On Wed, 25 Mar 2026 19:02:25 GMT, Marius Hanl <[email protected]> wrote:

>> Looking at the code review for the commit that caused this, that is a 
>> perfect example of how diffs can mislead. The fact that it introduced dead 
>> code was not visible from the diffs, only if you looked at the whole context.
>
> And I even made a (IMO good) suggestion to improve the if condition 🙂
> -> I made this comment because I felt like the if-else condition is 
> confusing. So turns out, it is.
> 
> I would therefore recommend to improve it now and remove the else block. It 
> is empty anyway.

I think it's out of scope for this PR, for several reasons.

This code relies on current, undocumented behavior of the caret shape being 
either a single line or two lines encoded as `MoveTo`-`LineTo` pairs.  There 
are plans to change the caret shape for certain BIDI/RTL scenarios to indicate 
the direction, so the old assumptions might not hold anymore.

Also, in jfx25 we've added `LayoutInfo`/`CaretInfo` APIs in the `Text` and 
`TextFlow`, which should provide the caret information without relying on 
implementation details.

related umbrellas:
https://bugs.openjdk.org/browse/JDK-8343557
https://bugs.openjdk.org/browse/JDK-8300569

-------------

PR Review Comment: https://git.openjdk.org/jfx/pull/2128#discussion_r2996129050

Reply via email to