Bug 110322 – Element boundaries prohibit Japanese line break opportunities in case of Gaiji(image)

• Component: Layout and Rendering
• Status: RESOLVED CONFIGURATION CHANGED

[Zheng Xu]

Created attachment 189277 [details]
test case (compressed with html and png)

EXPECTED:
The gaiji(image) should be at second line.

REASON (SPEC): 
The closing character of any set of paired punctuation should be kept with the preceding character.
(From http://www.unicode.org/reports/tr14/ , CL: Close Punctuation (XB))

OTHERS:
I have also tried patch in https://bugs.webkit.org/show_bug.cgi?id=17427 and that patch can not fix this issue, so that I register a new one.

[Glenn Adams]

there is no Close Punctuation (CL) character in your test example

[Glenn Adams]

(In reply to comment #1)
> there is no Close Punctuation (CL) character in your test example

scratch that comment; i see U+3002 IDEOGRAPHIC FULL STOP is considered a CL by UAX14; however, the inline image (gaiji character) is NOT a character for the purposes of UAX#14, so the text you cite below doesn't apply

perhaps this needs to be discussed with the UTC to determine if UAX#14 needs to explicitly address this scenario?

[Koji Ishii]

(In reply to comment #2)
> the inline image (gaiji character) is NOT a character for the purposes of UAX#14, so the text you cite below doesn't apply
> perhaps this needs to be discussed with the UTC to determine if UAX#14 needs to explicitly address this scenario?

CSS already defines that replaced elements must be OBJECT REPLACEMENT CHARACTER U+FFFD, so no need to go to UTC.

[Glenn Adams]

(In reply to comment #3)
> (In reply to comment #2)
> > the inline image (gaiji character) is NOT a character for the purposes of UAX#14, so the text you cite below doesn't apply
> > perhaps this needs to be discussed with the UTC to determine if UAX#14 needs to explicitly address this scenario?
> 
> CSS already defines that replaced elements must be OBJECT REPLACEMENT CHARACTER U+FFFD, so no need to go to UTC.

ok, I see that CSS3 Text says: "The line breaking behavior of a replaced element or other atomic inline is equivalent to that of the Object Replacement Character (U+FFFC)."; however, it says nothing else, which reduces to the following for the included example:

ID CB CL ID ID

so, at least by UAX#14, breaks are permitted at each boundary in the example except CB CL

ok, that's reasonable; the question is then, does ICU currently implement this semantic and WK doesn't treat the image as FFFC? in any case, this bug will depend on bug 17427 since the image represents and element boundary

btw, koji, why doesn't the CSS3 line-break explicitly include the CL class characters in its semantics, such as U+3002?

[Zheng Xu]

Created attachment 191124 [details]
Patch v1 (based on "Proposed Patch v6" in 17427). 

This patch is based on patch in https://bugs.webkit.org/show_bug.cgi?id=17427.
I will submit it until patch of 17427 committed.

[Zheng Xu]

(In reply to comment #5)
> I will submit it until patch of 17427 committed.
It means, I will not apply it to EWS and set to r? before patch for 17427 is committed.

[Zheng Xu]

Created attachment 193727 [details]
Patch

[Zheng Xu]

Created attachment 193729 [details]
Patch

[Zheng Xu]

Created attachment 193739 [details]
Patch

[WebKit Review Bot]

Comment on attachment 193739 [details]
Patch

Attachment 193739 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://webkit-commit-queue.appspot.com/results/17160621

New failing tests:
css3/filters/effect-blur-hw.html
css1/text_properties/vertical_align.html
css2.1/t0803-c5501-imrgn-t-00-b-ag.html
css3/filters/effect-brightness-clamping.html
css1/formatting_model/height_of_lines.html
css2.1/t0803-c5504-imrgn-l-06-b-ag.html
css3/filters/effect-blur.html
http/tests/cache/subresource-failover-to-network.html
css3/filters/composited-during-animation-layertree.html
css2.1/t0803-c5505-imrgn-00-a-ag.html
css2.1/t0803-c5504-imrgn-l-05-b-ag.html
css3/filters/composited-during-transition-layertree.html
css3/filters/effect-brightness-hw.html
css2.1/t0804-c5507-ipadn-r-04-b-ag.html
css2.1/t0803-c5502-imrgn-r-06-b-ag.html
css2.1/t0804-c5509-ipadn-l-04-f-ag.html
accessibility/img-fallsback-to-title.html
css2.1/t0803-c5503-imrgn-b-00-b-a.html
http/tests/cache/subresource-fragment-identifier.html
css3/filters/effect-brightness-clamping-hw.html
css2.1/t0803-c5505-mrgn-00-b-ag.html
css2.1/t0803-c5502-imrgn-r-05-b-ag.html
http/tests/css/cross-fade-reload.html
compositing/masks/direct-image-mask.html
css2.1/t0803-c5505-mrgn-03-c-ag.html
css2.1/t0804-c5510-padn-00-b-ag.html
compositing/color-matching/image-color-matching.html
compositing/geometry/repaint-foreground-layer.html
compositing/reflections/simple-composited-reflections.html
css2.1/t0803-c5502-mrgn-r-00-c-ag.html

[Build Bot]

Comment on attachment 193739 [details]
Patch

Attachment 193739 [details] did not pass mac-ews (mac):
Output: http://webkit-commit-queue.appspot.com/results/17243074

New failing tests:
css3/filters/effect-blur-hw.html
css3/filters/effect-combined.html
css3/filters/effect-brightness-clamping.html
css3/filters/custom/custom-filter-shader-cache.html
css1/formatting_model/height_of_lines.html
css3/filters/effect-blur.html
css3/filters/effect-invert-hw.html
css3/filters/effect-grayscale.html
css1/text_properties/vertical_align.html
css3/filters/composited-during-animation-layertree.html
css3/filters/effect-brightness-hw.html
css3/filters/effect-hue-rotate.html
css3/filters/effect-opacity-hw.html
css3/filters/effect-invert.html
css3/filters/effect-combined-hw.html
accessibility/img-fallsback-to-title.html
css3/filters/effect-hue-rotate-hw.html
css3/filters/effect-contrast.html
css3/filters/effect-drop-shadow.html
css3/filters/effect-brightness.html
http/tests/cache/subresource-fragment-identifier.html
css3/filters/effect-brightness-clamping-hw.html
css3/filters/effect-drop-shadow-hw.html
css3/filters/effect-opacity.html
css3/filters/effect-grayscale-hw.html
http/tests/css/cross-fade-reload.html
compositing/masks/direct-image-mask.html
compositing/color-matching/image-color-matching.html
css3/filters/effect-contrast-hw.html
compositing/reflections/simple-composited-reflections.html

[Build Bot]

Comment on attachment 193739 [details]
Patch

Attachment 193739 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-commit-queue.appspot.com/results/17234397

New failing tests:
css3/filters/effect-blur-hw.html
css3/filters/effect-combined.html
css2.1/t0803-c5501-imrgn-t-00-b-ag.html
css3/filters/effect-brightness-clamping.html
css1/formatting_model/height_of_lines.html
css3/filters/effect-blur.html
css3/filters/effect-grayscale.html
css1/text_properties/vertical_align.html
css3/filters/composited-during-animation-layertree.html
http/tests/inspector/network/cached-resource-destroyed-too-big-discarded.html
css3/filters/effect-brightness-hw.html
css3/filters/effect-drop-shadow.html
css3/filters/effect-combined-hw.html
accessibility/img-fallsback-to-title.html
css3/filters/effect-hue-rotate-hw.html
css3/filters/effect-contrast.html
css3/filters/effect-brightness.html
http/tests/cache/subresource-fragment-identifier.html
css3/filters/effect-brightness-clamping-hw.html
css3/filters/effect-drop-shadow-hw.html
css3/filters/effect-grayscale-hw.html
http/tests/inspector/network/cached-resource-destroyed-moved-to-storage.html
http/tests/inspector/network/image-as-text-loading-data-url.html
canvas/philip/tests/2d.text.draw.fontface.notinpage.html
http/tests/css/cross-fade-reload.html
compositing/masks/direct-image-mask.html
compositing/color-matching/image-color-matching.html
compositing/geometry/repaint-foreground-layer.html
css3/filters/effect-contrast-hw.html
compositing/reflections/simple-composited-reflections.html

[Ahmad Saleem]

Created attachment 460808 [details]
Safari 15.5 matches other browsers

I am unable to reproduce this bug in Safari 15.5 on macOS after downloading test case locally, I don't have any Japanese font installed (if it was any pre-requisite).

Since all browsers are matching each other as shown in the picture and it is not matching with Expected Behavior. I am not sure whether it can be closed or whether it needs to be further speced or clarified? Please mark this accordingly (either putting it in RADAR) or closing as "RESOLVED CONFIGURATION CHANGED" or "RESOLVED INVALID". Thanks!