Closed Bug 1516429 Opened 6 years ago Closed 6 years ago

[wpt-sync] Sync PR 14673 - [css-text] Arabic cursive joining and ZWJ

Categories

(Core :: CSS Parsing and Computation, enhancement, P4)

Product:
Core
Component:
CSS Parsing and Computation
Type:
enhancement

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla69
Tracking Flags:
Tracking Status
firefox69 --- fixed

People

(Reporter: wpt-sync, Unassigned)

References

(
URL
)

Details

(Whiteboard: [wptsync downstream])

Sync web-platform-tests PR 14673 into mozilla-central (this bug is closed when the sync is complete). PR: https://github.com/web-platform-tests/wpt/pull/14673 Details from upstream follow. Florian Rivoal <git@florian.rivoal.net> wrote: > [css-text] Arabic cursive joining and ZWJ > > This is a tentative test. Review from @r12a, @fantasai or @behdad very much appreciated. > > As far as I understand: > > 1. putting a U+200D ZERO WIDTH JOINER after an arabic letter that is otherwise not followed by more arabic letters can be used force it to take it's medial form instead of its final form > 2. putting a U+200D ZERO WIDTH JOINER before an arabic letter than is otherwise not preceeded by more arabic letters can be used to force it to take its media form instead of its initial form. > > As was working on other tests, It looked like all browsers were correctly doing 1, but Chrome and Safari were not doing 2. > > So I though I'd write a test for that, and that's what this PR is doing. > > However: > 1. Is this actually mandated by a spec? If yes, which? I couldn't identify the link I should put in the `<link rel="help" href="?????">` > 2. This test fails in Chrome and Safari due to not doing 2, which is what I expected. See this screenshot: > <img width="507" alt="screenshot 2018-12-26 23 44 42" src="https://user-images.githubusercontent.com/113268/50449084-44618000-0968-11e9-8bdf-f5024efa003f.png"> > > > 3. It also fails in Firefox for reasons that are less clear to me, since I don't actually read any arabic-script language. As far as I can tell, Firefox is doing the right thin in terms of shaping, but uses slightly different gliphs for `U+FE98 : ARABIC LETTER TEH MEDIAL FORM / U+FEE6 : ARABIC LETTER NOON FINAL FORM` vs `U+062A : ARABIC LETTER TEH / U+0646 : ARABIC LETTER NOON`. Presumably that may be coming from the font, which would make it conforming even if annoying, but that's a bit surprising since the same font (Geeza Pro) doesn't do that in in Chrome. Is that what's going on? If it's not what's going on, what could be? Does this look like a bug in Firefox, or is my test relying on something that cannot be relied on? > ![screenshot 2018-12-26 23 54 32](https://user-images.githubusercontent.com/113268/50449280-9c4cb680-0969-11e9-9f75-d1e0cc5d80b0.png) > > >
Component: web-platform-tests → CSS Parsing and Computation
Product: Testing → Core
Whiteboard: [wptsync downstream] → [wptsync downstream error]
Whiteboard: [wptsync downstream error] → [wptsync downstream]
Whiteboard: [wptsync downstream] → [wptsync downstream error]
Whiteboard: [wptsync downstream error] → [wptsync downstream]
Failed to get results from try push
Pushed by james@hoppipolla.co.uk: https://hg.mozilla.org/integration/mozilla-inbound/rev/59bf3ea6bc54 [wpt PR 14673] - [css-text] Arabic cursive joining and ZWJ, a=testonly

https://hg.mozilla.org/mozilla-central/rev/59bf3ea6bc54

Status: NEW → RESOLVED
Closed: 6 years ago
status-firefox69: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla69
You need to log in before you can comment on or make changes to this bug.