1325807 - Ideographic space (U+3000) should provide a soft wrap opportunity even with "word-break: keep-all"

Status: RESOLVED DUPLICATE of bug 1719535

(Core :: Layout: Text and Fonts, defect, P3)

Description

[Xidorn Quan [:xidorn] UTC+11]

Attachment: testcase

In the attached testcase, ideographic space (U+3000) is used to separate words, and "word-break: keep-all" is applied.

In this case, there should be a soft wrap opportunity after each ideographic space here.

In Blink, the text is broken after ideographic spaces, while WebKit behaves the same as Gecko. But I think it should provide wrap opportunity here, since it is a whitespace anyway.

I guess this behavior was changed by bug 765166.

Comment 1

[Xidorn Quan [:xidorn] UTC+11]

It seems Blink doesn't break line between ideographic space. So basically Blink reverts the WebKit's behavior to what Gecko had before bug 765166 :/

Comment 2

[Xidorn Quan [:xidorn] UTC+11]

Attachment: testcase 2

Oh, Blink also provide soft wrap opportunity for punctunations, which also makes sense to me...

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Screenshot of Edge

Edge behaves very odd...

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Screenshot of Chromium's testcase

Hmm, they test keep-all with punctuation. So, it must be *their* expected behavior.

Issue: https://bugs.chromium.org/p/chromium/issues/detail?id=141792
Review: https://codereview.chromium.org/1094863007

According to the patch and comment 11, they ban some break opportunities. Ishii-san said "what the spec[1] says is not that complicated; just disable soft wrap opportunities between "typographic letter units" (which is L*/N* grapheme clusters)".

Example 6 of the spec is what this case is?
https://www.w3.org/TR/css-text-3/#word-break

Comment 5

[Xidorn Quan [:xidorn] UTC+11]

(In reply to Xidorn Quan [:xidorn] (UTC+10) from comment #1)
> It seems Blink doesn't break line between ideographic space. So basically
> Blink reverts the WebKit's behavior to what Gecko had before bug 765166 :/

Ah, I mean, *WebKit* doesn't break line between ideographic space. Blink does so.

Comment 6

[Xidorn Quan [:xidorn] UTC+11]

(In reply to Masayuki Nakano [:masayuki] (Mozilla Japan) from comment #4)
> Created attachment 8821840 [details]
> Screenshot of Chromium's testcase
> 
> Hmm, they test keep-all with punctuation. So, it must be *their* expected
> behavior.

I think their expected behavior makes sense, although I have one concern that some punctuation in languages like English may not be a word separator, e.g. "e.g.", or "U.S.A". So I guess half-width punctuations probably should not provide break opportunity.

> According to the patch and comment 11, they ban some break opportunities.
> Ishii-san said "what the spec[1] says is not that complicated; just disable
> soft wrap opportunities between "typographic letter units" (which is L*/N*
> grapheme clusters)".

The spec is quite loose on this. It forces several rules, and impls are free to do whatever they want otherwise.

Comment 7

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

(In reply to Xidorn Quan [:xidorn] (UTC+10) from comment #6)
> although I have one concern
> that some punctuation in languages like English may not be a word separator,
> e.g. "e.g.", or "U.S.A". So I guess half-width punctuations probably should
> not provide break opportunity.

That's what I was troubled when I worked on breaking lines in URL. (E.g., smilies too)

At that time, we took CONSERVATIVE_BREAK_RANGE approach:
https://dxr.mozilla.org/mozilla-central/rev/1156db49e976173fc3cf90d2126456fd1e2bae4b/intl/lwbrk/nsJISx4051LineBreaker.cpp#624-632

Comment 8

[Makoto Kato [:m_kato]]

Legacy line segmenter breaks before and after U+0x3000 now. But ICU4X segmenter is expected behavior.

If I am wrong, please reopen this.