CI: Electron SIGSEGV crash in "Linux / Remote" smoke tests (Data Loss + Multiroot)

[ulugbekna]

Summary

The Linux / Remote smoke test job is failing with an Electron SIGSEGV crash and a follow-on timeout.

Observed in run #26519275250 on PR #318571, but the failures are unrelated to the PR changes — they appear to be infrastructure/flakiness issues.

Failures

1. Data Loss — verifies editors can save and restore

electron.launch: WebSocket error: socket hang up
[pid=10974][err] DevTools listening on ws://127.0.0.1:39813/devtools/browser/...
<ws error> socket hang up
<ws connect error> socket hang up
[pid=10974] <process did exit: exitCode=null, signal=SIGSEGV>

Electron crashes with SIGSEGV before the DevTools WebSocket can be established, so the smoke driver never connects.

2. Multiroot — "before all" hook for "shows results from all folders"

Error: Timeout of 120000ms exceeded.

This times out in the before all hook — likely a cascading failure caused by the Electron crash in the preceding test leaving the environment in a broken state.

Reproduction

• Job: Linux / Remote #78105290056
• Platform: linux-x64, Electron smoke tests (remote mode)
• Relevant log snippet:

  [pid=10974][err] [0527/150628.039401:ERROR:crashpad/util/file/file_io_posix.cc:145]
      open /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq: No such file or directory
  [pid=10974] <process did exit: exitCode=null, signal=SIGSEGV>
  

The cpufreq errors suggest the runner lacks /sys/devices/system/cpu/cpu0/cpufreq/, which is a known missing sysfs path in some containerized CI environments and can affect Electron's crashpad initialisation.

Impact

Unrelated PRs are blocked from auto-merging because this flaky job poisons the required checks.