Remove many unnecessary manual link resolves from library

[camelid]

Now that #76934 has merged, we can remove a lot of these! E.g, this is
no longer necessary:

[`Vec<T>`]: Vec

cc @jyn514

[rust-highfive]

r? @Mark-Simulacrum

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

[camelid]

If anyone's interested, I used this kind of hacky Python script (gist):

import re
from pathlib import Path
import sys

pat = re.compile(r'\[`?(.*)<.*>(::([A-Za-z0-9_]+))?(\(\))?`?\]: \1(::\3)?')

for path in Path(sys.argv[1]).rglob('*.rs'):
  print(path)

  inf = open(path, 'r')
  lines = inf.readlines()
  inf.close()

  outf = open(path, 'w')
  prev_line_empty = False
  prev_prev_line_empty = False
  for line in lines:
    if not pat.search(line):
      if not (prev_line_empty and prev_prev_line_empty and line.strip('/ \n') == ''):
        outf.write(line)
      prev_prev_line_empty = prev_line_empty
      prev_line_empty = line.strip('/ \n') == ''
    else:
      print('match!')
      prev_prev_line_empty = prev_line_empty
      prev_line_empty = True  # empty because was not written to file
  outf.close()

There were a few spots where it was over-eager in removing blank lines, but they were easy to fix manually.

[jyn514]

r? @jyn514

[jyn514]

This is blocked until the next beta bump, I think, we intentionally test the standard library is documented with --stage 0 so library contributors don't have to build the compiler: #75368 (comment)

[camelid]

Oops, forgot :)

Oh well, at least the PR will be ready then.

[camelid]

Rebased. I think this is ready now!

[jyn514]

@camelid I think this is ready to go once you rebase (which should fix the linkchecker error).

[camelid]

Don't we have to wait for the next beta bump though?

[jyn514]

@camelid I think I might have been wrong about that - since the links fail silently, they won't actually cause x.py doc --stage 0 to fail, the links will just be broken. Since I don't expect many people to be looking at every link in a stage 0 build, I think it's ok to have them broken for a cycle - doc.rust-lang.org/nightly/ will still have the correct links.

[camelid]

@jyn514 Actually they don't fail silently (at least not in CI):

 std/primitive.never.html:44: broken intra-doc link - [<code>Result&lt;String, !&gt;</code>]
std/primitive.never.html:51: broken intra-doc link - [<code>Result&lt;T, !&gt;</code>]
std/primitive.never.html:55: broken intra-doc link - [<code>Result&lt;T, !&gt;</code>]
std/primitive.never.html:56: broken intra-doc link - [<code>Result&lt;T, !&gt;</code>]
std/primitive.never.html:57: broken intra-doc link - [<code>Result&lt;!, E&gt;</code>]
thread 'main' panicked at 'found some broken links', src/tools/linkchecker/main.rs:109:9
std/primitive.never.html:81: broken intra-doc link - [<code>Result&lt;!, E&gt;</code>]

[jyn514]

@camelid I'm pretty sure that's from before #79321 landed, it's from a --stage 1 build, not a --stage 0 build.

[camelid]

@jyn514 I just told CI to re-run, so we'll see what happens :)

[camelid]

CI still failed with the same linkchecker errors.

[camelid]

Probably this should just wait until the next beta bump – it's not urgent, although it is annoying to have to wait another 5 weeks :/

[camelid]

Only a few more days and then we can finally merge this! 😂

[camelid]

Okay, the beta bump happened! Re-running CI...

[camelid]

Hmm, I don't see the button to re-run the workflow. Maybe because it's too old.

@jyn514 do you want to r+?

[jyn514]

@camelid can you rebase and push manually to make sure the links actually work?

[camelid]

@jyn514 CI passed. Should be ready to go!

[jyn514]

@bors r+ rollup

Thanks for sticking with this! I know waiting for the release trains is a pain.

[bors]

📌 Commit 0506789 has been approved by jyn514

[bors]

⌛ Testing commit 0506789 with merge 0876f59...

[bors]

☀️ Test successful - checks-actions
Approved by: jyn514
Pushing 0876f59 to master...