SpanlessEq improvements

[Jarcho]

fixes #9775

Also includes a simplification to consts::constant's interface since I was already touching the code.

At the start of eq_expr the check:

if !self.inner.allow_side_effects && left.span.ctxt() != right.span.ctxt() {
    return false;
}

was removed. This was added in 49e2501 to handle cfg macros. This is better handled by the newly added check_ctxt.

changelog: [various lints]: Don't consider different cfg! expansions to be the same unless they are for the same config.
changelog: [various lints]: Don't consider the expansion of two different macros to be equal, even when they expand to the same token sequence.
changelog: [various lints]: Don't consider two blocks to be equal if they contain disabled code or empty macro expansions, unless those section contain the exact same token sequence.

[rustbot]

r? @xFrednet

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

[xFrednet]

Hey, I have a busy week ahead of me. I'll review this next week, when I have a bit more time. Hope that's okay with you :)

Edit: It took a month, sorry!

[bors]

☔ The latest upstream changes (presumably #10528) made this pull request unmergeable. Please resolve the merge conflicts.

[xFrednet]

Looks good to me. Just two small questions, one is posted as a comment.

Have you measured, how much this change effects performance? I'd guess that it should not be significant, but it would still be good to check before merging it.

Also, thanks for being patient when it comes to the review! I had exams and really needed a break afterwards. The next review or r+ should be much faster 🙃

[xFrednet]

@Jarcho Have you seen the question in my review comment?

Have you measured, how much this change effects performance? I'd guess that it should not be significant, but it would still be good to check before merging it.

Once that is answered and the PR has been rebased, it should be ready for bors :)

[xFrednet]

Ping @Jarcho,

I'm currently triaging my PRs. This one was already approved and only requires a rebase. Would you mind doing so? Then you can r=me :)

[Jarcho]

Sorry. I kept telling myself I would deal with it 'later', and later has finally come. Don't ask me about when 'someday' is coming though. I'm not looking forward to that one.

Dogfood runs in approximately the same amount of time. So no obvious perf issues.

[Jarcho]

@bors r=xFrednet

[bors]

📌 Commit 58132cb has been approved by xFrednet

It is now in the queue for this repository.

[bors]

⌛ Testing commit 58132cb with merge 1e8d2c5...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: xFrednet
Pushing 1e8d2c5 to master...