Support for Windows Hidden Files

[alamb3142]

This PR adds a step into Files::next_visible_file to check for the hidden attribute on Windows systems. This causes it to behave more like a Windows user would expect it to.

Previously eza only hid dot files when not specifying -a or -A, it now hides dot files and all hidden files. Some types of links are hidden (file junctions are, not sure about hard links) because they're marked as hidden, but symlinks still display with or without the all flag.

[cafkafk]

The commit summary needs to conform to conventional commits. See this for more info.

[alamb3142]

Alternative to unpacking & repacking:

                #[cfg(windows)]
                if !self.dotfiles && file.as_ref().is_ok_and(|f| f.attributes().hidden) {
                    continue;
                }

                return Some(file);

May close this PR so I can start again with commits that adhere to the guidelines above

[cafkafk]

May close this PR so I can start again with commits that adhere to the guidelines above

That is an option but shouldn't be nescesarry, you can git rebase -i <commit before yours> and reword them. Then you just have to force push and it will be fixed.

[alamb3142]

Rebased & edited the commit to follow guidelines. Also implemented the change suggested by cfxegbert

[alamb3142]

@cafkafk I let GitHub merge the latest changes in, would you rather I rebase again so it's just the one commit?

[PThorpe92]

No worries, on merge, github gives you an option to remove it from commit history 👍 pretty sure automatically

[cafkafk]

I marked this as draft. Feel free to make it "ready for review" when you want us to take a look at it again.

[alamb3142]

Hey @cafkafk I'm ready for the code to be reviewed, it looks like it's just blocked by the comment about commit messages, which has been addressed

[PThorpe92]

Has this been tested on windows? @daviessm sorry to keep pinging you for windows stuff but you kinda brought that on yourself by being the only one with a windows box 😅 it is super appreciated tho ❤️ cuz having to deal with windows just isn't any fun
(btw you should come hang out on the gitter channel)

Otherwise, LGTM 👍

[daviessm]

Tested working on Windows, code looks sensible. Just one minor spelling nit in comments.

[cafkafk]

CI and Integration tests pass, and local testing confirms this doesn't break linux so LGTM 👍

[cafkafk]

as pointed out by @cfxegbert on element, is_ok_and was first stabilized in 1.70 (rust-lang/rust#110019), but we have to support 1.65 for gentoo, so unfortunately this will have to be reverted temporarily

[cafkafk]

@alamb3142 could you make a new pull request with the changes, then we can look into changing is_ok_and?

[alamb3142]

@cafkafk Yeah I'm happy to start a new PR, but is there any real risk with is_ok_and given that it's only called inside a #[cfg(windows)] block? Is that not sufficient to stop it from breaking on 1.65 Linux builds?

[cafkafk]

@cafkafk Yeah I'm happy to start a new PR, but is there any real risk with is_ok_and given that it's only called inside a #[cfg(windows)] block? Is that not sufficient to stop it from breaking on 1.65 Linux builds?

I mean, currently we specify MSRV as 1.65, and any windows user that below 1.70 will experience this right?

But also, we're discussing bumping the MSRV on element right now, maybe this isn't an issue after all, but to be determined.

[alamb3142]

Ah yeah that is a good point. I know most Windows users would prefer Scoop over Cargo for an install, but that's still potentially breaking it for some Windows users. Maybe I could commit the version that unpacks the result & repacks it, then submit a PR with this version for later when the MSRV is bumped?

[cafkafk]

Ah yeah that is a good point. I know most Windows users would prefer Scoop over Cargo for an install, but that's still potentially breaking it for some Windows users. Maybe I could commit the version that unpacks the result & repacks it, then submit a PR with this version for later when the MSRV is bumped?

That sounds like a good idea, I'll await it