Repository navigation
bug(transport): disable SSE reconnect without overflowing the timer #11
Description
Expected behavior
A disabled reconnect policy exits the loop without scheduling another connection. Enabled delays are bounded and abortable.
Source finding and impact
autoReconnectSse=false passes Number.MAX_SAFE_INTEGER as retryMs; SseClient forwards it to setTimeout. This is outside the native timer range and does not represent disabled reconnect.
Evidence and environment
Inspected source at 4f2ef5de7c80ad7a6199dc09869ea3b728752f0e on 2026-09-12 in the local macOS workspace. This finding has not yet been reproduced through a running application or the package integration suite.
Regression and acceptance
Use actual Node runtime plus wrapper/fake transport coverage: one failed stream with reconnect disabled causes no second fetch, timer overflow warning or sleeping retry. Validate restart cancels the old loop.
The resolving PR must protect the observable contract and document any remaining uncertainty. This planning pass makes no runtime changes.
Refs Juliusolsson05/agent-code#918; planning finding F05.
Additional local primitive evidence
On Node v24.14.1, an isolated setTimeout call with Number.MAX_SAFE_INTEGER produced internal delay 1 and TimeoutOverflowWarning; the timer was immediately cleared. No network retry or native OpenCode process was involved. Public-wrapper integration coverage is still required. Planning PR: Juliusolsson05/agent-code#931.
Activity