[css-text][text-spacing] Propose to make allow-end as the default behavior for the line end

[kojiishi]

Currently, the default behavior of line-end for the text-spacing-trim property is defined as:

set fullwidth closing punctuation with half-width glyphs (flush) at the end of each line

If the author specifies the allow-end:

Set fullwidth closing punctuation with half-width glyphs (flush) at the end of each line if it does not otherwise fit prior to justification, else set the punctuation with full-width glyphs.

From the dogfood feedback, I'd like to propose to change the default behavior to allow-end. We can add a value for the current default behavior if needed.

This proposal is because when you're typing in editable content. The current default behavior trims the last character as you type, making them difficult to distinguish whether it's full-width code point or half-width code point, until users type the next character. Some IMEs on desktop platforms remember which was used last time, making this symptom more difficult.

This happens quite frequently for Japanese/Simplified Chinese/Korean, where the comma and full stops (periods) are in the "closing" shape. Not as frequent for Traditional Chinese.

I don't know the motivation and the past discussions to make the current behavior as the default. If it has big reasons and the WG decided to make a trade-off, could someone please share?

From my current perspective, the feedback looks reasonable to me, and the difference between the current default and allow-end looks hardly noticeable, so this looks like a reasonable change.

Thoughts? /cc @fantasai @frivoal @nt1m @vitorroriz @Clqsin45 @MurakamiShinyu @himorin @kidayasuo

[MurakamiShinyu]

The current default behavior of line-end was my suggestion:

• [css-text-4] text-spacing: trim-end is better than allow-end for the normal value #7055

However, it is fine to me changing this default, as I commented in #9511 (comment).

[frivoal]

The change adopted in #9511 means that the initial value is now normal, and it does have an allow-end behavior. I think we can consider this issue solved?

[kojiishi]

I think we can consider this issue solved?

Yes, thank you.