Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (f4c85d5) to head (1d3f916).
⚠️ Report is 2 commits behind head on main.

@@            Coverage Diff             @@
##             main   #64341      +/-   ##
==========================================
- Coverage   90.24%   90.23%   -0.02%     
==========================================
  Files         741      741              
  Lines      240976   240982       +6     
  Branches    45408    45401       -7     
==========================================
- Hits       217477   217448      -29     
- Misses      15070    15120      +50     
+ Partials     8429     8414      -15     

I don't think we should land this. The more checks we have, the more performance impact it would create. Do you have any examples out there that is impacted by this? Does people really pass UTF8? If not, let's close this.

Hi @anonrig thanks for the review!

I completely understand the concern regarding the performance impact on the fast path. I picked this up to resolve issue #49888, as there seemed to be some inconsistency in how variations like utf-8 (which is common in HTTP headers/HTML meta tags) bypass the C++ fast path and fall back to the slower route.

That said, if the consensus is that the branch/check overhead for the 99% of users passing standard utf8 outweighs the benefit for the edge cases, I am perfectly happy to close this PR. Let me know what you think is best for the module!