[miguel-heygen]

What

check, snapshot and render no longer fail on Windows with EBUSY: resource busy or locked, unlink '.../chrome-headless-shell-win64/debug.log' when a Chrome from the managed install is still running.

Why

Every cleanup of a set-aside or killed-install browser folder in browser/manager.ts was a bare recursive rmSync. A running chrome-headless-shell keeps debug.log open and Windows refuses to unlink it, so the removal threw and the command died before it did any work. The failed removal left the folder behind, so the next command tried again and failed again.

Related work

#5698 and #6506 retry other locked files on Windows (ffmpeg executables, undo/redo); neither touches the browser cache.

How

One helper, removeLeftover, now does both removals that run during normal commands: the sweep of killed-install folders (before every download, and before a render) and the removal of the folder an upgrade sets aside. It retries a few times, then leaves a folder that Windows still holds (EBUSY/EPERM) for the next sweep. Any other error still throws. hyperframes browser clear uses the same sweep, so it also leaves a folder Windows still holds instead of throwing.

render shares the path: it calls the same sweep (sweepKilledInstalls) and ensureBrowser as check and snapshot.

Test plan

• Unit tests added/updated
• Manual testing performed
• Documentation updated (if applicable)
• Comments follow CONTRIBUTING.md "Comments"

fails without the fix: with the EBUSY/EPERM tolerance removed, does not fail a run when Windows holds a leftover's debug.log fails with promise rejected "Error: EBUSY: resource busy or locked, un…" instead of resolving.

What I measured

manager.test.ts: 72 passed on three runs on Linux; oxfmt and oxlint clean. The test makes the removal of a set-aside folder throw EBUSY and checks the sweep, a forced download and a forced re-download all resolve and the folder stays.

What I did NOT execute

A real Windows machine with a running Chrome. The cause is reasoned from the reported error and the code, not reproduced on Windows.

[github-actions]

Edit accuracy: accurate 271 (base branch 271), smooth 118 of those

The gate passes.
Smoothness is reported in the artifact, not gated. A case fails only if it fails 2 of 3 runs.

Quarantined, measured but not gated (0)

Unstable (4)

• seqrotatenudge-none-center-r0-nested-z100: tracking 0.02, pressJump 0, drop 0, reload 3.03, render 0.04, renderKey -, undo false, teleport true / tracking 0.02, pressJump 0, drop 0, reload 0, render 0.05, renderKey -, undo true, teleport true / tracking 0.02, pressJump 0, drop 0, reload 0, render 0.05, renderKey -, undo true, teleport true
• seqfastkeys-none-px-r0-root-z100: tracking 0.07, pressJump 0.05, drop 0, reload 0, render 0.02, renderKey -, undo false, teleport true / tracking 0.07, pressJump 0.05, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true / tracking 0.06, pressJump 0.05, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true
• seqplainkeys-none-pct-r0-root-z100: tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo false, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true
• seqplainkeys-none-pct-r0-nested-z100: tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo false, teleport true / tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true / tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true

[heygengenesis]

Verdict: APPROVE · 0 blocker · 0 important · 1 nit · head cc92a087

• nit: the body says hyperframes browser clear is unchanged, but clearBrowser (manager.ts:950) goes through removeInstallLeftovers, so it now swallows EBUSY/EPERM too. When the only thing left is a held .chrome-replaced-* folder, browser clear prints "Removed cached browser" and the folder is still there (before this change it threw). Worth a body fix, or keeping clearBrowser strict.

Checked

• Ran manager.test.ts at head: 72 passed. With the EBUSY/EPERM tolerance in removeLeftover replaced by a rethrow, the new test fails with promise rejected "Error: EBUSY…" instead of resolving, so the fails-without-the-fix claim holds. The second forced ensureBrowser covers the set-aside path in moveStagedInstallIntoCache, and the sweep covers removeInstallLeftovers.
• render reaches the same sweep through commands/render/execute.ts:201. Staging cleanup is still a bare rmSync, which is fine because a staged browser never runs.
• oxlint/oxfmt are clean. CI is green on this head, including the Windows render and test lanes.
• Not covered: real Windows behaviour (the body says the same).