No changes needing a change description found.

You can try these changes here

@copilot fix format, spellcheck and run benchmark ci

@copilot fix format, spellcheck and run benchmark ci

Addressed in commit c226b35.

• Fixed formatting issues (prettier --check now passes).
• Resolved spellcheck by renaming unflattenRuntime to expandRuntimeMetrics.
• Ran benchmark CI-equivalent checks:
  • pnpm -r --filter "@azure-tools/typespec-benchmark..." build
  • pnpm --filter @azure-tools/typespec-benchmark test
  • benchmark run command with CI flags completed successfully.

@copilot run benchmark workflow still failing

@copilot run benchmark workflow still failing

Fixed in commit 093bcfe.

Root cause was Node engine mismatch in the benchmark workflow (pnpm install failed with ERR_PNPM_UNSUPPORTED_ENGINE on Node 24.11.1). I updated .github/workflows/benchmark.yml to use Node 24.15.0, which satisfies the dependency engine requirement.

@copilot benchmark run into OOM

  Benchmarking: azure-core-dataplane
    Warmup 1/3...
    Warmup 2/3...
    Warmup 3/3...
    Iteration 1/25...
    Iteration 2/25...
    Iteration 3/25...
    Iteration 4/25...
    Iteration 5/25...
    Iteration 6/25...
    Iteration 7/25...
    Iteration 8/25...

<--- Last few GCs --->

[4813:0x9db3000]   429954 ms: Scavenge (interleaved) 4048.6 (4076.5) -> 4041.3 (4114.7) MB, pooled: 0 MB, 34.43 / 0.00 ms  (average mu = 0.217, current mu = 0.229) allocation failure; 
[4813:0x9db3000]   431946 ms: Mark-Compact (reduce) 4062.8 (4116.2) -> 4047.8 (4061.7) MB, pooled: 0 MB, 145.02 / 0.00 ms  (+ 1717.1 ms in 328 steps since start of marking, biggest step 8.1 ms, walltime since start of marking 1992 ms) (average mu = 0.211,
FATAL ERROR: Ineffective mark-compacts near heap limit Allocation failed - JavaScript heap out of memory
----- Native stack trace -----

 1: 0x73f8c4 node::OOMErrorHandler(char const*, v8::OOMDetails const&) [node]
 2: 0xc06f90  [node]
 3: 0xc0707f  [node]
 4: 0xeaa885  [node]
 5: 0xeaa8b2  [node]
 6: 0xeaabaa  [node]
 7: 0xebb8aa  [node]
 8: 0xebfc50  [node]
 9: 0x1953f71  [node]
/home/runner/work/_temp/a4a90c35-7d83-4657-bd54-f22379613fdf.sh: line 9:  4813 Aborted                 (core dumped) node packages/benchmark/dist/src/cli.js run --specs-dir packages/benchmark/specs --iterations 25 --warmup 3 --noise-cv-threshold 0.08 --max-reruns 1 --rerun-iterations 10 --commit 6f075c146cd770a97dca3d420e095e7220bdeb57 --output /tmp/benchmark-results.json
Error: Process completed with exit code 134.

@copilot benchmark run into OOM

...

Fixed in 1248489.

I updated the benchmark workflow to run the benchmark CLI with a larger Node heap (--max-old-space-size=6144) for both run and backfill, which addresses the OOM seen during high-iteration runs.

@copilot still OOM, increasing memory doesn't seem to be the right solution anyway, what is causing this increase in use, is there a memory leak with an emitter or the benchmark code?

@copilot still OOM, increasing memory doesn't seem to be the right solution anyway, what is causing this increase in use, is there a memory leak with an emitter or the benchmark code?

Root cause is memory accumulation in the benchmark process, not a single emitter regression: repeated in-process compile() calls keep heap growth across iterations/specs until the run OOMs. I fixed this by isolating each iteration compile in a short-lived child process so memory is reclaimed between iterations while keeping the same benchmark output/metrics (3f2a5f9). I also validated benchmark build/tests and reran the high-iteration benchmark command successfully.