Regression in `text_input`

[edwloef]

Is your issue REALLY a bug?

• My issue is indeed a bug!
• I am not crazy! I will not fill out this form just to ask a question or request a feature. Pinky promise.

Is there an existing issue for this?

• I have searched the existing issues.

Is this issue related to iced?

• My hardware is compatible and my graphics drivers are up-to-date.

What happened?

text_input treats any cursor that's in-bounds of the widget but out-of-bounds of the actual text as being at the end of that text. This is incorrect for non-left-aligned text input. Try clicking within the text_input but left of the contained text:

fn main() -> iced::Result {
    iced::run((), view)
}

fn view(_: &()) -> impl Into<iced::Element<'_, iced::Never>> {
    iced::widget::text_input("placeholder", "value")
        .width(100)
        .align_x(iced::Center)
        .on_input(|_| todo!())
}

What is the expected behavior?

Cursors that are in-bounds of the widget but left of the actual text should be treated as being at the beginning of that text.

Version

master

Operating System

Linux

Do you have any log output?

[hecrj]

This seems like a cosmic-text bug to me; Buffer::hit returns the end of buffer incorrectly.

[edwloef]

That would surprise me, since this worked before the big text::Input refactor 🤔 was the old code manually working around this fact?

[hecrj]

Exactly. We weren't using cosmic-text for editing logic.

[edwloef]

In that case, another regression where I'm now unsure who's at fault: double-click-drag selection doesn't seem to work at all anymore. E.g. here:

fn main() -> iced::Result {
    iced::run((), view)
}

fn view(_: &()) -> impl Into<iced::Element<'_, iced::Never>> {
    iced::widget::text_input("placeholder", "some multi word text here")
        .width(300)
        .on_input(|_| todo!())
}

[hecrj]

@edwloef I think d188d6f should fix that.

[edwloef]

Yup, working now :)

[edwloef]

This seems like a cosmic-text bug to me; Buffer::hit returns the end of buffer incorrectly.

This should, I think, be fixed by pop-os/cosmic-text#528.