Lintcheck: Add URL to lint emission place in diff

[xFrednet]

This PR adds links to the emission code in our lintcheck CI. When reviewing changes, I would like to be able to see the bigger context, which isn't always included in the lint message. This PR adds a nice link to the lintcheck diff that allows for simple investigation of the code in question.

At first, I wanted to simply use the doc.rs links and call it a day, but then I figured out that some crates might have their source files remapped. Cargo was the crate that made me notice this. As a response, I made the link configurable. (See rust-lang/docs.rs#2551 for a detailed explanation and possible solution to remove this workaround in the future.)

It's probably easiest to review the individual commits. The last one is just a dummy to showcase the change.

🖼️ rendered 🖼️

---

r? @Alexendoo

changelog: none

---

That's it, I hope that everyone who's reading this has a beautiful day :D

[xFrednet]

It looks like the cargo link is sadly still incorrect 🤔

[xFrednet]

The cargo links work now 🎉

Some links are currently not 100% correct as the versions we use for lintcheck are so old that they used a previous version of docs.rs link structure. I considered fixing this, but decided to hold of on it, until I update our default and CI test set (Which is on my todo list for next week)

[Alexendoo]

A few times I've forgotten if I'm reading the Added or Removed sections because they look the same, this would add a convenient place to put a plus/minus or similar, what do you think?

[xFrednet]

I like the idea. A simple + or - would be made into a list by GH's Markdown. I now used the title as a prefix, which constructs messages like this:

Added clippy::cool_lint at https://cataas.com/cat
Removed clippy::cool_lint at https://cataas.com/cat
Changed clippy::cool_lint at https://cataas.com/cat

Edit: 🖼️ rendered 🖼️

[Alexendoo]

Could remove these now since the file_link would cover it well enough

[xFrednet]

The byte position is still used in the key() function on line 19. We could replace it with the link, but I think having it separately would make more sense. While all links currently contain line labels, there is no guarantee that they will for manually configured lints. And links also don't contain the column, which is included in this field.

[Alexendoo]

Since they're displayed under [lint] at [url] sections I think those making up the key makes sense. Removing information from the key has a benefit - it would mean tweaks to a span's start column/end line/end column show as a change rather than

Added `x` at y#5
...
Removed `x` at y#5

I didn't think about links without line numbers, if that happened more things would appear as changes than necessary but reasonably I don't think we'll have such links

[Alexendoo]

Not critical though, either way it's ready for r+ after ditching the test commit

[xFrednet]

Removing information from the key has a benefit - it would mean tweaks to a span's start column/end line/end column show as a change rather than

The tweaked span would still result in a added and removed message, since the lint message displays the span and is also used as part of the span. I think here it's safer to display them separately and leave it to the reviewer to merge them.

[Alexendoo]

The message isn't part of the key

[xFrednet]

Right, I mixed it up with the lint name 😅 But I'd still prefer it to be separate. Some lints might also be emitted multiple times per line.

[xFrednet]

I removed the dummy commit and renamed a variable to avoid clippy::similar_names. This should be good to go from my side

[Alexendoo]

🐰

@bors r+

[bors]

📌 Commit 0e3d197 has been approved by Alexendoo

It is now in the queue for this repository.

[bors]

⌛ Testing commit 0e3d197 with merge b31bce4...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: Alexendoo
Pushing b31bce4 to master...