build: ignore changes in docs

[peixotoleonardo]

When only documentation is changed, running tests may not be necessary. Perhaps we could consider skipping them in such cases.

[nodejs-github-bot]

Review requested:

• @nodejs/actions

[richardlau]

We have tests that check the docs for consistency, e.g.

node/test/parallel/test-cli-node-options-docs.js

Lines 13 to 14 in 3178a76

	const cliMd = path.join(rootDir, 'doc', 'api', 'cli.md');
	const cliText = fs.readFileSync(cliMd, { encoding: 'utf8' });

[peixotoleonardo]

@richardlau

This test ensures that all CLI options are documented but does not validate other parts of the documentation.

I put !doc/api/cli.md to run the test if this file change.

[richardlau]

@richardlau

This test ensures that all CLI options are documented but does not validate other parts of the documentation.

I put !doc/api/cli.md to run the test if this file change.

There are others, e.g.

node/test/parallel/test-release-changelog.js

Line 39 in 3178a76

const changelogPath = `doc/changelogs/CHANGELOG_V${major}.md`;

(I don't have time to find them all right now -- the point is that we do have tests that are looking at the docs).

[richardlau]

The addons tests also extract the code snippets from the documentation to test those:

node/Makefile

Lines 369 to 387 in 3178a76

	DOCBUILDSTAMP_PREREQS = tools/doc/addon-verify.mjs doc/api/addons.md
	ifeq ($(OSTYPE),aix)
	DOCBUILDSTAMP_PREREQS := $(DOCBUILDSTAMP_PREREQS) out/$(BUILDTYPE)/node.exp
	endif
	ifeq ($(OSTYPE),os400)
	DOCBUILDSTAMP_PREREQS := $(DOCBUILDSTAMP_PREREQS) out/$(BUILDTYPE)/node.exp
	endif
	node_use_openssl_and_icu = $(call available-node,"-p" \
	"process.versions.openssl != undefined && process.versions.icu != undefined")
	test/addons/.docbuildstamp: $(DOCBUILDSTAMP_PREREQS) tools/doc/node_modules
	@if [ "$(shell $(node_use_openssl_and_icu))" != "true" ]; then \
	echo "Skipping .docbuildstamp (no crypto and/or no ICU)"; \
	else \
	$(RM) -r test/addons/??_*/; \
	[ -x $(NODE) ] && $(NODE) $< || node $< ; \
	[ $$? -eq 0 ] && touch $@; \
	fi

[peixotoleonardo]

@richardlau

This case I will close the PR, sorry about my English (It's in progress).