Handle QCollator with locale C by delegating to QString

Previously, the C locale was treated as English because each back-end
takes the locale's bcp47Name(), which maps C to en. However, the C
locale has its own rules; which QString helpfully implements; so we
can delegate to it in this case. Extended this to sort keys, where
possible. Clean up existing implementations in the process.

Extended tst_QCollator::compare() with some cases to check this. That
required wrapping the test's calls to collator.compare() in a sign
canonicalizer, since it can return any -ve for < or +ve for >, not
just -1 and +1 for these cases (and it'd be rash to hard-code specific
negative and positive values, as they may vary between backends).

[ChangeLog][QtCore][QCollator] Added support for collation in the C
locale, albeit this is only well-defined for ASCII. Collation sort
keys remain unsupported on Darwin.

Fixes: QTBUG-58621
Change-Id: I327010d90f09bd1b1816f5590cb124e3d423e61d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

6 files changed, 57 insertions, 19 deletions

diff --git a/src/corelib/tools/qcollator.cpp b/src/corelib/tools/qcollator.cpp
index 1cf223aae6..4315d35a52 100644
--- a/src/corelib/tools/qcollator.cpp
+++ b/src/corelib/tools/qcollator.cpp
@@ -79,7 +79,6 @@ QT_BEGIN_NAMESPACE
 QCollator::QCollator(const QLocale &locale)
     : d(new QCollatorPrivate(locale))
 {
-    d->init();
 }
 
 /*!
@@ -323,6 +322,8 @@ bool QCollator::ignorePunctuation() const
     methods directly. But if the string is compared repeatedly (e.g. when sorting
     a whole list of strings), it's usually faster to create the sort keys for each
     string and then sort using the keys.
+
+    \note Not supported with the C (a.k.a. POSIX) locale on Darwin.
  */
 
 /*!
diff --git a/src/corelib/tools/qcollator_icu.cpp b/src/corelib/tools/qcollator_icu.cpp
index fd621983d3..ab45b9a1a1 100644
--- a/src/corelib/tools/qcollator_icu.cpp
+++ b/src/corelib/tools/qcollator_icu.cpp
@@ -55,6 +55,8 @@ QT_BEGIN_NAMESPACE
 void QCollatorPrivate::init()
 {
     cleanup();
+    if (isC())
+        return;
 
     UErrorCode status = U_ZERO_ERROR;
     QByteArray name = QLocalePrivate::get(locale)->bcp47Name('_');
@@ -140,6 +142,8 @@ QCollatorSortKey QCollator::sortKey(const QString &string) const
 {
     if (d->dirty)
         d->init();
+    if (d->isC())
+        return QCollatorSortKey(new QCollatorSortKeyPrivate(string.toUtf8()));
 
     if (d->collator) {
         QByteArray result(16 + string.size() + (string.size() >> 2), Qt::Uninitialized);
diff --git a/src/corelib/tools/qcollator_macx.cpp b/src/corelib/tools/qcollator_macx.cpp
index 9aa59a81dc..42e67e0c12 100644
--- a/src/corelib/tools/qcollator_macx.cpp
+++ b/src/corelib/tools/qcollator_macx.cpp
@@ -55,6 +55,15 @@ QT_BEGIN_NAMESPACE
 void QCollatorPrivate::init()
 {
     cleanup();
+    /*
+      LocaleRefFromLocaleString() will accept "POSIX" as the locale name, but
+      the locale it produces (named "pos") doesn't implement the [A-Z] < [a-z]
+      behavior we expect of the C locale.  We can use QStringView to get round
+      that for collation, but this leaves no way to do a sort key.
+    */
+    if (isC())
+        return;
+
     LocaleRef localeRef;
     int rc = LocaleRefFromLocaleString(QLocalePrivate::get(locale)->bcp47Name().constData(), &localeRef);
     if (rc != 0)
@@ -92,6 +101,8 @@ int QCollator::compare(const QChar *s1, int len1, const QChar *s2, int len2) con
 {
     if (d->dirty)
         d->init();
+    if (!d->collator)
+        return QStringView(s1, len1).compare(QStringView(s2, len2), caseSensitivity());
 
     SInt32 result;
     Boolean equivalent;
@@ -104,6 +115,7 @@ int QCollator::compare(const QChar *s1, int len1, const QChar *s2, int len2) con
         return 0;
     return result < 0 ? -1 : 1;
 }
+
 int QCollator::compare(const QString &str1, const QString &str2) const
 {
     return compare(str1.constData(), str1.size(), str2.constData(), str2.size());
@@ -118,6 +130,11 @@ QCollatorSortKey QCollator::sortKey(const QString &string) const
 {
     if (d->dirty)
         d->init();
+    if (!d->collator) {
+        // What should (or even *can*) we do here ? (See init()'s comment.)
+        qWarning("QCollator doesn't support sort keys for the C locale on Darwin");
+        return QCollatorSortKey(nullptr);
+    }
 
     //Documentation recommends having it 5 times as big as the input
     QVector<UCCollationValue> ret(string.size() * 5);
@@ -136,6 +153,9 @@ QCollatorSortKey QCollator::sortKey(const QString &string) const
 
 int QCollatorSortKey::compare(const QCollatorSortKey &key) const
 {
+    if (!d.data())
+        return 0;
+
     SInt32 order;
     UCCompareCollationKeys(d->m_key.data(), d->m_key.size(),
                            key.d->m_key.data(), key.d->m_key.size(),
diff --git a/src/corelib/tools/qcollator_p.h b/src/corelib/tools/qcollator_p.h
index e89c08447c..361c3fb987 100644
--- a/src/corelib/tools/qcollator_p.h
+++ b/src/corelib/tools/qcollator_p.h
@@ -110,6 +110,7 @@ public:
 
     QCollatorPrivate(const QLocale &locale) : locale(locale) {}
     ~QCollatorPrivate() { cleanup(); }
+    bool isC() { return locale.language() == QLocale::C; }
 
     void clear() {
         cleanup();
diff --git a/src/corelib/tools/qcollator_posix.cpp b/src/corelib/tools/qcollator_posix.cpp
index 42413a4a82..81f97a02e1 100644
--- a/src/corelib/tools/qcollator_posix.cpp
+++ b/src/corelib/tools/qcollator_posix.cpp
@@ -48,10 +48,12 @@ QT_BEGIN_NAMESPACE
 
 void QCollatorPrivate::init()
 {
-    if (locale != QLocale())
-        qWarning("Only default locale supported with the posix collation implementation");
-    if (caseSensitivity != Qt::CaseSensitive)
-        qWarning("Case insensitive sorting unsupported in the posix collation implementation");
+    if (!isC()) {
+        if (locale != QLocale())
+            qWarning("Only C and default locale supported with the posix collation implementation");
+        if (caseSensitivity != Qt::CaseSensitive)
+            qWarning("Case insensitive sorting unsupported in the posix collation implementation");
+    }
     if (numericMode)
         qWarning("Numeric mode unsupported in the posix collation implementation");
     if (ignorePunctuation)
@@ -73,14 +75,16 @@ static void stringToWCharArray(QVarLengthArray<wchar_t> &ret, const QString &str
 
 int QCollator::compare(const QChar *s1, int len1, const QChar *s2, int len2) const
 {
-    QVarLengthArray<wchar_t> array1, array2;
-    stringToWCharArray(array1, QString(s1, len1));
-    stringToWCharArray(array2, QString(s2, len2));
-    return std::wcscoll(array1.constData(), array2.constData());
+    return compare(QString::fromRawData(s1, len1), QString::fromRawData(s2, len2));
 }
 
 int QCollator::compare(const QString &s1, const QString &s2) const
 {
+    if (d->isC())
+        return s1.compare(s2, caseSensitivity());
+    if (d->dirty)
+        d->init();
+
     QVarLengthArray<wchar_t> array1, array2;
     stringToWCharArray(array1, s1);
     stringToWCharArray(array2, s2);
@@ -89,10 +93,7 @@ int QCollator::compare(const QString &s1, const QString &s2) const
 
 int QCollator::compare(const QStringRef &s1, const QStringRef &s2) const
 {
-    if (d->dirty)
-        d->init();
-
-    return compare(s1.constData(), s1.size(), s2.constData(), s2.size());
+    return compare(s1.toString(), s2.toString());
 }
 
 QCollatorSortKey QCollator::sortKey(const QString &string) const
@@ -102,14 +103,18 @@ QCollatorSortKey QCollator::sortKey(const QString &string) const
 
     QVarLengthArray<wchar_t> original;
     stringToWCharArray(original, string);
-    QVector<wchar_t> result(string.size());
-    size_t size = std::wcsxfrm(result.data(), original.constData(), string.size());
-    if (size > uint(result.size())) {
+    QVector<wchar_t> result(original.size());
+    if (d->isC()) {
+        std::copy(original.cbegin(), original.cend(), result.begin());
+    } else {
+        size_t size = std::wcsxfrm(result.data(), original.constData(), string.size());
+        if (size > uint(result.size())) {
+            result.resize(size+1);
+            size = std::wcsxfrm(result.data(), original.constData(), string.size());
+        }
         result.resize(size+1);
-        size = std::wcsxfrm(result.data(), original.constData(), string.size());
+        result[size] = 0;
     }
-    result.resize(size+1);
-    result[size] = 0;
     return QCollatorSortKey(new QCollatorSortKeyPrivate(std::move(result)));
 }
 
diff --git a/src/corelib/tools/qcollator_win.cpp b/src/corelib/tools/qcollator_win.cpp
index 5a838c1b50..35142bb8b8 100644
--- a/src/corelib/tools/qcollator_win.cpp
+++ b/src/corelib/tools/qcollator_win.cpp
@@ -60,6 +60,8 @@ extern LCID qt_inIsoNametoLCID(const char *name);
 void QCollatorPrivate::init()
 {
     collator = 0;
+    if (isC())
+        return;
 
 #ifndef USE_COMPARESTRINGEX
     localeID = qt_inIsoNametoLCID(QLocalePrivate::get(locale)->bcp47Name().constData());
@@ -86,6 +88,9 @@ void QCollatorPrivate::cleanup()
 
 int QCollator::compare(const QChar *s1, int len1, const QChar *s2, int len2) const
 {
+    if (d->isC())
+        return QString::compare_helper(s1, len1, s2, len2, d->caseSensitivity);
+
     if (d->dirty)
         d->init();
 
@@ -119,6 +124,8 @@ QCollatorSortKey QCollator::sortKey(const QString &string) const
 {
     if (d->dirty)
         d->init();
+    if (d->isC())
+        return QCollatorSortKey(new QCollatorSortKeyPrivate(string));
 
 #ifndef USE_COMPARESTRINGEX
     int size = LCMapStringW(d->localeID, LCMAP_SORTKEY | d->collator,