Prevent heap allocations in FiveTuple Fingerprint

[paulwe]

No description provided.

[codecov]

Codecov Report

Attention: Patch coverage is 88.88889% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 68.27%. Comparing base (8958518) to head (9616801).

Files	Patch %	Lines
internal/allocation/five_tuple.go	88.23%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #393      +/-   ##
==========================================
- Coverage   68.40%   68.27%   -0.14%     
==========================================
  Files          43       43              
  Lines        2333     2348      +15     
==========================================
+ Hits         1596     1603       +7     
- Misses        573      578       +5     
- Partials      164      167       +3     

Flag	Coverage Δ
go	68.27% <88.88%> (-0.14%)	⬇️
wasm	27.93% <72.22%> (+0.29%)	⬆️

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.

[davidzhao]

nice, this will be a solid speedup

[rg0now]

We've already tried to track this down with @stv0g, see #299 and #304. This seems like a much better solution and the immediate effect is indeed massive performance improvement in GetAllocation based on my benchmarks (go test -cpuprofile cpu_895851.prof -bench=. -run=XXX -benchtime=200000x).

GetAllocation before the merge: UDPAddr.String dominates performance.

GetAllocation after the merge: FiveTupeFingerprint got supercheap.