[WilliamBergamin]

Summary

This pull request fixes a close-handshake timer that could outlive its connection in SlackWebSocket.disconnect().

Fixes #2771. Follow-up to #2710, which added the timer.

The bug: disconnect() starts a 30s timer when it sends a close frame. In two cases that timer survived cleanup, then fired 30s later: a false "Peer did not complete the close handshake" warning and a second close event, which can trigger a duplicate reconnect.

1. The connection drops after we send a close frame (the reported case). undici fires error with the socket already CLOSED, so disconnect() started a second timer and lost the first.
   • 🟢 disconnect() on a CLOSED socket now cleans up right away.
2. disconnect() is called before the connection opens. undici closes a connecting socket synchronously inside close(), so cleanup ran before the timer was started.
   • 🟢 The timer now starts before close().

Note: when the connection drops, the client now receives close before error.

Not in this PR:

• Emitting close only once per instance: client.disconnect() during a reconnect backoff relies on that extra close to resolve.
• Logging the error at debug level while closing: a separate PR.

Testing

Each case has a new unit test that fails on main.

To check by hand against a build of this branch:

• Run the script from socket-mode: an error while closing re-arms closeHandshakeTimeout and orphans the first timer (false close-handshake warning, spurious close) #2771: the false warning ~20s after the drop is gone.
• Call client.disconnect() while the socket is still connecting and wait 31s: one close, no warning.

Requirements

• I've read and understood the Contributing Guidelines and have done my best effort to follow them.
• I've read and agree to the Code of Conduct.

🤖 Generated with Claude Code

[changeset-bot]

🦋 Changeset detected

Latest commit: 94e0050

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@slack/socket-mode	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.39%. Comparing base (1fabb45) to head (94e0050).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2777      +/-   ##
==========================================
- Coverage   89.40%   89.39%   -0.02%     
==========================================
  Files          65       65              
  Lines       10545    10548       +3     
  Branches      478      480       +2     
==========================================
+ Hits         9428     9429       +1     
- Misses       1086     1087       +1     
- Partials       31       32       +1     

Flag	Coverage Δ
cli-hooks	89.28% <100.00%> (-0.07%)	⬇️
cli-test	89.28% <100.00%> (-0.07%)	⬇️
logger	89.28% <100.00%> (-0.07%)	⬇️
oauth	89.28% <100.00%> (-0.07%)	⬇️
socket-mode	89.28% <100.00%> (-0.07%)	⬇️
web-api	89.28% <100.00%> (-0.07%)	⬇️
webhook	89.28% <100.00%> (-0.07%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[zimeg]

🧠 @WilliamBergamin Thanks for handling all of these cases and state possibilities!