feat: support child window on linux, closes #273

[keiya01]

What kind of change does this PR introduce?

• Bugfix
• Feature
• Docs
• New Binding issue #___
• Code style update
• Refactor
• Build-related changes
• Other, please describe:

Does this PR introduce a breaking change?

• Yes, and the changes were approved in issue #___
• No

Checklist

• When resolving issues, they are referenced in the PR's title (e.g fix: remove a typo, closes #___, #___)
• A change file is added if any packages will require a version bump due to this PR per the instructions in the readme.
• I have added a convincing reason for adding this feature, if necessary

Other information

[wusyong]

Tested it and looking great! Thanks!
I think at this rate we could bring this method to general window builder method?
This feature isn't relevant to mobile afaik.

[vdvman1]

If it is made a general window builder method, it would probably be best to document the behaviour differences for each platform, if there are any.
I know at the very least that "child window" on MacOS and Windows have different behaviour, where the "child window" behaviour on MacOS is more akin to "owned window" in Windows, AFAIK

[keiya01]

I fixed these review.
Thank you for reviewing!

[vdvman1]

Might be good to use OwnedBy on windows rather than ChildOf, as I believe that would be more consistent window behaviour with the other platforms (I'm not sure though, that should probably be tested, I haven't/can't test how it behaves on each platform myself)

[wusyong]

It's look good to me but I feel like @amrbashir is best to review the difference between OwnedBy and ChildOf

[amrbashir]

as @vdvman1 said, ChildOf on Windows doesn't behave like other platforms. OwnedBy is the closest to other platforms.

With that said, I don't think it is a good idea to provide a general builder method for all 3 platforms. Instead we should have them under extensions trait so macOS would have WindowBuilderExtMacOS::with_parent, and Windows would have WindowBuilderExtWindows::with_parent and WindowBuilderExtWindows::with_owner and Linux would have WindowBuilderExtLinux::with_transient_for.

[keiya01]

I see.
There are some slight difference between these three methods.
If we split these method for each platform, users may not be confused.

[FabianLars]

On a somewhat off-topic note, can you do normal pushes instead of force-pushing in the future? Would make it much easier to follow along :)

Imo we should add this to our contributing guidelines too 🤔

[keiya01]

On a somewhat off-topic note, can you do normal pushes instead of force-pushing in the future? Would make it much easier to follow along :)

Oh, sure.
Thank you for your advice 🙇‍♂️

Imo we should add this to our contributing guidelines too

yeah even if we separate these method, I think we should keep a context as comment.

[wusyong]

@keiya01 So sorry you have to revert. I think this is the part I'll usually miss some context. 🙇‍♂️

@FabianLars I'm okay with both as long as commits are clear. I also will do some force push if commits are all just want to make clippy and fmt happy.

[keiya01]

I reverted in Revert "fix: ganeral window builder".
BTW by bundling commit with force push, I somtimes be eased to revert lol

[keiya01]

I will add some description later.

[keiya01]

I wonder why does android fail...

[wusyong]

Probably github CI being funny again. I'll rerun it tomorrow.

[wusyong]

Looks like it's because of rust-mobile/ndk#286
Let's ignore this for now. I'll update ndk later.
@amrbashir Does this look good to you?