wai dev's test: the stand-in room service listens without a reverse lookup #663

Merged

dcharlot merged 2 commits from fix/fake-stack-reverse-lookup into main 2026-11-06 13:15:50 -05:00

Conversation 0 Commits 2 Files changed 1 +10 -1

dcharlot commented 2026-11-05 18:25:59 -05:00

Owner

Copy link

Replaces #661, whose commit message and first body stated things I have since measured to be wrong.

tests/cli_studio.rs failed on every run on one macOS runner and passed on every other, with "the room service's clock on the relay: not within 30 s". Across 22 runs of that job on 5 November it was 3 of 3 on that runner and 19 of 19 passing elsewhere.

Cause. The stand-in room service starts an HTTPServer, whose server_bind calls socket.getfqdn before the server listens. On that host, under the runner's Python 3.15.2, getfqdn("127.0.0.1") takes 35 s, past the 30 s wai dev waits. The stand-in relay opens a plain socket, so it was up in about 0.1 s and only the room service stalled. A second macOS host shows the same 35 s; two others answer the same call in milliseconds on the same Python version. Why those hosts are slow is not established, and this change doesn't depend on it.

Nothing in the fixture reads the name, so it no longer asks for it.

Tested, on that host and Python, with a stack dump armed:

Stand-in	Time to answer	Where it was
Unchanged	35.5 s	socket.getfqdn, from http/server.py's server_bind
This change	0.12 s over three rounds

On another host, a socket.getfqdn planted to stall for 40 s fails the test with the same message at 30.18 s, and passes with this change in 0.30 s.

CI checks out each PR's own head, so a PR already open keeps the old fixture until it carries this commit. The same commit is on #640, #656 and #658.

Changes the test fixture only, so there is no changelog entry.

🤖 Generated with Claude Code

Replaces #661, whose commit message and first body stated things I have since measured to be wrong. `tests/cli_studio.rs` failed on every run on one macOS runner and passed on every other, with "the room service's clock on the relay: not within 30 s". Across 22 runs of that job on 5 November it was 3 of 3 on that runner and 19 of 19 passing elsewhere. **Cause.** The stand-in room service starts an `HTTPServer`, whose `server_bind` calls `socket.getfqdn` before the server listens. On that host, under the runner's Python 3.15.2, `getfqdn("127.0.0.1")` takes 35 s, past the 30 s `wai dev` waits. The stand-in relay opens a plain socket, so it was up in about 0.1 s and only the room service stalled. A second macOS host shows the same 35 s; two others answer the same call in milliseconds on the same Python version. Why those hosts are slow is not established, and this change doesn't depend on it. Nothing in the fixture reads the name, so it no longer asks for it. **Tested, on that host and Python, with a stack dump armed:** | Stand-in | Time to answer | Where it was | |---|---|---| | Unchanged | 35.5 s | `socket.getfqdn`, from `http/server.py`'s `server_bind` | | This change | 0.12 s over three rounds | | On another host, a `socket.getfqdn` planted to stall for 40 s fails the test with the same message at 30.18 s, and passes with this change in 0.30 s. CI checks out each PR's own head, so a PR already open keeps the old fixture until it carries this commit. The same commit is on #640, #656 and #658. Changes the test fixture only, so there is no changelog entry. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

dcharlot added 1 commit 2026-11-05 18:25:59 -05:00

wai dev's test: the stand-in room service listens without a reverse lookup ... 576511d62b

tests/cli_studio.rs failed on every run on one macOS runner and passed on
every other runner, 3 of 3 against 19 of 19 on 5 November, with "the room
service's clock on the relay: not within 30 s".

The stand-in room service starts an HTTPServer, whose server_bind calls
socket.getfqdn before the server listens. On that host, under the runner's
Python 3.15.2, getfqdn("127.0.0.1") takes 35 s, past the 30 s wai dev waits, and
a stack dump of the stalled stand-in shows it there. The stand-in relay opens a
plain socket and answers in 0.1 s. Nothing here reads the name, so the fixture
no longer asks for it.

On that host and Python the room stand-in took 35.5 s and now takes 0.12 s. A
socket.getfqdn planted to stall for 40 s fails the test without this change,
with the same message, and passes with it.

Changelog: none

Co-Authored-By: Claude Opus 5.5 <[email protected]>

dcharlot referenced this pull request 2026-11-05 18:27:58 -05:00

wai dev's test: the stand-in room service listens without a reverse lookup #661

dcharlot added 1 commit 2026-11-06 07:16:43 -05:00

wai dev's test fixture: the change has no changelog entry, by its trailer ... 55e17b4f6a

The fixture commit before this one carried its no-entry trailer in a paragraph
of its own, so it was not in that message's trailer block and the changelog
check read none. This commit states it where the check reads it. It changes
no file.

Changelog: none
Co-Authored-By: Claude Opus 5.5 <[email protected]>

dcharlot merged commit e2774f7f83 into main 2026-11-06 13:15:50 -05:00

dcharlot referenced this pull request from a commit 2026-11-06 13:15:52 -05:00

Merge pull request 'wai dev's test: the stand-in room service listens without a reverse lookup' (#663) from fix/fake-stack-reverse-lookup into main

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

Subscribe

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!663

No description provided.