log: escape terminal control characters

[steadytao]

Fixes #1112

Summary

Escape terminal control characters in filenames and diagnostics written to standard output, standard error or rsync log files. The output filter preserves printable text in the active locale while escaping:

• C0 controls except tab
• DEL
• C1 controls
• existing \#NNN sequences which would otherwise be ambiguous

Invalid or incomplete multibyte input falls back to conservative byte filtering. Generic output no longer trusts leading or trailing carriage returns. The three intentional progress displays emit their intended carriage return through a dedicated helper.

Compatibility

Character boundaries are decoded using the active locale. This preserves valid multibyte text including UTF-8 and EUC-JP while still escaping C1 bytes in single-byte locales such as ISO-8859-1. Systems without wchar.h or mbrtowc() use a conservative byte-oriented fallback. Default filtering without --8-bit-output remains byte-oriented.

[steadytao]

@rubo77 please review

[seks99x]

I have this on my radar, just need some proper time reviewing this complex bounds checking and pointer math. Will come back to it

[steadytao]

Both locale concerns from @FusionPow were valid. The filter was treating permitted high bytes as UTF-8 regardless of the active locale. Reworked this to instead use the active locale's decoder to identify boundaries. ISO-8859-1 d8 9b preserves d8 and escapes the C1 byte while EUC-JP 8e b1 remains intact. Invalid or incomplete input falls back to byte filtering.

I also tightened the output bounds before expansion and added regressions for UTF-8, ISO-8859-1, EUC-JP and LC_ALL=C. Thanks for checking these cases.

[steadytao]

And @rubo77, I would report a case only where raw control characters demonstrably reach a terminal. Raw bytes in redirected or machine-readable output can be intentional so this should not be reported broadly without reproducing a terminal path.

[steadytao]

Updated description; will address any CI failures.

[steadytao]

@seks99x: I have this on my radar, just need some proper time reviewing this complex bounds checking and pointer math. Will come back to it

Also. Appreciated.

[seks99x]

Great job @steadytao. LGTM