Harden flaky transport CI tests

[quinnj]

Summary

This hardens two timing-sensitive transport tests that failed on a recent release-adjacent master CI run even though the immediately following master run passed.

• Replaces the early-final-response test's wall-clock assertion with synchronization that proves roundtrip! returns while the upload callback is blocked.
• Makes the keep-alive idle-timeout raw socket probe wait for the first response byte, with a longer Windows budget, matching the helper behavior already used by other raw probes.

Why

The failed CI run showed two test-fragility signals rather than a product regression:

• Ubuntu missed elapsed < 0.75 by about 13 ms.
• Windows intermittently read an empty first response, and that same idle-timeout test had failed in an earlier master run.

Validation

• julia --startup-file=no --project=. test/http_client_transport_tests.jl
• julia --startup-file=no --project=. test/http_server_http1_tests.jl
• git diff --check

Co-authored by Codex

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.96%. Comparing base (50b121c) to head (b98d88e).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1323   +/-   ##
=======================================
  Coverage   87.96%   87.96%           
=======================================
  Files          30       30           
  Lines       11807    11807           
=======================================
  Hits        10386    10386           
  Misses       1421     1421           

☔ 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.