fix: accept an empty flags string in regexp_replace

[jackylee-ch]

Which issue does this PR close?

• N/A

Rationale for this change

regexp_replace('foobarbaz', 'b..', 'X', '') errors with regex parse error: (?)b instead of returning fooXbaz. regexp_count and regexp_instr accept an empty flags string through compile_regex, and so does Postgres. 'gg' hits the same error, since stripping g leaves an empty flags string.

What changes are included in this PR?

Both call sites omit the (?flags) prefix when no flags remain after g is stripped, matching compile_regex in the same module.

What is the testing strategy for this PR?

Three queries in regexp_replace.slt and two unit tests in regexpreplace.rs. All of them fail on main with repetition operator missing expression and pass with this change.

Are there any user-facing changes?

regexp_replace with an empty flags string now replaces instead of erroring. No API change.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.67%. Comparing base (5b389eb) to head (2542278).

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #24987      +/-   ##
==========================================
- Coverage   81.67%   81.67%   -0.01%     
==========================================
  Files        1126     1126              
  Lines      414524   414545      +21     
  Branches   414524   414545      +21     
==========================================
+ Hits       338558   338571      +13     
- Misses      56054    56057       +3     
- Partials    19912    19917       +5     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.