[webkit-changes] [263998] trunk

Mon, 06 Jul 2020 17:00:20 -0700

Title: [263998] trunk
Revision
263998
Author
mmaxfi...@apple.com
Date
2020-07-06 16:59:23 -0700 (Mon, 06 Jul 2020)

Log Message

Locale-specific quotes infrastructure needs to compare locale strings properly
https://bugs.webkit.org/show_bug.cgi?id=213827

Reviewed by Darin Adler.

Source/WebCore:

Before this patch, WebKit is selecting which quotes to display on <q>
elements by doing a raw strcmp() on the locale string with a big table
of locale strings. strcmp() is the wrong way to compare locale strings.

The HTML spec has a list of locales and their associated quotes[1].
It is formulated in terms of CSS using the "lang()" pseudoclass.
The spec of the lang() pseudoclass[2] describes that locale comparison
needs to be done according to section 3.3.2 in RFC4647[3].

This algorithm is a pretty general algorithm, and implementing it naively
would mean turning our O(log(n)) algorithm into a O(n) algorithm, which
would be unfortunate. Instead, we can use a few observations about the
set of locale strings we are comparing against, in order to preserve the
O(log(n)) runtime:
- All the locales have either 1 or 2 subtags
- None of the subtags in any of the ranges are wildcards
- The list is sorted, so a locale string that is a prefix of another one
      is listed before it.

[1] https://html.spec.whatwg.org/multipage/rendering.html#quotes
[2] https://drafts.csswg.org/selectors-4/#the-lang-pseudo
[3] https://tools.ietf.org/html/rfc4647#page-10

Test: fast/css-generated-content/quotes-lang-2.html

* WebCore.xcodeproj/xcshareddata/xcschemes/WebCore.xcscheme:
* rendering/RenderQuote.cpp:
(WebCore::subtagCompare):
(WebCore::quoteTableLanguageComparisonFunction):
(WebCore::quotesForLanguage):
(WebCore::RenderQuote::computeText const):

LayoutTests:

* fast/css-generated-content/quotes-lang-2-expected.html: Added.
* fast/css-generated-content/quotes-lang-2.html: Added.

Modified Paths

Added Paths

Diff

Modified: trunk/LayoutTests/ChangeLog (263997 => 263998)


--- trunk/LayoutTests/ChangeLog	2020-07-06 23:32:05 UTC (rev 263997)
+++ trunk/LayoutTests/ChangeLog	2020-07-06 23:59:23 UTC (rev 263998)
@@ -1,3 +1,13 @@
+2020-07-06  Myles C. Maxfield  <mmaxfi...@apple.com>
+
+        Locale-specific quotes infrastructure needs to compare locale strings properly
+        https://bugs.webkit.org/show_bug.cgi?id=213827
+
+        Reviewed by Darin Adler.
+
+        * fast/css-generated-content/quotes-lang-2-expected.html: Added.
+        * fast/css-generated-content/quotes-lang-2.html: Added.
+
 2020-07-06  Zalan Bujtas  <za...@apple.com>
 
         [Win] No need to mark tables/mozilla_expected_failures/other/empty_cells.html with [Failure] anymore.

Added: trunk/LayoutTests/fast/css-generated-content/quotes-lang-2-expected.html (0 => 263998)


--- trunk/LayoutTests/fast/css-generated-content/quotes-lang-2-expected.html	                        (rev 0)
+++ trunk/LayoutTests/fast/css-generated-content/quotes-lang-2-expected.html	2020-07-06 23:59:23 UTC (rev 263998)
@@ -0,0 +1,38 @@
+<!DOCTYPE html>
+
+<style>
+q {
+    font-family: "Times";
+}
+</style>
+
+<p>
+    You should see the quotes for the specified language on each line below,
+    and not just basic " and ' characters unless no language is specified.
+</p>
+
+<q style="quotes: '\201e' '\201d' '\201a' '\2019';"><q>agq</q></q>
+<q style=""><q>ag</q></q>
+<q style=""><q>gq</q></q>
+<q style=""><q>az</q></q>
+<q style="quotes: '\201c' '\201d' '\2018' '\2019';"><q>zh-han</q></q>
+<q style="quotes: '\201c' '\201d' '\2018' '\2019';"><q>zh-ant</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>az-cyrl-abc</q></q>
+<q style="quotes: '\201e' '\201c' '\201a' '\2018';"><q>de-abc</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>de-ch-abc</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>de-abc-ch</q></q>
+<q style="quotes: '\00ab' '\00bb' '\00ab' '\00bb';"><q>fr-abc</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-ca-abc</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-ch-abc</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-abc-ca</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-abc-ch</q></q>
+<q style="quotes: '\201c' '\201d' '\2018' '\2019';"><q>af-x</q></q>
+<q style="quotes: '\201e' '\201c' '\201a' '\2018';"><q>de-x</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>de-ch-x</q></q>
+<q style="quotes: '\201e' '\201c' '\201a' '\2018';"><q>de-x-ch</q></q>
+<q style="quotes: '\00ab' '\00bb' '\00ab' '\00bb';"><q>fr-x</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-ca-x</q></q>
+<q style="quotes: '\00ab' '\00bb' '\2039' '\203a';"><q>fr-ch-x</q></q>
+<q style="quotes: '\00ab' '\00bb' '\00ab' '\00bb';"><q>fr-x-ca</q></q>
+<q style="quotes: '\00ab' '\00bb' '\00ab' '\00bb';"><q>fr-x-ch</q></q>
+<q style="quotes: '\201c' '\201d' '\2018' '\2019';"><q>en-ch</q></q>

Added: trunk/LayoutTests/fast/css-generated-content/quotes-lang-2.html (0 => 263998)


--- trunk/LayoutTests/fast/css-generated-content/quotes-lang-2.html	                        (rev 0)
+++ trunk/LayoutTests/fast/css-generated-content/quotes-lang-2.html	2020-07-06 23:59:23 UTC (rev 263998)
@@ -0,0 +1,38 @@
+<!DOCTYPE html>
+
+<style>
+q {
+    font-family: "Times";
+}
+</style>
+
+<p>
+    You should see the quotes for the specified language on each line below,
+    and not just basic " and ' characters unless no language is specified.
+</p>
+
+<q lang="agq"><q>agq</q></q>
+<q lang="ag"><q>ag</q></q>
+<q lang="gq"><q>gq</q></q>
+<q lang="az"><q>az</q></q>
+<q lang="zh-han"><q>zh-han</q></q>
+<q lang="zh-ant"><q>zh-ant</q></q>
+<q lang="az-cyrl-abc"><q>az-cyrl-abc</q></q>
+<q lang="de-abc"><q>de-abc</q></q>
+<q lang="de-ch-abc"><q>de-ch-abc</q></q>
+<q lang="de-abc-ch"><q>de-abc-ch</q></q>
+<q lang="fr-abc"><q>fr-abc</q></q>
+<q lang="fr-ca-abc"><q>fr-ca-abc</q></q>
+<q lang="fr-ch-abc"><q>fr-ch-abc</q></q>
+<q lang="fr-abc-ca"><q>fr-abc-ca</q></q>
+<q lang="fr-abc-ch"><q>fr-abc-ch</q></q>
+<q lang="af-x"><q>af-x</q></q>
+<q lang="de-x"><q>de-x</q></q>
+<q lang="de-ch-x"><q>de-ch-x</q></q>
+<q lang="de-x-ch"><q>de-x-ch</q></q>
+<q lang="fr-x"><q>fr-x</q></q>
+<q lang="fr-ca-x"><q>fr-ca-x</q></q>
+<q lang="fr-ch-x"><q>fr-ch-x</q></q>
+<q lang="fr-x-ca"><q>fr-x-ca</q></q>
+<q lang="fr-x-ch"><q>fr-x-ch</q></q>
+<q lang="en-ch"><q>en-ch</q></q>

Modified: trunk/Source/WebCore/ChangeLog (263997 => 263998)


--- trunk/Source/WebCore/ChangeLog	2020-07-06 23:32:05 UTC (rev 263997)
+++ trunk/Source/WebCore/ChangeLog	2020-07-06 23:59:23 UTC (rev 263998)
@@ -1,3 +1,42 @@
+2020-07-06  Myles C. Maxfield  <mmaxfi...@apple.com>
+
+        Locale-specific quotes infrastructure needs to compare locale strings properly
+        https://bugs.webkit.org/show_bug.cgi?id=213827
+
+        Reviewed by Darin Adler.
+
+        Before this patch, WebKit is selecting which quotes to display on <q>
+        elements by doing a raw strcmp() on the locale string with a big table
+        of locale strings. strcmp() is the wrong way to compare locale strings.
+
+        The HTML spec has a list of locales and their associated quotes[1].
+        It is formulated in terms of CSS using the "lang()" pseudoclass.
+        The spec of the lang() pseudoclass[2] describes that locale comparison
+        needs to be done according to section 3.3.2 in RFC4647[3].
+
+        This algorithm is a pretty general algorithm, and implementing it naively
+        would mean turning our O(log(n)) algorithm into a O(n) algorithm, which
+        would be unfortunate. Instead, we can use a few observations about the
+        set of locale strings we are comparing against, in order to preserve the
+        O(log(n)) runtime:
+        - All the locales have either 1 or 2 subtags
+        - None of the subtags in any of the ranges are wildcards
+        - The list is sorted, so a locale string that is a prefix of another one
+              is listed before it.
+
+        [1] https://html.spec.whatwg.org/multipage/rendering.html#quotes
+        [2] https://drafts.csswg.org/selectors-4/#the-lang-pseudo
+        [3] https://tools.ietf.org/html/rfc4647#page-10
+
+        Test: fast/css-generated-content/quotes-lang-2.html
+
+        * WebCore.xcodeproj/xcshareddata/xcschemes/WebCore.xcscheme:
+        * rendering/RenderQuote.cpp:
+        (WebCore::subtagCompare):
+        (WebCore::quoteTableLanguageComparisonFunction):
+        (WebCore::quotesForLanguage):
+        (WebCore::RenderQuote::computeText const):
+
 2020-07-06  Daniel Bates  <daba...@apple.com>
 
         [iOS] WAKWindow should override -resignFirstResponder and clear state

Modified: trunk/Source/WebCore/rendering/RenderQuote.cpp (263997 => 263998)


--- trunk/Source/WebCore/rendering/RenderQuote.cpp	2020-07-06 23:32:05 UTC (rev 263997)
+++ trunk/Source/WebCore/rendering/RenderQuote.cpp	2020-07-06 23:59:23 UTC (rev 263998)
@@ -89,8 +89,55 @@
 
 #endif // ASSERT_ENABLED
 
+struct SubtagComparison {
+    size_t keyLength;
+    size_t keyContinue;
+    size_t rangeLength;
+    size_t rangeContinue;
+    int comparison;
+};
+
+static SubtagComparison subtagCompare(const char* key, const char* range)
+{
+    SubtagComparison result;
+
+    result.keyLength = strlen(key);
+    result.keyContinue = result.keyLength;
+    if (auto* hyphenPointer = strchr(key, '-')) {
+        result.keyLength = hyphenPointer - key;
+        result.keyContinue = result.keyLength + 1;
+    }
+
+    result.rangeLength = strlen(range);
+    result.rangeContinue = result.rangeLength;
+    if (auto* hyphenPointer = strchr(range, '-')) {
+        result.rangeLength = hyphenPointer - range;
+        result.rangeContinue = result.rangeLength + 1;
+    }
+
+    if (result.keyLength == result.rangeLength)
+        result.comparison = memcmp(key, range, result.keyLength);
+    else
+        result.comparison = strcmp(key, range);
+
+    return result;
+}
+
+// These strings need to be compared according to "Extended Filtering", as in Section 3.3.2 in RFC4647.
+// https://tools.ietf.org/html/rfc4647#page-10
+//
+// The "checkFurther" field is needed in one specific situation.
+// In the quoteTable below, there are lines like:
+// { "de"   , 0x201e, 0x201c, 0x201a, 0x2018 },
+// { "de-ch", 0x00ab, 0x00bb, 0x2039, 0x203a },
+// Let's say the binary search arbitrarily decided to test our key against the upper line "de" first.
+// If the key we're testing against is "de-ch", then we should report "greater than",
+// so the binary search will keep searching and eventually find the "de-ch" line.
+// However, if the key we're testing against is "de-de", then we should report "equal to",
+// because these are the quotes we should use for all "de" except for "de-ch".
 struct QuotesForLanguage {
     const char* language;
+    UChar checkFurther;
     UChar open1;
     UChar close1;
     UChar open2;
@@ -99,166 +146,217 @@
 
 static int quoteTableLanguageComparisonFunction(const void* a, const void* b)
 {
-    return strcmp(static_cast<const QuotesForLanguage*>(a)->language,
-        static_cast<const QuotesForLanguage*>(b)->language);
+    // These strings need to be compared according to "Extended Filtering", as in Section 3.3.2 in RFC4647.
+    // https://tools.ietf.org/html/rfc4647#page-10
+    //
+    // We can exploit a few things here to improve perf:
+    // 1. The first subtag must be matched exactly
+    // 2. All the ranges have either 1 or 2 subtags
+    // 3. None of the subtags in any of the ranges are wildcards
+    //
+    // Also, see the comment just above the QuotesForLanguage struct.
+
+    auto* key = static_cast<const QuotesForLanguage*>(a);
+    auto* range = static_cast<const QuotesForLanguage*>(b);
+
+    auto firstSubtagComparison = subtagCompare(key->language, range->language);
+
+    if (firstSubtagComparison.keyLength != firstSubtagComparison.rangeLength)
+        return firstSubtagComparison.comparison;
+
+    if (firstSubtagComparison.comparison)
+        return firstSubtagComparison.comparison;
+
+    for (UChar i = 1; i <= range->checkFurther; ++i) {
+        if (!quoteTableLanguageComparisonFunction(key, range + i)) {
+            // Tell the binary search to check later in the array of ranges, to eventually find the match we just found here.
+            return 1;
+        }
+    }
+
+    for (size_t keyOffset = firstSubtagComparison.keyContinue; ;) {
+        auto nextSubtagComparison = subtagCompare(key->language + keyOffset, range->language + firstSubtagComparison.rangeContinue);
+
+        if (!nextSubtagComparison.rangeLength) {
+            // E.g. The key is "zh-Hans" and the range is "zh".
+            return 0;
+        }
+
+        if (!nextSubtagComparison.keyLength) {
+            // E.g. the key is "zh" and the range is "zh-Hant".
+            return nextSubtagComparison.comparison;
+        }
+
+        if (nextSubtagComparison.keyLength == 1) {
+            // E.g. the key is "zh-x-Hant" and the range is "zh-Hant".
+            // We want to try to find the range "zh", so tell the binary search to check earlier in the array of ranges.
+            return -1;
+        }
+
+        if (nextSubtagComparison.keyLength == nextSubtagComparison.rangeLength && !nextSubtagComparison.comparison) {
+            // E.g. the key is "de-Latn-ch" and the range is "de-ch".
+            return 0;
+        }
+
+        keyOffset += nextSubtagComparison.keyContinue;
+    }
 }
 
 static const QuotesForLanguage* quotesForLanguage(const String& language)
 {
     // Table of quotes from http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#quotes
+    // FIXME: This table is out-of-date.
     static const QuotesForLanguage quoteTable[] = {
-        { "af",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "agq",        0x201e, 0x201d, 0x201a, 0x2019 },
-        { "ak",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "am",         0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "ar",         0x201d, 0x201c, 0x2019, 0x2018 },
-        { "asa",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "az-cyrl",    0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "bas",        0x00ab, 0x00bb, 0x201e, 0x201c },
-        { "bem",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "bez",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "bg",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "bm",         0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "bn",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "br",         0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "brx",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "bs-cyrl",    0x201e, 0x201c, 0x201a, 0x2018 },
-        { "ca",         0x201c, 0x201d, 0x00ab, 0x00bb },
-        { "cgg",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "chr",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "cs",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "da",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "dav",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "de",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "de-ch",      0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "dje",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "dua",        0x00ab, 0x00bb, 0x2018, 0x2019 },
-        { "dyo",        0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "dz",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ebu",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ee",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "el",         0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "en",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "en-gb",      0x201c, 0x201d, 0x2018, 0x2019 },
-        { "es",         0x201c, 0x201d, 0x00ab, 0x00bb },
-        { "et",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "eu",         0x201c, 0x201d, 0x00ab, 0x00bb },
-        { "ewo",        0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "fa",         0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "ff",         0x201e, 0x201d, 0x201a, 0x2019 },
-        { "fi",         0x201d, 0x201d, 0x2019, 0x2019 },
-        { "fr",         0x00ab, 0x00bb, 0x00ab, 0x00bb },
-        { "fr-ca",      0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "fr-ch",      0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "gsw",        0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "gu",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "guz",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ha",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "he",         0x0022, 0x0022, 0x0027, 0x0027 },
-        { "hi",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "hr",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "hu",         0x201e, 0x201d, 0x00bb, 0x00ab },
-        { "id",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ig",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "it",         0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "ja",         0x300c, 0x300d, 0x300e, 0x300f },
-        { "jgo",        0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "jmc",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "kab",        0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "kam",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "kde",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "kea",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "khq",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ki",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "kkj",        0x00ab, 0x00bb, 0x2039, 0x203a },
-        { "kln",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "km",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "kn",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ko",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ksb",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ksf",        0x00ab, 0x00bb, 0x2018, 0x2019 },
-        { "lag",        0x201d, 0x201d, 0x2019, 0x2019 },
-        { "lg",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ln",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "lo",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "lt",         0x201e, 0x201c, 0x201e, 0x201c },
-        { "lu",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "luo",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "luy",        0x201e, 0x201c, 0x201a, 0x2018 },
-        { "lv",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mas",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mer",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mfe",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mg",         0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "mgo",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mk",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "ml",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mr",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ms",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "mua",        0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "my",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "naq",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "nb",         0x00ab, 0x00bb, 0x2018, 0x2019 },
-        { "nd",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "nl",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "nmg",        0x201e, 0x201d, 0x00ab, 0x00bb },
-        { "nn",         0x00ab, 0x00bb, 0x2018, 0x2019 },
-        { "nnh",        0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "nus",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "nyn",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "pl",         0x201e, 0x201d, 0x00ab, 0x00bb },
-        { "pt",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "pt-pt",      0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "rn",         0x201d, 0x201d, 0x2019, 0x2019 },
-        { "ro",         0x201e, 0x201d, 0x00ab, 0x00bb },
-        { "rof",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ru",         0x00ab, 0x00bb, 0x201e, 0x201c },
-        { "rw",         0x00ab, 0x00bb, 0x2018, 0x2019 },
-        { "rwk",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "saq",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "sbp",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "seh",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ses",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "sg",         0x00ab, 0x00bb, 0x201c, 0x201d },
-        { "shi",        0x00ab, 0x00bb, 0x201e, 0x201d },
-        { "shi-tfng",   0x00ab, 0x00bb, 0x201e, 0x201d },
-        { "si",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "sk",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "sl",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "sn",         0x201d, 0x201d, 0x2019, 0x2019 },
-        { "so",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "sq",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "sr",         0x201e, 0x201c, 0x201a, 0x2018 },
-        { "sr-latn",    0x201e, 0x201c, 0x201a, 0x2018 },
-        { "sv",         0x201d, 0x201d, 0x2019, 0x2019 },
-        { "sw",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "swc",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ta",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "te",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "teo",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "th",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "ti-er",      0x2018, 0x2019, 0x201c, 0x201d },
-        { "to",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "tr",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "twq",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "tzm",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "uk",         0x00ab, 0x00bb, 0x201e, 0x201c },
-        { "ur",         0x201d, 0x201c, 0x2019, 0x2018 },
-        { "vai",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "vai-latn",   0x201c, 0x201d, 0x2018, 0x2019 },
-        { "vi",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "vun",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "xh",         0x2018, 0x2019, 0x201c, 0x201d },
-        { "xog",        0x201c, 0x201d, 0x2018, 0x2019 },
-        { "yav",        0x00ab, 0x00bb, 0x00ab, 0x00bb },
-        { "yo",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "zh",         0x201c, 0x201d, 0x2018, 0x2019 },
-        { "zh-hant",    0x300c, 0x300d, 0x300e, 0x300f },
-        { "zu",         0x201c, 0x201d, 0x2018, 0x2019 },
+        { "af",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "agq",        0, 0x201e, 0x201d, 0x201a, 0x2019 },
+        { "ak",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "am",         0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "ar",         0, 0x201d, 0x201c, 0x2019, 0x2018 },
+        { "asa",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "az-cyrl",    0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "bas",        0, 0x00ab, 0x00bb, 0x201e, 0x201c },
+        { "bem",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "bez",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "bg",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "bm",         0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "bn",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "br",         0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "brx",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "bs-cyrl",    0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "ca",         0, 0x201c, 0x201d, 0x00ab, 0x00bb },
+        { "cgg",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "chr",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "cs",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "da",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "dav",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "de",         1, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "de-ch",      0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "dje",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "dua",        0, 0x00ab, 0x00bb, 0x2018, 0x2019 },
+        { "dyo",        0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "dz",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ebu",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ee",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "el",         0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "en",         1, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "en-gb",      0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "es",         0, 0x201c, 0x201d, 0x00ab, 0x00bb },
+        { "et",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "eu",         0, 0x201c, 0x201d, 0x00ab, 0x00bb },
+        { "ewo",        0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "fa",         0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "ff",         0, 0x201e, 0x201d, 0x201a, 0x2019 },
+        { "fi",         0, 0x201d, 0x201d, 0x2019, 0x2019 },
+        { "fr",         2, 0x00ab, 0x00bb, 0x00ab, 0x00bb },
+        { "fr-ca",      0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "fr-ch",      0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "gsw",        0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "gu",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "guz",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ha",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "he",         0, 0x0022, 0x0022, 0x0027, 0x0027 },
+        { "hi",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "hr",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "hu",         0, 0x201e, 0x201d, 0x00bb, 0x00ab },
+        { "id",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ig",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "it",         0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "ja",         0, 0x300c, 0x300d, 0x300e, 0x300f },
+        { "jgo",        0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "jmc",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "kab",        0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "kam",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "kde",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "kea",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "khq",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ki",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "kkj",        0, 0x00ab, 0x00bb, 0x2039, 0x203a },
+        { "kln",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "km",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "kn",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ko",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ksb",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ksf",        0, 0x00ab, 0x00bb, 0x2018, 0x2019 },
+        { "lag",        0, 0x201d, 0x201d, 0x2019, 0x2019 },
+        { "lg",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ln",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "lo",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "lt",         0, 0x201e, 0x201c, 0x201e, 0x201c },
+        { "lu",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "luo",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "luy",        0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "lv",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mas",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mer",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mfe",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mg",         0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "mgo",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mk",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "ml",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mr",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ms",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "mua",        0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "my",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "naq",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "nb",         0, 0x00ab, 0x00bb, 0x2018, 0x2019 },
+        { "nd",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "nl",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "nmg",        0, 0x201e, 0x201d, 0x00ab, 0x00bb },
+        { "nn",         0, 0x00ab, 0x00bb, 0x2018, 0x2019 },
+        { "nnh",        0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "nus",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "nyn",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "pl",         0, 0x201e, 0x201d, 0x00ab, 0x00bb },
+        { "pt",         1, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "pt-pt",      0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "rn",         0, 0x201d, 0x201d, 0x2019, 0x2019 },
+        { "ro",         0, 0x201e, 0x201d, 0x00ab, 0x00bb },
+        { "rof",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ru",         0, 0x00ab, 0x00bb, 0x201e, 0x201c },
+        { "rw",         0, 0x00ab, 0x00bb, 0x2018, 0x2019 },
+        { "rwk",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "saq",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "sbp",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "seh",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ses",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "sg",         0, 0x00ab, 0x00bb, 0x201c, 0x201d },
+        { "shi",        1, 0x00ab, 0x00bb, 0x201e, 0x201d },
+        { "shi-tfng",   0, 0x00ab, 0x00bb, 0x201e, 0x201d },
+        { "si",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "sk",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "sl",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "sn",         0, 0x201d, 0x201d, 0x2019, 0x2019 },
+        { "so",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "sq",         0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "sr",         1, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "sr-latn",    0, 0x201e, 0x201c, 0x201a, 0x2018 },
+        { "sv",         0, 0x201d, 0x201d, 0x2019, 0x2019 },
+        { "sw",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "swc",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ta",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "te",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "teo",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "th",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "ti-er",      0, 0x2018, 0x2019, 0x201c, 0x201d },
+        { "to",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "tr",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "twq",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "tzm",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "uk",         0, 0x00ab, 0x00bb, 0x201e, 0x201c },
+        { "ur",         0, 0x201d, 0x201c, 0x2019, 0x2018 },
+        { "vai",        1, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "vai-latn",   0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "vi",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "vun",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "xh",         0, 0x2018, 0x2019, 0x201c, 0x201d },
+        { "xog",        0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "yav",        0, 0x00ab, 0x00bb, 0x00ab, 0x00bb },
+        { "yo",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "zh",         1, 0x201c, 0x201d, 0x2018, 0x2019 },
+        { "zh-hant",    0, 0x300c, 0x300d, 0x300e, 0x300f },
+        { "zu",         0, 0x201c, 0x201d, 0x2018, 0x2019 },
     };
 
-    const unsigned maxLanguageLength = 8;
-
 #if ASSERT_ENABLED
     // One time check that the table meets the constraints that the code below relies on.
 
@@ -270,8 +368,6 @@
         checkNumberOfDistinctQuoteCharacters(apostrophe);
 
         for (unsigned i = 0; i < WTF_ARRAY_LENGTH(quoteTable); ++i) {
-            ASSERT(strlen(quoteTable[i].language) <= maxLanguageLength);
-
             if (i)
                 ASSERT(strcmp(quoteTable[i - 1].language, quoteTable[i].language) < 0);
 
@@ -287,19 +383,19 @@
 #endif // ASSERT_ENABLED
 
     unsigned length = language.length();
-    if (!length || length > maxLanguageLength)
-        return 0;
+    if (!length)
+        return nullptr;
 
-    char languageKeyBuffer[maxLanguageLength + 1];
+    Vector<char> languageKeyBuffer(length + 1);
     for (unsigned i = 0; i < length; ++i) {
         UChar character = toASCIILower(language[i]);
         if (!(isASCIILower(character) || character == '-'))
-            return 0;
+            return nullptr;
         languageKeyBuffer[i] = static_cast<char>(character);
     }
     languageKeyBuffer[length] = 0;
 
-    QuotesForLanguage languageKey = { languageKeyBuffer, 0, 0, 0, 0 };
+    QuotesForLanguage languageKey = { languageKeyBuffer.data(), 0, 0, 0, 0, 0 };
 
     return static_cast<const QuotesForLanguage*>(bsearch(&languageKey,
         quoteTable, WTF_ARRAY_LENGTH(quoteTable), sizeof(quoteTable[0]), quoteTableLanguageComparisonFunction));
@@ -378,11 +474,12 @@
         isOpenQuote = true;
         FALLTHROUGH;
     case QuoteType::CloseQuote:
-        if (const QuotesData* quotes = style().quotes())
+        if (const auto* quotes = style().quotes())
             return isOpenQuote ? quotes->openQuote(m_depth).impl() : quotes->closeQuote(m_depth).impl();
-        if (const QuotesForLanguage* quotes = quotesForLanguage(style().specifiedLocale()))
+        if (const auto* quotes = quotesForLanguage(style().computedLocale()))
             return stringForQuoteCharacter(isOpenQuote ? (m_depth ? quotes->open2 : quotes->open1) : (m_depth ? quotes->close2 : quotes->close1));
         // FIXME: Should the default be the quotes for "en" rather than straight quotes?
+        // (According to https://html.spec.whatwg.org/multipage/rendering.html#quotes, the answer is "yes".)
         return m_depth ? apostropheString() : quotationMarkString();
     }
     ASSERT_NOT_REACHED();

_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
https://lists.webkit.org/mailman/listinfo/webkit-changes

Reply via email to