AX: Add tests for RTL/LTR bidi text #46728
Conversation
|
EWS run on current version of this PR (hash 0be54d7) |
🧪 style|
EWS run on previous version of this PR (hash 0be54d7)
|
| staticTextTextMarkerRange = staticText.textMarkerRangeForElement(staticText); | ||
| output += expect("staticText.textMarkerRangeLength(staticTextTextMarkerRange)", "28"); | ||
|
|
||
| output += expectRectWithVariance("staticText.boundsForRange(0, 28)", -1, -1, 188, 18, 1); |
There was a problem hiding this comment.
expectRectWithVariance is good, but it still does feel like this test might tend to break if things shift by a couple of pixels.
Would it be possible to instead iterate over the characters and, using the bounds, just assert whether or not each index moved to the left or to the right?
0be54d7 to
7ebbb87
Compare
|
EWS run on previous version of this PR (hash 7ebbb87)
|
7ebbb87 to
a3dcd16
Compare
|
EWS run on previous version of this PR (hash a3dcd16) |
|
EWS run on previous version of this PR (hash dab773c)
|
dab773c to
1a47eec
Compare
|
EWS run on previous version of this PR (hash 1a47eec) |
1a47eec to
412b05d
Compare
|
EWS run on previous version of this PR (hash 412b05d) |
| <meta charset="utf-8"> | ||
| <script src="../../resources/js-test.js"></script> | ||
| <script src="../../resources/accessibility-helper.js"></script> | ||
| <style> | ||
| .unicode-bidi-normal { | ||
| unicode-bidi: normal; | ||
| } | ||
| .unicode-bidi-embed { | ||
| unicode-bidi: embed; | ||
| } | ||
| .unicode-bidi-bidi-override { | ||
| unicode-bidi: bidi-override; | ||
| } | ||
| .unicode-bidi-plaintext { | ||
| unicode-bidi: plaintext; | ||
| } | ||
| .unicode-bidi-isolate { | ||
| unicode-bidi: isolate; | ||
| } | ||
| .unicode-bidi-isolate-override { | ||
| unicode-bidi: isolate-override; | ||
| } | ||
| </style> |
There was a problem hiding this comment.
I think typically in other accessibility layout tests, we don't indent things in the <head>.
| <meta charset="utf-8"> | ||
| <script src="../../resources/js-test.js"></script> | ||
| <script src="../../resources/accessibility-helper.js"></script> | ||
| <style> | ||
| .unicode-bidi-normal { | ||
| unicode-bidi: normal; | ||
| } | ||
| .unicode-bidi-embed { | ||
| unicode-bidi: embed; | ||
| } | ||
| .unicode-bidi-bidi-override { | ||
| unicode-bidi: bidi-override; | ||
| } | ||
| .unicode-bidi-plaintext { | ||
| unicode-bidi: plaintext; | ||
| } | ||
| .unicode-bidi-isolate { | ||
| unicode-bidi: isolate; | ||
| } | ||
| .unicode-bidi-isolate-override { | ||
| unicode-bidi: isolate-override; | ||
| } | ||
| </style> |
There was a problem hiding this comment.
I think typically in other accessibility layout tests, we don't indent things in the <head>.
|
Your commit message (and PR description) needs to follow a very specific format, or else it will get rejected by the commit queue: |
|
|
||
| PASS: rangeLength === 28 | ||
| PASS: maxBounds === 188 | ||
| PASS: currentBounds > lastBounds === true |
There was a problem hiding this comment.
To better catch failures down the line, I wonder if each pass should print which character is being evaluated, in addition to whether this condition passes?
There was a problem hiding this comment.
Even better, how about just count the number of characters going each direction.
I'd like something along the lines of:
PASS: characters 0...20 are LTR
PASS: characters 20...26 are RTL
etc.
| This test verifies that a combination of Arabic and English is read correctly by Accessibility, when using CSS unicode-bidi property. | ||
|
|
||
| PASS: rangeLength === 28 | ||
| PASS: maxBounds === 188 |
There was a problem hiding this comment.
Maybe delete this line. It looks like it's failing on the bots because we don't get exactly the same number of pixels with and without ITM.
412b05d to
4a93a5a
Compare
|
EWS run on previous version of this PR (hash 4a93a5a) |
4a93a5a to
8625023
Compare
|
EWS run on previous version of this PR (hash 8625023) |
8625023 to
1d6fda0
Compare
|
EWS run on previous version of this PR (hash 1d6fda0) |
1d6fda0 to
34e6ce6
Compare
|
EWS run on previous version of this PR (hash 34e6ce6) |
34e6ce6 to
939c13f
Compare
|
EWS run on previous version of this PR (hash 939c13f) |
939c13f to
540df70
Compare
|
EWS run on previous version of this PR (hash 540df70) |
540df70 to
47dbfb7
Compare
|
EWS run on previous version of this PR (hash 47dbfb7) |
47dbfb7 to
f6d2cd5
Compare
|
EWS run on previous version of this PR (hash f6d2cd5) |
|
Commit message contains (OOPS!), blocking PR #46728. Details: Build #23436 |
f6d2cd5 to
199a68d
Compare
|
EWS run on current version of this PR (hash 199a68d) |
https://bugs.webkit.org/show_bug.cgi?id=294460 rdar://153327595 Reviewed by Tyler Wilcock. Add tests for bidi text: English-Arabic and English-Hebrew. This test counts the number of LTR and RTL positions within strings based on textmarkers. There are six strings tested, each with a unique value applied to its unicode-bidi css property. * LayoutTests/accessibility/text-marker/text-marker-bidi-element-arabic-expected.txt: Added. * LayoutTests/accessibility/text-marker/text-marker-bidi-element-arabic.html: Added. * LayoutTests/accessibility/text-marker/text-marker-bidi-element-hebrew-expected.txt: Added. * LayoutTests/accessibility/text-marker/text-marker-bidi-element-hebrew.html: Added. * LayoutTests/platform/ios/TestExpectations: Canonical link: https://commits.webkit.org/296764@main
199a68d to
c275e78
Compare
|
Committed 296764@main (c275e78): https://commits.webkit.org/296764@main Reviewed commits have been landed. Closing PR #46728 and removing active labels. |
c275e78
199a68d
🛠 playstation