[sylvestre]

• s/x/y/Ng and s/x/y/gN are now accepted, like GNU: replace the Nth match and every one after. They were rejected with "multiple 'g' or numeric flags in substitute command" (and unit tests asserted that).
• Repeated flags now use GNU's messages, multiple g' options to s' command and multiple number options to s' command. The latter points at the second number (3ins/a/b/2g3, char 9 like GNU) instead of at the g`.
• An invalid back-reference (s/\(a\)/\2/) was reported after the flags were parsed, so the UUTILS_DIAG caret sat past the end of the command. It now points at the \2, including when the replacement spans several lines.

$ echo aaaa | sed 's/a/b/2g'
abbb
$ UUTILS_DIAG=always sed 's/\(a\)/\2/'
sed: <script argument 1>:1:9: error: invalid reference \2 on `s' command's RHS
   ╭─[ <script argument 1>:1:9 ]
   │
 1 │ s/\(a\)/\2/
   │         ─
───╯

[Copilot]

🟢 Approval recommended

The focused implementation and regression coverage support the intended behavior, with no unresolved findings.

No open issues 🎉

What changed in this PR

Improves sed’s GNU compatibility for substitution flags and makes error locations more precise.

Changes:

• Accepts Ng and gN to replace the Nth match and all later matches.
• Uses GNU-style duplicate-flag messages and points diagnostics at offending flags or back-references.
• Adds regression tests for substitution behavior and diagnostic locations.

File	Description
tests/​by-util/​test_sed.rs	Tests combined flags and diagnostic carets.
src/​sed/​processor.rs	Applies global replacement from the selected occurrence.
src/​sed/​compiler.rs	Parses combined flags and records precise error locations.
src/​sed/​command.rs	Stores global mode and back-reference locations.

_{Review effort: Balanced}

[codspeed]

Merging this PR will improve performance by 7.68%

⚠️ 1 benchmark spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• access_log_append

⚡ 2 improved benchmarks
✅ 9 untouched benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	number_fix	1.3 s	1.2 s	+12.08%
⚡	access_log_subst	2.2 s	2.1 s	+3.46%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing sylvestre:subst-flag-diagnostics (294b1a1) with main (b1ea0dc)}

[codecov]

Codecov Report

❌ Patch coverage is 97.61905% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 94.84%. Comparing base (b1ea0dc) to head (294b1a1).

Files with missing lines	Patch %	Lines
src/sed/compiler.rs	97.36%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #608      +/-   ##
==========================================
- Coverage   94.85%   94.84%   -0.01%     
==========================================
  Files          14       14              
  Lines        7230     7241      +11     
  Branches      432      432              
==========================================
+ Hits         6858     6868      +10     
- Misses        371      372       +1     
  Partials        1        1              

Flag	Coverage Δ
macos_latest	96.38% <97.61%> (-0.01%)	⬇️
ubuntu_latest	96.32% <97.61%> (-0.01%)	⬇️
windows_latest	0.00% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sylvestre]

Split into #609, #610 and #611.