Website: exclude files that would cause problems for prebuild or jekyll

[zygoloid]

Exclude some files from the website and prebuild steps that aren't part of the git repository, but may exist in a checkout, and if present will cause the website prebuild or build to misbehave or break.

[jonmeow]

Approving from the response, but might still want to look at rmtree

[zygoloid]

I've followed the suggestion to exclude things rather than deleting them. I think that's better; deleting the node_modules in particular is a bit rude given that they would still be wanted for (eg) ongoing development of the vscode extension. There are three different cases here:

• .jekyll-cache and _site are not renamed from checkout/website/ to checkout/ because they might already exist in the destination and in any case wouldn't contain the right things.
• **/node_modules/**/*.md and **/dist/**/*.md are not included when building the navigation tree because they're not part of the repository. (This only really affects utils/vscode because it has a custom gitignore for these, but I've excluded them globally for simplicity.) They formerly caused the python script to break because parsing the title failed.
• The .bazel symlinks, as well as node_modules and dist, are now excluded from jekyll's build because they can be or contain broken symlinks as well as markdown files we don't want to include in the build. Same for .cache, but I've excluded all dot-files for safety.

[jonmeow]

Suggested change

	children = [
	x
	for x in subdir.glob("**/*.md")
	if x != readme
	and "/node_modules/" not in str(x)
	and "/dist/" not in str(x)
	]
	children = []
	for x in subdir.glob("**/*.md"):
	if x != readme and not ("/node_modules/" in str(x) or "/dist/" in str(x)):
	children.append(x)

This just because of the difficulty of reading multi-line comprehension conditions

[jonmeow]

Hmm, or moving the not (this is bikeshedding at this point):

Suggested change

	children = [
	x
	for x in subdir.glob("**/*.md")
	if x != readme
	and "/node_modules/" not in str(x)
	and "/dist/" not in str(x)
	]
	children = []
	for x in subdir.glob("**/*.md"):
	if not (x == readme or "/node_modules/" in str(x) or "/dist/" in str(x)):
	children.append(x)

[zygoloid]

Factored out the predicate as a function.