[Shopify Sync] Fix failure dump with invalid UTF-8

[dkeegan-figma]

Summary

Cherry-picks Shopify's UTF-8 encoding fix (commit ed4ae26). Prevents JSON dump failures when test output contains invalid UTF-8 bytes, ensuring failure reports are always written successfully.

Changes

Modified Files

• ruby/lib/minitest/queue/failure_formatter.rb - Added UTF-8 encoding with replacement

New Files

• ruby/test/minitest/queue/failure_formatter_test.rb - Test for UTF-8 handling

The Bug

When test output contains invalid UTF-8 bytes (e.g., from binary data, malformed strings), the Ruby JSON library raises an exception because it validates UTF-8 encoding:

JSON.generate({output: "\x80\x81\x82"})
# => Encoding::UndefinedConversionError

This prevents the entire failure report from being written, losing valuable debugging information.

The Fix

Encode output as UTF-8 before JSON serialization, replacing invalid bytes with empty string:

@output = output.encode('UTF-8', invalid: :replace, undef: :replace, replace: '')

Example

# Before: Crash on invalid UTF-8
"\x80\x81\x82" # => JSON exception

# After: Invalid bytes replaced
"\x80\x81\x82" # => "" (invalid bytes removed)
"Test \x80 output" # => "Test  output"

Conflict Resolution

✅ Applied cleanly - No conflicts!

Benefits

• Reliability: Failure reports always written, even with binary data
• Debugging: Never lose failure information due to encoding issues
• Robustness: Handles edge cases gracefully

Real-World Scenarios

This fixes failures when tests interact with:

• Binary files
• Network responses with incorrect encoding
• External processes with corrupted output
• Legacy systems with non-UTF-8 data

Testing

• ✅ Ruby syntax checks pass
• ✅ UTF-8 encoding logic verified
• ✅ New test file included
• ⏳ Full test suite requires Redis (run in CI)

Original Shopify PR

• Shopify PR: #365
• Original commit: ed4ae26
• Author: Hartley McGuire skipkayhil@gmail.com

Risk Assessment

Risk Level: Very Low

• Defensive fix, only activates on invalid input
• No behavior change for valid UTF-8
• Well-tested encoding approach
• Small, focused change

Additional Notes

The fix uses replace: '' (empty string) to remove invalid bytes. Alternative would be replace: '?' to show where bytes were replaced, but empty string is cleaner for most use cases.

---

cc: @Dkeegan