Fix memory consumption on empty/small messages (#13393) (#13545)

[Dreamsorcerer]

(cherry picked from commit 5e54037)

[greptile-apps]

Confidence Score: 5/5

The PR appears safe to merge, with no concrete blocking or independently actionable non-blocking issue identified.

The new queue accounting, parser stall/resume behavior, and endpoint ownership lifecycle are coordinated across client and server paths and are exercised by focused parser and transport-level tests.

_{Reviews (1): Last reviewed commit: "Fix memory consumption on empty/small me..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.24%. Comparing base (e3640af) to head (191c944).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             3.14   #13546      +/-   ##
==========================================
+ Coverage   96.20%   96.24%   +0.03%     
==========================================
  Files         160      160              
  Lines       51133    51529     +396     
  Branches     2778     2809      +31     
==========================================
+ Hits        49195    49593     +398     
+ Misses       1756     1755       -1     
+ Partials      182      181       -1     

Flag	Coverage Δ
CI-GHA	96.23% <100.00%> (+0.03%)	⬆️
OS-Linux	96.02% <100.00%> (+0.03%)	⬆️
OS-Windows	93.82% <96.52%> (+0.13%)	⬆️
OS-macOS	95.29% <96.52%> (+0.11%)	⬆️
Py-3.10	95.40% <100.00%> (+0.04%)	⬆️
Py-3.11	95.64% <100.00%> (+0.04%)	⬆️
Py-3.12	95.72% <100.00%> (+0.04%)	⬆️
Py-3.13	95.70% <100.00%> (+0.03%)	⬆️
Py-3.14	95.82% <100.00%> (+0.04%)	⬆️
Py-3.14t	94.90% <96.52%> (+0.11%)	⬆️
Py-pypy-3.11	94.67% <100.00%> (+0.04%)	⬆️
VM-macos	95.29% <96.52%> (+0.11%)	⬆️
VM-ubuntu	96.02% <100.00%> (+0.03%)	⬆️
VM-windows	93.82% <96.52%> (+0.13%)	⬆️
cython-coverage	79.13% <81.88%> (+0.19%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

[codspeed-hq]

Merging this PR will not alter performance

✅ 80 untouched benchmarks
⏩ 7 skipped benchmarks¹

---

_{Comparing patchback/backports/3.14/b62321b6fa264a89f4aca084ad60fc38a59ffa51/pr-13545 (191c944) with 3.14 (e3640af)}

Footnotes

1. 7 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩