Restore 304 performance after fixing FileResponse replace race

[bdraco]

Followup to #10101 to address the feedback in #10112

We can avoid opening the file is we are going to send a 304.

fixes #10112

[codspeed-hq]

CodSpeed Performance Report

Merging #10113 will improve performances by 17.63%

_{Comparing 304_perf (360dd40) with master (9456955)}

Summary

⚡ 1 improvements
✅ 46 untouched benchmarks

Benchmarks breakdown

	Benchmark	master	304_perf	Change
⚡	test_simple_web_file_response_not_modified[pyloop]	67.6 ms	57.5 ms	+17.63%

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.76%. Comparing base (9456955) to head (360dd40).
Report is 1 commits behind head on master.

✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #10113      +/-   ##
==========================================
+ Coverage   98.33%   98.76%   +0.42%     
==========================================
  Files         122      122              
  Lines       36951    36961      +10     
  Branches     4399     4410      +11     
==========================================
+ Hits        36337    36503     +166     
+ Misses        439      311     -128     
+ Partials      175      147      -28     

Flag	Coverage Δ
CI-GHA	98.64% <100.00%> (+0.39%)	⬆️
OS-Linux	98.33% <100.00%> (+0.15%)	⬆️
OS-Windows	96.18% <100.00%> (?)
OS-macOS	97.44% <100.00%> (+0.09%)	⬆️
Py-3.10.11	97.28% <100.00%> (+0.29%)	⬆️
Py-3.10.15	97.82% <100.00%> (+<0.01%)	⬆️
Py-3.11.10	97.86% <100.00%> (+0.65%)	⬆️
Py-3.11.9	97.33% <100.00%> (?)
Py-3.12.7	98.39% <100.00%> (+1.18%)	⬆️
Py-3.13.0	97.20% <100.00%> (-0.01%)	⬇️
Py-3.13.1	98.30% <100.00%> (?)
Py-3.9.13	97.20% <100.00%> (?)
Py-3.9.20	97.78% <100.00%> (?)
Py-pypy7.3.16	97.35% <100.00%> (?)
VM-macos	97.44% <100.00%> (+0.09%)	⬆️
VM-ubuntu	98.33% <100.00%> (+0.15%)	⬆️
VM-windows	96.18% <100.00%> (?)

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

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

[patchback]

Backport to 3.11: 💚 backport PR created

✅ Backport PR branch: patchback/backports/3.11/0130213423d8d055ece27c36a4dcc96e524cbc1c/pr-10113

Backported as #10117

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.

[patchback]

Backport to 3.12: 💚 backport PR created

✅ Backport PR branch: patchback/backports/3.12/0130213423d8d055ece27c36a4dcc96e524cbc1c/pr-10113

Backported as #10118

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.

[steverep]

Nice! I might have guessed maybe 5% penalty from the buffered open. How big is the benchmark file (at least 4K I hope)?

[bdraco]

The test uses this 11.1KB file https://github.com/aio-libs/aiohttp/blob/master/tests/sample.txt

The size of the file shouldn't make much (or maybe any) difference with most file systems unless its doing some type of read-ahead in the background since for most file systems, reading the metadata about the file doesn't require reading the contents.