benchmark,tools: fix napi benchmarks on GHA

[nigrosimone]

The GHA benchmark cannot run any napi/* benchmark: every run prints "Binding failed to load" (for example on #66395). Two reasons:

• the workflow builds Node, but never the benchmark addons
• the build uses --debug-node (the default of shell.nix), so process.features.debug is true and benchmark/common.js looks for the addon in build/Debug, while node-gyp builds it in build/Release

Now the workflow runs make bench-addons-build when the category has napi, and benchmark/common.js takes the build type from process.config, as test/common does.

Tested on my fork, napi/make_callback with the same code on both sides (x86_64-linux only): https://github.com/nigrosimone/node/actions/runs/36763821317

                              confidence improvement accuracy (*)   (**)  (***)
napi/make_callback n=1000000                  1.51 %       ±1.66% ±2.49% ±3.95%
napi/make_callback n=10000000          *      1.53 %       ±1.25% ±1.83% ±2.75%

Disclosure: I used Opus 5.5 (Max) as coding assistant

[nodejs-github-bot]

Review requested:

• @nodejs/actions
• @nodejs/performance

[panva]

This conflicts with #66351 (already in commit-queue PRs queued for automated landing through the Commit Queue. )

[nigrosimone]

@panva rebased on #66351

[aduh95]

I don't understand why we need two make calls, can't we merge them?

Suggested change

	make build-ci -j4 V=1 ${{ contains(format(' {0} ', inputs.category), ' napi ') && '&& make bench-addons-build' || '' }}
	make build-ci ${{ contains(format(' {0} ', inputs.category), ' napi ') && 'bench-addons-build' || '' }} -j4 V=1

[nigrosimone]

They can't be merged safely: with -j4 make builds both goals concurrently, and nothing orders bench-addons-build after build-ci.
On a clean checkout $(NODE_EXE) can hit config.gypi before configure runs, or race with build-ci's inner make on out/Release. && guarantees the full build finishes first.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.23%. Comparing base (9067cc4) to head (235ee9b).
⚠️ Report is 146 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #66423      +/-   ##
==========================================
+ Coverage   90.42%   93.23%   +2.80%     
==========================================
  Files         790      422     -368     
  Lines      275435   193715   -81720     
  Branches    52825    32844   -19981     
==========================================
- Hits       249074   180611   -68463     
+ Misses      16772    12796    -3976     
+ Partials     9589      308    -9281     

see 513 files with indirect coverage changes

🚀 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.