Fix known hosts parsing

[TanmayArya-1p]

Fixes #16595

Previously, ssh.example.com would match the following line in known_hosts
*example.com,!*h.example.com ssh-ed25519 AAAAC...
But it should not match according to the OpenBSD manual. Trying this with git-fetch-with-cli = true prompts the user to add the new fingerprint. (But without git cli, Cargo does not throw an error even though it should)

How to test and review this PR?

I added a test to known_hosts.rs that demonstrates the new behaviour

[TanmayArya-1p]

This is a draft implementation should fix the linked issue. I want to try to go through OpenSSH's parsing logic just to make sure we haven't missed any other critical cases here.

[TanmayArya-1p]

Looked at the unit tests for openssh:
ref: https://github.com/openssh/openssh-portable/blob/master/regress/unittests/hostkeys/testdata/known_hosts

The parser correctly handles the test cases defined there as far as i can see.
I'm marking this PR as ready for review.

[rustbot]

r? @epage

rustbot has assigned @epage.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• People who recently interacted with files modified in this PR: @ehuss, @epage, @weihanglo
• @ehuss, @epage, @weihanglo expanded to ehuss, epage, weihanglo
• Random selection from ehuss, epage, weihanglo

[weihanglo]

We should also have test for this bracketed ports. Also, do you have any examples for this?

[TanmayArya-1p]

A host that is bracketed with ports was already tested here so this works as an example:

cargo/src/cargo/sources/git/known_hosts.rs

Line 774 in b9f0d83

assert!(khs[2].host_matches("[example.net]:2222"));

I also noticed another bug in the know_hosts parsing that was not mentioned in the original issue:
Say we have the following know_hosts file:

# known_hosts
[example.net]:2222 ssh-dss AAAAB3N..

e:2222 will successfully match because [example.com]:2222 is treated as a glob. In glob matching syntax, [ ] matches if any single character within the brackets matches.
Take a look at this for reference:
https://www.digitalocean.com/community/tools/glob?comments=true&glob=%5Bexample.net%5D%3A2222&matches=false&tests=e%3A2222

I have also added a test that validates against this behaviour here

Here is the block of code responsible:

cargo/src/cargo/sources/git/known_hosts.rs

Lines 612 to 623 in b9f0d83

	let is_glob = is_glob_pattern(&pattern);
	if is_glob {
	match glob::Pattern::new(&pattern) {
	Ok(glob) => match_found |= glob.matches(&host),
	Err(e) => {
	tracing::warn!(
	"failed to interpret hostname `{pattern}` as glob pattern: {e}"
	)
	}
	}
	}

Just for my future reference, The OpenBSD manual says this about this format:

A hostname or address may optionally be enclosed within ‘[’ and ‘]’ brackets then followed by ‘:’ and a non standard port number.

I can squash and rearrange the commits into 2 atomic commits if you wish :)

[TanmayArya-1p]

squashed into 2 atomic commits

[weihanglo]

Thanks!

View changes since this review