perf(client): parse h1 content-length statelessly

[trivikr]

This relates to...

N/A

Rationale

Avoid allocating and parsing Content-Length response headers through strings.
Since llhttp already validates Content-Length as 1*DIGIT and rejects duplicate values, the H1 parser can parse the numeric value directly from the header buffer and use -1 to represent absence.

Changes

• Parse H1 response Content-Length directly from the header buffer.
• Use -1 as the parser sentinel for missing Content-Length.
• Check response body length against the parsed numeric value on message completion.

Features

N/A

Bug Fixes

N/A

Breaking Changes and Deprecations

N/A

Status

• I have read and agreed to the Developer's Certificate of Origin
• Tested
• Benchmarked (optional)
• Documented
• Review ready
• In review
• Merge ready

Assisted-by: openai:gpt-5.4

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.13%. Comparing base (5c96f7d) to head (676da7e).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5124      +/-   ##
==========================================
- Coverage   93.14%   93.13%   -0.01%     
==========================================
  Files         110      110              
  Lines       35898    36109     +211     
==========================================
+ Hits        33436    33629     +193     
- Misses       2462     2480      +18     

☔ View full report in Codecov by Sentry.
📢 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.

[trivikr]

The benchmark improved after using for loop instead of array.reduce

benchmark                   avg (min … max) p75 / p99    (min … top 1%)
------------------------------------------- -------------------------------
• parse content-length
------------------------------------------- -------------------------------
legacy                       392.90 ns/iter 390.98 ns  █                   
                      (376.62 ns … 1.64 µs) 434.16 ns ██▇▃ ▄               
                    ( 95.61  b … 310.16  b) 104.99  b ██████▆▅▂▃▂▂▃▁▁▁▂▁▁▁▁

stateless                     38.53 ns/iter  38.28 ns               █      
                     (28.36 ns … 104.78 ns)  42.76 ns               █      
                    (  0.09  b …  80.11  b)   0.26  b ▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▃▁▁▁▁▁

Marking this PR as ready for review.

[Uzlopak]

so basically parseInt?

[trivikr]

Yes. But takes 10X less time #5124 (comment)

[mcollina]

lgtm