QByteArray: inline QByteArray::indexOf() and use a char overload

QByteArray::indexOf() had the best implementation of the two *because*
it was out-of-line, in qbytearray.cpp. The compiler could thus see the
body of the QtPrivate::findByteArray() function and inline it, with
constant-propagation, so only findCharHelper() was expanded and we had a
fast path towards memchr().

On the other hand, QByteArrayView::indexOf() was inline, so the
compiler emitted the call to QtPrivate::findByteArray() in user code,
causing the full function to be executed. We fix that by adding a char
overload for QtPrivate::findByteArray() and lastIndexOf(), so we get to
memchr() more quickly.

Also, inline QByteArray::indexOf() and lastIndexOf().

Before:
QByteArray::indexOf():
     8.83287376 nsecs per iteration
     22.01766832 CPU cycles per iteration, 2.49 GHz
     62.00000330 instructions per iteration, 2.816 instr/cycle
     21.00000281 branch instructions per iteration, 2.38 G/sec
QByteArrayView::indexOf():
     9.64034694 nsecs per iteration
     24.03001151 CPU cycles per iteration, 2.49 GHz
     68.00000355 instructions per iteration, 2.830 instr/cycle
     23.00000306 branch instructions per iteration, 2.39 G/sec

After (same result for both, requires recompilation):
     8.83207052 nsecs per iteration
     22.01568546 CPU cycles per iteration, 2.49 GHz
     60.00000331 instructions per iteration, 2.725 instr/cycle
     21.00000281 branch instructions per iteration, 2.38 G/sec

The inlining appears to have cut 2 instructions, but the effect is not
measurable in this benchmark (the entire code is in cache and the Branch
Predictor is probably primed).

Fixes: QTBUG-119750
Change-Id: Ica7a43f6147b49c187ccfffd179e1cb4b306fc62
Reviewed-by: Marc Mutz <marc.mutz@qt.io>

src/corelib/compat/removed_api.cpp

@@ -625,7 +625,7 @@ QBitArray QBitArray::operator~() const
     return QBitArray(*this).inverted_inplace();
 }
 
-#include "qbytearray.h"
+#include "qbytearray.h"     // also includes inlined API
 
 QByteArray QByteArray::left(qsizetype len)  const
 {

src/corelib/text/qbytearray.cpp

@@ -2656,6 +2656,11 @@ static inline qsizetype findCharHelper(QByteArrayView haystack, qsizetype from,
     return -1;
 }
 
+qsizetype QtPrivate::findByteArray(QByteArrayView haystack, qsizetype from, char needle) noexcept
+{
+    return findCharHelper(haystack, from, needle);
+}
+
 qsizetype QtPrivate::findByteArray(QByteArrayView haystack, qsizetype from, QByteArrayView needle) noexcept
 {
     const auto ol = needle.size();
@@ -2690,6 +2695,7 @@ qsizetype QtPrivate::findByteArray(QByteArrayView haystack, qsizetype from, QByt
 */
 
 /*!
+    \fn qsizetype QByteArray::indexOf(char ch, qsizetype from) const
     \overload
 
     Returns the index position of the start of the first occurrence of the
@@ -2702,11 +2708,6 @@ qsizetype QtPrivate::findByteArray(QByteArrayView haystack, qsizetype from, QByt
     \sa lastIndexOf(), contains()
 */
 
-qsizetype QByteArray::indexOf(char ch, qsizetype from) const
-{
-    return qToByteArrayViewIgnoringNull(*this).indexOf(ch, from);
-}
-
 static qsizetype lastIndexOfHelper(const char *haystack, qsizetype l, const char *needle,
                                    qsizetype ol, qsizetype from)
 {
@@ -2760,6 +2761,11 @@ static inline qsizetype lastIndexOfCharHelper(QByteArrayView haystack, qsizetype
     return -1;
 }
 
+qsizetype QtPrivate::lastIndexOf(QByteArrayView haystack, qsizetype from, char needle) noexcept
+{
+    return lastIndexOfCharHelper(haystack, from, needle);
+}
+
 qsizetype QtPrivate::lastIndexOf(QByteArrayView haystack, qsizetype from, QByteArrayView needle) noexcept
 {
     if (haystack.isEmpty()) {
@@ -2813,6 +2819,7 @@ qsizetype QtPrivate::lastIndexOf(QByteArrayView haystack, qsizetype from, QByteA
 */
 
 /*!
+    \fn qsizetype QByteArray::lastIndexOf(char ch, qsizetype from) const
     \overload
 
     Returns the index position of the start of the last occurrence of byte \a ch
@@ -2826,11 +2833,6 @@ qsizetype QtPrivate::lastIndexOf(QByteArrayView haystack, qsizetype from, QByteA
     \sa indexOf(), contains()
 */
 
-qsizetype QByteArray::lastIndexOf(char ch, qsizetype from) const
-{
-    return qToByteArrayViewIgnoringNull(*this).lastIndexOf(ch, from);
-}
-
 static inline qsizetype countCharHelper(QByteArrayView haystack, char needle) noexcept
 {
     qsizetype num = 0;

src/corelib/text/qbytearray.h

@@ -133,10 +133,12 @@ public:
     [[nodiscard]] char back() const { return at(size() - 1); }
     [[nodiscard]] inline char &back();
 
+    QT_CORE_INLINE_SINCE(6, 7)
     qsizetype indexOf(char c, qsizetype from = 0) const;
     qsizetype indexOf(QByteArrayView bv, qsizetype from = 0) const
     { return QtPrivate::findByteArray(qToByteArrayViewIgnoringNull(*this), from, bv); }
 
+    QT_CORE_INLINE_SINCE(6, 7)
     qsizetype lastIndexOf(char c, qsizetype from = -1) const;
     qsizetype lastIndexOf(QByteArrayView bv) const
     { return lastIndexOf(bv, size()); }
@@ -688,6 +690,16 @@ bool QByteArray::isNull() const noexcept
     return d->isNull();
 }
 #endif
+#if QT_CORE_INLINE_IMPL_SINCE(6, 7)
+qsizetype QByteArray::indexOf(char ch, qsizetype from) const
+{
+    return qToByteArrayViewIgnoringNull(*this).indexOf(ch, from);
+}
+qsizetype QByteArray::lastIndexOf(char ch, qsizetype from) const
+{
+    return qToByteArrayViewIgnoringNull(*this).lastIndexOf(ch, from);
+}
+#endif
 
 #if !defined(QT_NO_DATASTREAM) || defined(QT_BOOTSTRAPPED)
 Q_CORE_EXPORT QDataStream &operator<<(QDataStream &, const QByteArray &);

src/corelib/text/qbytearrayalgorithms.h

@@ -25,9 +25,15 @@ bool startsWith(QByteArrayView haystack, QByteArrayView needle) noexcept;
 [[nodiscard]] Q_CORE_EXPORT Q_DECL_PURE_FUNCTION
 bool endsWith(QByteArrayView haystack, QByteArrayView needle) noexcept;
 
+[[nodiscard]] Q_CORE_EXPORT Q_DECL_PURE_FUNCTION
+qsizetype findByteArray(QByteArrayView haystack, qsizetype from, char needle) noexcept;
+
 [[nodiscard]] Q_CORE_EXPORT Q_DECL_PURE_FUNCTION
 qsizetype findByteArray(QByteArrayView haystack, qsizetype from, QByteArrayView needle) noexcept;
 
+[[nodiscard]] Q_CORE_EXPORT Q_DECL_PURE_FUNCTION
+qsizetype lastIndexOf(QByteArrayView haystack, qsizetype from, char needle) noexcept;
+
 [[nodiscard]] Q_CORE_EXPORT Q_DECL_PURE_FUNCTION
 qsizetype lastIndexOf(QByteArrayView haystack, qsizetype from, QByteArrayView needle) noexcept;
 

src/corelib/text/qbytearrayview.h

@@ -263,7 +263,7 @@ public:
     [[nodiscard]] qsizetype indexOf(QByteArrayView a, qsizetype from = 0) const noexcept
     { return QtPrivate::findByteArray(*this, from, a); }
     [[nodiscard]] qsizetype indexOf(char ch, qsizetype from = 0) const noexcept
-    { return QtPrivate::findByteArray(*this, from, QByteArrayView(&ch, 1)); }
+    { return QtPrivate::findByteArray(*this, from, ch); }
 
     [[nodiscard]] bool contains(QByteArrayView a) const noexcept
     { return indexOf(a) != qsizetype(-1); }
@@ -275,7 +275,7 @@ public:
     [[nodiscard]] qsizetype lastIndexOf(QByteArrayView a, qsizetype from) const noexcept
     { return QtPrivate::lastIndexOf(*this, from, a); }
     [[nodiscard]] qsizetype lastIndexOf(char ch, qsizetype from = -1) const noexcept
-    { return QtPrivate::lastIndexOf(*this, from, QByteArrayView(&ch, 1)); }
+    { return QtPrivate::lastIndexOf(*this, from, ch); }
 
     [[nodiscard]] qsizetype count(QByteArrayView a) const noexcept
     { return QtPrivate::count(*this, a); }