Issue Summary

During a test/format run, unexpected garbled/non-ASCII characters are appearing in the log output around WT_SESSION.verify lines. The output includes sequences such as:

[2026/09/24 12:36:43.937] Аҧ�� WT_SESSION.verify: 1876
[2026/09/24 12:40:38.962] 0�:��� WT_SESSION.verify: 3110
[2026/09/24 12:45:23.864] (�:��� WT_SESSION.verify: 2937

This looks like a logging/encoding issue in the test harness or runtime output and needs investigation.

Context

• The issue is visible in captured output from multiple timestamps, suggesting it is reproducible within the same run.
• The affected lines are associated with WT_SESSION.verify and appear alongside normal test output such as t: successful run completed.
• The unexpected characters are not expected in the log stream and may indicate corrupted encoding, terminal/locale mismatch, or some form of binary data being emitted to stdout/stderr.

Proposed Solution

• Investigate the source of the output for WT_SESSION.verify in the test/format pipeline.
• Determine whether the characters are caused by:
  • an encoding/locale mismatch,
  • a logging library emitting non-UTF-8 bytes,
  • a terminal/console rendering issue, or
  • corrupted output from the underlying process.
• Reproduce with the same test/format command and inspect raw logs/artifacts to identify the exact byte sequence being emitted.
• Fix the emitting component or sanitize/normalize the output so logs remain valid text.

Original Slack thread: https://mongodb.slack.com/archives/C0AKD3HPD8T/p1790219226010279
This ticket was generated by AI from a Slack thread.