[bob sira (bobsira)]

Summary

Fixes #121 for UTF-8 child-process output.

The workaround introduced by #122 allowed processing to continue but replaced high-bit bytes with ?. The current standard formatter retained that behavior, corrupting valid multibyte characters in JSON and XML process logs.

• Preserve complete 2-, 3-, and 4-byte UTF-8 sequences, validating them with the existing Utility::IsTextUTF8 helper.
• Retain ? replacement for invalid bytes and NUL in the formatter, while preserving valid characters elsewhere in the line.
• Leave ASCII handling, custom formatting, pipe line framing, and the output schema unchanged.
• Add regression coverage and document the expected child-process encoding.

This does not introduce legacy Windows code-page detection/conversion or repair text already corrupted by the child application.

Validation

• Built LogMonitor and LogMonitorTests in x64 Release.
• Before the fix, the standard-format UTF-8 regression failed; ASCII/NUL handling and custom UTF-8 output passed.
• After the fix, all 7 selected process-formatting and utility tests passed. Coverage includes mixed 2/3/4-byte characters, long lines, malformed/truncated sequences, and valid characters around invalid bytes.
• Real child-process checks compared the preserved baseline executable with the candidate. The baseline corrupted UTF-8 as expected; the candidate preserved it in JSON, XML, and custom output, including six split-sequence cases at 4096-byte write boundaries and stderr. JSON/XML also preserved valid characters around malformed bytes while replacing the malformed bytes with ?.
• The targeted lint run reports the same pre-existing missing <string> include diagnostic on both unchanged main and this branch; no new lint diagnostics were introduced.

Validation was scoped to process formatting and related utilities. The broader EventMonitor replay crash tracked in #238 is outside this change.

[Copilot]

🟡 Changes recommended

Newly preserved U+FFFE and U+FFFF characters produce invalid XML and need XML-specific replacement.

1 open issue

• Replace XML-forbidden U+FFFE and U+FFFF with question marks · New

What changed in this PR

This PR replaces the earlier byte-replacement workaround with UTF-8 preservation in standard process logs, addressing #121 while leaving custom formatting and pipe framing unchanged.

Changes:

• Preserve valid multibyte UTF-8 sequences while replacing invalid bytes and NUL.
• Add regression coverage for standard and custom formatting.
• Document the expected child-process encoding and lack of code-page conversion.

File	Description
LogMonitor/​src/​LogMonitor/​ProcessMonitor.cpp	Validate and preserve UTF-8 sequences in standard output.
LogMonitor/​LogMonitorTests/​LogMonitorTests.cpp	Cover UTF-8 preservation, malformed bytes, and existing formatting behavior.
LogMonitor/​docs/​README.md	Explain UTF-8 expectations for child-process output.

_{Review effort: Balanced}