Use intra-doc links in str and BTreeSet

[Manishearth]

Partial fix for #32129 and #32130

A slight degradation in quality is that the #method.foo links would previously link to the same page on String's documentation, and now they will navigate to str. Not a big deal IMO, and we can also try to improve that.

[rust-highfive]

r? @kennytm

(rust_highfive has picked a reviewer for you, use r? to override)

[Manishearth]

This was previously reviewed in #74453

@bors r=jyn514

[bors]

📌 Commit 748634e has been approved by jyn514

[bors]

🌲 The tree is currently closed for pull requests below priority 5, this pull request will be tested once the tree is reopened

[jyn514]

@Manishearth why did you open a new PR? AFAIK the concerns in #74453 (comment) were never addressed.

@bors r-

cc @ehuss

[Manishearth]

@jyn514 I accidentally deleted some branches and then force pushed

It did not seem like those were blocking concerns. They cannot, at any rate, be fixed in this PR.

[jyn514]

To be clear, it is better to fix the broken links on the String page, but I consider it a backstep in other regards that should (eventually) be fixed.

Hmm ... Would it makes sense to use intra-doc links only for things that aren't primitives? Or is it not possible to use relative links because this module is re-exported in both std and core?

[Manishearth]

Hmm ... Would it makes sense to use intra-doc links only for things that aren't primitives?

This wouldn't fix the original problem though. Note that it's not just primitives, it's methods of primitives, etc.

[jyn514]

Ok I misunderstood what happened - currently these links work in libcore but are broken in libstd. This PR makes it so they work in both crates, but link externally from libcore (they'll go to rust-lang.org/nightly). The second bit is unfortunate but IMO better than the status quo.

We should open an issue about the primitives, this has popped up a few different times now.

@bors r+

[bors]

📌 Commit 748634e has been approved by jyn514

[bors]

🌲 The tree is currently closed for pull requests below priority 5, this pull request will be tested once the tree is reopened

[Manishearth]

@jyn514 I was starting to draft an issue about the primitives, but I remembered that there may not be a solution here because primitives do not exist in docs before std. The code that makes this an external link is deliberately written to handle this case.

[jyn514]

@Manishearth why are primitives in libstd and not libcore? That seems weird.

[Manishearth]

(#73423 btw, cc @ehuss)

[ollie27]

Fixes #32129, fixes #32130

Those issues can't be closed until the exceptions in linkchecker are removed:

rust/src/tools/linkchecker/main.rs

Lines 122 to 127 in d3df851

	// FIXME(#32129)
	if file.ends_with("std/io/struct.IoSlice.html")
	|| file.ends_with("std/string/struct.String.html")
	{
	return None;
	}

and

rust/src/tools/linkchecker/main.rs

Lines 132 to 141 in d3df851

	// FIXME(#32130)
	if file.ends_with("alloc/collections/btree_map/struct.BTreeMap.html")
	|| file.ends_with("alloc/collections/btree_set/struct.BTreeSet.html")
	|| file.ends_with("std/collections/btree_map/struct.BTreeMap.html")
	|| file.ends_with("std/collections/btree_set/struct.BTreeSet.html")
	|| file.ends_with("std/collections/hash_map/struct.HashMap.html")
	|| file.ends_with("std/collections/hash_set/struct.HashSet.html")
	{
	return None;
	}

[jyn514]

@bors r-

[Manishearth]

@ollie27 hmm. Unfortunately not all of those exceptions can be removed, because some of the links are forward references, which is out of scope for intra doc links

[Manishearth]

@ollie27 would you be okay if we can land this, fixing most of the cases, and unregistering the concern on intra doc stabilization?

[Manishearth]

@bors r=jyn514

removed the "fixes foo" keyword from the PR body

[bors]

📌 Commit 748634e has been approved by jyn514