1358019 - word-break: break-all shouldn't create breaks between punctuation

Status: RESOLVED FIXED

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

Description

[fantasai]

'word-break: break-all' is not supposed to affect breaking between/around punctuation, only between pairs of letters (or symbols that behave like letters for line breaking).
Testcase: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cdiv%20style%3D%22word-break%3A%20break-all%20%3B%20font-size%3A%20200px%22%3EABC)%3C%2Fdiv%3E
Spec: https://www.w3.org/TR/css-text-3/#break-all

“In addition to ‘normal’ soft wrap opportunities, lines may break between any two **letters** (except where forbidden by the ‘line-break’ property). Hyphenation is not applied. This option is used mostly in a context where the text is predominantly using CJK characters with few non-CJK excerpts and it is desired that the text be better distributed on each line.”

Comment 2

[fantasai]

Copying over metadata from duplicate bug 1472598.

Xidorn points out that we might want to implement `line-break: anywhere` as part of this, to make sure that the use case has a solution.

Comment 3

[Karl Dubost💡 :karlcow]

Attachment: firefox, safari, chrome

Probably related to Bug 1272059 

Using the test case http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cdiv%20style%3D%22word-break%3A%20break-all%20%3B%20font-size%3A%20200px%22%3EABC)%3C%2Fdiv%3E

Firefox, Safari: They break at the punctuation.
Chrome: It breaks at the character before the punctuation.

Comment 4

[fantasai]

Koji is reporting that Chrome is running into compat problems because of this bug, fwiw. Any plans for fixing yet?

Comment 5

[Emilio Cobos Álvarez (:emilio)]

(In reply to fantasai from comment #4)

Koji is reporting that Chrome is running into compat problems because of this bug, fwiw. Any plans for fixing yet?

Comment 6

[Emilio Cobos Álvarez (:emilio)]

Err, I meant to reply:

Jonathan has a better handle on all our line-breaking stuff. I'm happy to give it a shot at some point though.

Comment 7

[Jonathan Kew [:jfkthame]]

It looks like our current implementation of word-break:break-all is rather too naïve: it essentially just treats every character as a potential break position -- e.g. see https://searchfox.org/mozilla-central/rev/00c0d068ece99717bea7475f7dc07e61f7f35984/intl/lwbrk/LineBreaker.cpp#901-902. (Maybe an older spec said something more like that?)

Anyhow, as currently spec'd, break-all doesn't simply mean "break anywhere"; it's defined in terms of revised line-break categories for certain classes of character. We should probably revise the impl in LineBreaker.cpp to better match the spec.

I've pushed a try job at https://treeherder.mozilla.org/#/jobs?repo=try&revision=012fd97eff0766a3add5c137777a867fe4db04e6 with an experimental patch; let's see what shakes out. Most likely there'll be a few oranges on the tree...

Comment 8

[Jonathan Kew [:jfkthame]]

Turns out there's an existing web-platform reftest (word-break-break-all-020.html) for this -- it's currently marked as failing, but if we fix break-all to follow the spec then we can remove that annotation.

Comment 9

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1358019 - Reimplement word-break:break-all to better match the CSS Text spec. r?masayuki

Comment 10

[Emilio Cobos Álvarez (:emilio)]

Nice! It would've taken me quite a bit to come up with that patch :-)

Comment 11

[fantasai]

@Emilio You could map the old behavior to line-break: anywhere? :)
@jfkthame Thank you~

Comment 12

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a93501fc7922
Reimplement word-break:break-all to better match the CSS Text spec. r=masayuki

Comment 13

[Cristian Brindusan [:cbrindusan]]

https://hg.mozilla.org/mozilla-central/rev/a93501fc7922