Support Cyrillic's equivalent of 'E' as exponent separator

Only Ukrainian is actually recorded in CLDR as using U+0415 as
exponent separator; all other Cyrillic-using locales officially use
plain ASCII 'E'. However, it seems reasonable, in all Cyrillic
locales, to recognize both (given that they look very similar).

Task-number: QTBUG-107801
Change-Id: I70a1e60a2d9fe7e254e01d32c5bad909ea4b8c76
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>

src/corelib/text/qlocale.cpp

@@ -3839,8 +3839,11 @@ inline QLocaleData::NumericData QLocaleData::numericData(QLocaleData::NumberMode
     result.plus = plus().viewData(single_character_data);
     if (mode != IntegerMode)
         result.decimal = decimalSeparator().viewData(single_character_data);
-    if (mode == DoubleScientificMode)
+    if (mode == DoubleScientificMode) {
         result.exponent = exponential().viewData(single_character_data);
+        // exponentCyrillic means "apply the Cyrrilic-specific exponent hack"
+        result.exponentCyrillic = m_script_id == QLocale::CyrillicScript;
+    }
 #ifndef QT_NO_SYSTEMLOCALE
     if (this == &systemLocaleData) {
         const auto getString = [sys = systemLocale()](QSystemLocale::QueryType query) {
@@ -4017,6 +4020,15 @@ char NumericTokenizer::nextToken()
         return ',';
     }
 
+    // Cyrillic has its own E, used by Ukrainian as exponent; but others
+    // writing Cyrillic may well use that; and Ukrainians might well use E.
+    // All other Cyrillic locales (officially) use plain ASCII E.
+    if (m_guide.exponentCyrillic // Only true in scientific float mode.
+        && (tail.startsWith(u"\u0415") || tail.startsWith(u"E"))) {
+        ++m_index;
+        return 'e';
+    }
+
     return 0;
 }
 } // namespace with no name

src/corelib/text/qlocale_p.h

@@ -322,6 +322,7 @@ public:
         char32_t zeroUcs = 0;
         qint8 zeroLen = 0;
         bool isC = false; // C locale sets this and nothing else.
+        bool exponentCyrillic = false; // True only for floating-point parsing of Cyrillic.
         void setZero(QStringView zero)
         {
             // No known locale has digits that are more than one Unicode
@@ -346,6 +347,8 @@ public:
         {
             if (isC)
                 return true;
+            if (exponentCyrillic && exponent != u"E" && exponent != u"\u0415")
+                return false;
             return (zeroLen == 1 || zeroLen == 2) && zeroUcs > 0
                 && (mode == IntegerMode || !decimal.isEmpty())
                 // group may be empty (user config in system locale)

tests/auto/corelib/text/qlocale/tst_qlocale.cpp

@@ -910,6 +910,14 @@ void tst_QLocale::toReal_data()
         << u"fa_IR"_s << u"\u06f4\u00d7\u200e\u2212\u06f0\u06f3"_s << false << 0.0;
     QTest::newRow("fa_IR 4x!3") // Only first character of exponent and sign
         << u"fa_IR"_s << u"\u06f4\u00d7\u200e\u06f0\u06f3"_s << false << 0.0;
+
+    // Cyrillic has its own E; only officially used by Ukrainian as exponent,
+    // with other Cyrillic locales using the Latin E. QLocale allows that there
+    // may be some cross-over between these.
+    QTest::newRow("uk_UA Cyrillic E") << u"uk_UA"_s << u"4\u0415-3"_s << true << 4e-3; // Official
+    QTest::newRow("uk_UA Latin E") << u"uk_UA"_s << u"4E-3"_s << true << 4e-3;
+    QTest::newRow("ru_RU Latin E") << u"ru_RU"_s << u"4E-3"_s << true << 4e-3; // Official
+    QTest::newRow("ru_RU Cyrillic E") << u"ru_RU"_s << u"4\u0415-3"_s << true << 4e-3;
 }
 
 void tst_QLocale::stringToDouble_data()