| Summary: | REGRESSION (Mountain Lion-Mavericks): Incorrect font used to render decomposed grapheme cluster in some cases | ||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | mitz | ||||||||||||||||||||||||
| Component: | Text | Assignee: | Myles C. Maxfield <mmaxfield> | ||||||||||||||||||||||||
| Status: | NEW --- | ||||||||||||||||||||||||||
| Severity: | Normal | CC: | ap, benjamin, cdumez, cmarcelo, dbates, ews-watchlist, mmaxfield, rniwa | ||||||||||||||||||||||||
| Priority: | P2 | Keywords: | Regression | ||||||||||||||||||||||||
| Version: | Safari Technology Preview | ||||||||||||||||||||||||||
| Hardware: | Unspecified | ||||||||||||||||||||||||||
| OS: | Unspecified | ||||||||||||||||||||||||||
| See Also: |
https://bugs.webkit.org/show_bug.cgi?id=185933 https://bugs.webkit.org/show_bug.cgi?id=68287 https://bugs.webkit.org/show_bug.cgi?id=186007 |
||||||||||||||||||||||||||
| Attachments: |
|
||||||||||||||||||||||||||
This got broken as soon as OS X 10.9 Mavericks. If this is a WebCore regression, this puts it between the tags Safari-536.30.2 and Safari-537.70. This doesn’t seem to be a WebKit regression: the bug happens in Safari 7.0 on Mavericks (537.71) but doesn’t happen in Safari 6.1.1 on Mountain Lion (537.73.11). Unless there’s some platform-specific #ifs in the code (which there don’t seem to be), the regression is probably due to a change in Core Text’s behavior between Mountain Lion and Mavericks. Results of CTLineCreateWithUniCharProvider:
Mountain Lion:
CTLine: run count = 1, string range = (0, 2), width = 12.8975, A/D/L = 47.2524/9.86816/0, glyph count = 1
{
CTRun: string range = (0, 2), string = "i\u0302", attributes = <CFBasicHash 0x7fcdca035400 [0x7fff77f34110]>{type = mutable dict, count = 3,
entries =>
0 : <CFString 0x7fff76938810 [0x7fff77f34110]>{contents = "NSKern"} = <CFNumber 0x7fcdca0531b0 [0x7fff77f34110]>{value = +0.0, type = kCFNumberFloat32Type}
1 : <CFString 0x7fff76938830 [0x7fff77f34110]>{contents = "NSLigature"} = <CFNumber 0x87 [0x7fff77f34110]>{value = +0, type = kCFNumberSInt32Type}
2 : <CFString 0x7fff769387f0 [0x7fff77f34110]>{contents = "NSFont"} = CTFont <name: Verdana, size: 47.000000, matrix: 0x0>
CTFontDescriptor <attributes: <CFBasicHash 0x7fcdca0257a0 [0x7fff77f34110]>{type = mutable dict, count = 1,
entries =>
2 : <CFString 0x7fff76939ef0 [0x7fff77f34110]>{contents = "NSFontNameAttribute"} = <CFString 0x7fff77f182a0 [0x7fff77f34110]>{contents = "Verdana"}
}
>
}
}
Mavericks:
<CTLine: 0x7fd1b3c6f0d0>{run count = 1, string range = (0, 2), width = 56.2256, A/D/L = 40.8413/12.25/0.430664, glyph count = 1, runs = (
<CTRun: 0x7fd1b3c5bfd0>{string range = (0, 2), string = "i\u0302", attributes = <CFBasicHash 0x7fd1b3d54830 [0x7fff7c97aeb0]>{type = mutable dict, count = 3,
entries =>
0 : <CFString 0x7fff7b915b30 [0x7fff7c97aeb0]>{contents = "NSKern"} = <CFNumber 0x7fd1b3c46160 [0x7fff7c97aeb0]>{value = +0.0, type = kCFNumberFloat32Type}
1 : <CFString 0x7fff7b915b50 [0x7fff7c97aeb0]>{contents = "NSLigature"} = <CFNumber 0x27 [0x7fff7c97aeb0]>{value = +0, type = kCFNumberSInt32Type}
2 : <CFString 0x7fff7b915b10 [0x7fff7c97aeb0]>{contents = "NSFont"} = <CTFont: 0x7fd1b3c61a20>{name = LastResort, size = 49.000000, matrix = 0x0, descriptor = <CTFontDescriptor: 0x7fd1b3d36330>{attributes = <CFBasicHash 0x7fd1b3d37520 [0x7fff7c97aeb0]>{type = mutable dict, count = 1,
entries =>
2 : <CFString 0x7fff7b917390 [0x7fff7c97aeb0]>{contents = "NSFontNameAttribute"} = <CFString 0x7fff7bd5c150 [0x7fff7c97aeb0]>{contents = "LastResort"}
}
>}}
}
}
)
}
(Ignore the sizes being 47 and 49 above, I used versions of the test case that had those sizes instead of the default). I believe the right way to solve this is to normalize the string. (In reply to Myles C. Maxfield from comment #6) > I believe the right way to solve this is to normalize the string. In the example in https://bugs.webkit.org/show_bug.cgi?id=186010#c0 Verdana doesn't support U+0302, but it does support U+00EE (In reply to Myles C. Maxfield from comment #6) > I believe the right way to solve this is to normalize the string. Is that how Core Text arrives at the right result? Created attachment 341407 [details]
Mountain Lion
Created attachment 341408 [details]
Mavericks
Attached screenshots showing how even the test case from bug 68287 regressed between Mountain Lion and Mavericks. In Mountain Lion, Core Text had the ability to tell us that we could use Verdana for the combining character sequence i+circumflex, even though Verdana didn’t have a glyph for the combining circumflex alone. The fix for bug 68287 was to take advantage of this ability, which was already benefiting the Cocoa text system. In Mavericks, that Core Text ability was gone. It regressed WebKit, but also the Cocoa text system, as can be seen with TextEdit. (In reply to mitz from comment #8) > Is that how Core Text arrives at the right result? Used to arrive :( See also: <rdar://problem/40583639> Created attachment 341418 [details]
macOS Browsers
Created attachment 341419 [details]
Edge
(In reply to Myles C. Maxfield from comment #15) > Created attachment 341419 [details] > Edge 😑 Is it the same Verdana though? It’s possible that the version that ships with Windows these days does have a standalone glyph for combining circumflex. (In reply to mitz from comment #16) > (In reply to Myles C. Maxfield from comment #15) > > Created attachment 341419 [details] > > Edge > > 😑 > > Is it the same Verdana though? Nope! > It’s possible that the version that ships > with Windows these days does have a standalone glyph for combining > circumflex. You are correct - Windows's Verdana has a glyph for U+0302. I'll upload a new screenshot. Created attachment 341447 [details]
Edge
Created attachment 341469 [details]
WIP
Attachment 341469 [details] did not pass style-queue:
ERROR: Source/WebCore/ChangeLog:8: You should remove the 'No new tests' and either add and list tests, or explain why no new tests were possible. [changelog/nonewtests] [5]
Total errors found: 1 in 10 files
If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 341469 [details] WIP Attachment 341469 [details] did not pass mac-ews (mac): Output: http://webkit-queues.webkit.org/results/7841629 New failing tests: imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-form-errors-misc.html fast/text/midword-break-before-surrogate-pair-2.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-href-errors-misc.html fast/regex/dom/unicodeCaseInsensitive.html fast/url/idna2008.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-href-errors-misc.html imported/w3c/web-platform-tests/css/css-text/word-break/word-break-break-all-007.html imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-href-errors-misc.html imported/blink/fast/text/international/complex-text-leading-space-wrapping.html Created attachment 341470 [details]
Archive of layout-test-results from ews101 for mac-sierra
The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: ews101 Port: mac-sierra Platform: Mac OS X 10.12.6
Comment on attachment 341469 [details] WIP Attachment 341469 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.webkit.org/results/7841733 New failing tests: imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-form-errors-misc.html fast/text/midword-break-before-surrogate-pair-2.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-href-errors-misc.html fast/regex/dom/unicodeCaseInsensitive.html fast/url/idna2008.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-href-errors-misc.html imported/w3c/web-platform-tests/css/css-text/word-break/word-break-break-all-007.html imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-href-errors-misc.html imported/blink/fast/text/international/complex-text-leading-space-wrapping.html Created attachment 341472 [details]
Archive of layout-test-results from ews104 for mac-sierra-wk2
The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews104 Port: mac-sierra-wk2 Platform: Mac OS X 10.12.6
Comment on attachment 341469 [details] WIP Attachment 341469 [details] did not pass ios-sim-ews (ios-simulator-wk2): Output: http://webkit-queues.webkit.org/results/7841777 New failing tests: imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-href-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-href-errors-misc.html fast/regex/dom/unicodeCaseInsensitive.html fast/url/idna2008.html imported/w3c/web-platform-tests/css/css-text/word-break/word-break-break-all-007.html imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-href-errors-misc.html imported/blink/fast/text/international/complex-text-leading-space-wrapping.html Created attachment 341476 [details]
Archive of layout-test-results from ews123 for ios-simulator-wk2
The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews123 Port: ios-simulator-wk2 Platform: Mac OS X 10.13.4
Comment on attachment 341469 [details] WIP Attachment 341469 [details] did not pass mac-debug-ews (mac): Output: http://webkit-queues.webkit.org/results/7841795 New failing tests: imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-form-errors-misc.html fast/text/midword-break-before-surrogate-pair-2.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-href-errors-misc.html fast/regex/dom/unicodeCaseInsensitive.html fast/url/idna2008.html imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-href-errors-misc.html imported/blink/fast/text/international/complex-text-leading-space-wrapping.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-href-errors-misc.html imported/w3c/web-platform-tests/css/css-text/word-break/word-break-break-all-007.html Created attachment 341477 [details]
Archive of layout-test-results from ews117 for mac-sierra
The attached test failures were seen while running run-webkit-tests on the mac-debug-ews.
Bot: ews117 Port: mac-sierra Platform: Mac OS X 10.12.6
Comment on attachment 341469 [details] WIP Attachment 341469 [details] did not pass mac-debug-ews (mac): Output: http://webkit-queues.webkit.org/results/7842513 New failing tests: imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-form-errors-misc.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-form-errors-misc.html fast/text/midword-break-before-surrogate-pair-2.html imported/w3c/web-platform-tests/encoding/legacy-mb-tchinese/big5/big5-encode-href-errors-misc.html fast/regex/dom/unicodeCaseInsensitive.html fast/url/idna2008.html imported/w3c/web-platform-tests/encoding/legacy-mb-korean/euc-kr/euckr-encode-href-errors-misc.html imported/w3c/web-platform-tests/css/css-text/word-break/word-break-break-all-007.html imported/w3c/web-platform-tests/encoding/legacy-mb-japanese/shift_jis/sjis-encode-href-errors-misc.html imported/blink/fast/text/international/complex-text-leading-space-wrapping.html Created attachment 341481 [details]
Archive of layout-test-results from ews116 for mac-sierra
The attached test failures were seen while running run-webkit-tests on the mac-debug-ews.
Bot: ews116 Port: mac-sierra Platform: Mac OS X 10.12.6
|
In the following test case: <div style="font-family: verdana, times;"> î and i̿ </div> the first grapheme cluster (lowercase I with combining circumflex) is rendered in Times, but it should be rendered in Verdana. This works correctly in OS X Mountain Lion 10.8.5, after <https://trac.webkit.org/r95391>, the fix for bug 68287, but got broken some time between then and r232220. The problem is that Font::canRenderCombiningCharacterSequence() started returning false for Verdana and the i+circumflex sequence.