fix: PINNED_BACKREFERENCE bugs from review

[jbachorik]

What does this PR do?

Fixes correctness bugs and remaining review findings in the PINNED_BACKREFERENCE strategy (a single-forward-scan matcher for backreference patterns like (\w+)\s+\1 where the group's charset is provably disjoint from what follows it):

• Non-capturing-group-wrapped separator quantifiers (e.g. (?:\s+)) were silently treated as exactly-one-occurrence instead of honoring their real bounds.
• A capturing group nested inside the pinned group's body or its separator silently broke totalGroupCount(), causing group(n) to be inaccessible or throw.
• Removed an unnecessary, allocation-equal, -1-buggy custom matchBounded/shiftSpans override in favor of the correct base-class default.
• Tightened minCandidateLength() to use the real group/separator minimum lengths instead of hardcoded values.
• Fixed stale javadoc examples and a mislabeled test section that referenced patterns no longer routed to this strategy.
• Added detector unit tests that kill two previously-surviving mutants (||→&& and <1→<0 on the span/separator-length checks).

Motivation

Follow-up from the PINNED_BACKREFERENCE strategy review (PR #94) — a multi-stream code review surfaced a CRITICAL correctness bug and several HIGH/MEDIUM/LOW findings, all addressed here.

Related Issue(s)

Change Type

• Bug fix
• New feature
• Performance improvement
• Refactoring (no functional change)
• Documentation
• Test improvement
• Build/CI change

Checklist

• I have read the CONTRIBUTING.md guidelines
• All existing tests pass (./gradlew build)
• I have added tests for my changes
• I have updated documentation (if applicable)
• My commits are signed

Performance Impact

minCandidateLength() now prunes more candidates before scanning (uses real group/separator minimum lengths instead of hardcoded 1s); no other performance-relevant change.

Additional Notes

Both commits on this branch (PINNED_BACKREFERENCE fixes) were produced with AI assistance (Claude Code) and reviewed by me.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 82.69231% with 18 lines in your changes missing coverage. Please review.
✅ Project coverage is 83.1%. Comparing base (343e776) to head (8d58197).

Files with missing lines	Patch %	Lines
...doghq/reggie/codegen/analysis/PatternAnalyzer.java	67.8%	6 Missing and 12 partials ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##              main     #97     +/-   ##
=========================================
- Coverage     83.8%   83.1%   -0.7%     
  Complexity       1       1             
=========================================
  Files          134     134             
  Lines        41162   41223     +61     
  Branches      5522    5548     +26     
=========================================
- Hits         34500   34288    -212     
- Misses        4957    5217    +260     
- Partials      1705    1718     +13     

Files with missing lines	Coverage Δ
...ggie/codegen/analysis/PinnedBackreferenceInfo.java	100.0% <100.0%> (ø)
.../codegen/PinnedBackreferenceBytecodeGenerator.java	100.0% <100.0%> (ø)
...ggie/processor/ReggieMatcherBytecodeGenerator.java	73.6% <ø> (-0.1%)	⬇️
.../com/datadoghq/reggie/runtime/RuntimeCompiler.java	87.8% <ø> (-0.1%)	⬇️
...doghq/reggie/codegen/analysis/PatternAnalyzer.java	76.0% <67.8%> (-0.1%)	⬇️

... and 3 files with indirect coverage changes

---

Continue to review full report in Codecov by Harness.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 343e776...8d58197. Read the comment docs.

🚀 New features to boost your workflow:

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

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b0d29b2322

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Reject multi-character separator atoms before using bounds

When the separator quantifier wraps an atom that consumes more than one character, these fields are repetition counts, but generateSeparatorLengthCheck compares them to sepLen, the number of characters scanned. For example, (\w+)(?:--){2}\1 is still accepted by this detector because the first separator charset is -, but the generated matcher scans four hyphens and rejects it as 4 > max(2), so a valid input like ab----ab becomes a false negative. Either reject non-single-character separator atoms here or convert the bounds to character lengths before passing them to codegen.

Useful? React with 👍 / 👎.