anything this PR changes (`docs/CLI.md`, `engine/cli.js`, `research/running-config-end.md`, `test/running-config-end.test.js`). `link-roundtrip`'s `0 passed, 1 failed (no summary line)` at 1910 ms means the suite process died before printing its summary; the runner's ceiling is 300 s, so it is not a timeout. Caveat: this is the second consecutive head failing *only* Node 20, so it may be reproducible rather than luck. **Why it is still open:** Jay's standing rule is a PR never merges on a red leg, and `mergeable` is true. The rerun endpoint is 404 on this Forgejo build and the auto-rebase workflow rebases without producing a new head, so only a push re-triggers CI. Owner's call: push an empty commit for a fresh run, or raise the two wall-clock budgets to CI-realistic values. Same three-PR shape, for the record: #1739 is deterministic (`device-aliases` — `B1LDCD finds both shelves [3]` against a hardcoded 2, plus `palette-search-keys` same count), #1749 is its own new suite's 2,000-frame budget measured at 4051 ms in CI. — surgery watchdog
jaylouisw force-pushed rev/1117-0130 from 25a5806189
Some checks failed
CI / Classify the change (pull_request) Successful in 8s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 16s
CI / Server smoke test (pull_request) Successful in 7s
CI / Docker build (pull_request) Successful in 22s
CI / Test (Node 20) (pull_request) Failing after 56s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Test (Node 22) (pull_request) Successful in 1m3s
CI / Arm auto-merge (pull_request) Successful in 3s
to 34744fc906
All checks were successful
CI / Classify the change (pull_request) Successful in 8s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 18s
CI / Server smoke test (pull_request) Successful in 7s
CI / Docker build (pull_request) Successful in 24s
CI / Test (Node 20) (pull_request) Successful in 51s
CI / Test (Node 18) (pull_request) Successful in 1m1s
CI / Test (Node 22) (pull_request) Successful in 57s
CI / Arm auto-merge (pull_request) Successful in 15s
2026-11-17 02:18:32 +00:00
Compare
Sign in to join this conversation.
No reviewers
No milestone
No project
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
jaylouisw/packetforge!1768
No description provided.