Remove UTF16LE means swap compatibility hack

ufleisch · Dec 20, 2023 · 73ddb42 · 73ddb42
1 parent 1ee7493
commit 73ddb42
Show file tree

Hide file tree

Showing 4 changed files with 41 additions and 37 deletions.
diff --git a/taglib/toolkit/tstring.cpp b/taglib/toolkit/tstring.cpp
@@ -38,7 +38,7 @@ namespace
  using namespace TagLib;
 
  // Returns the native format of std::wstring.
- String::Type wcharByteOrder()
+ constexpr String::Type wcharByteOrder()
  {
  if(Utils::systemByteOrder() == Utils::LittleEndian)
  return String::UTF16LE;
@@ -172,35 +172,31 @@ String::String(const std::string &s, Type t) :
  }
 }
 
+String::String(const wstring &s) :
+ String(s, wcharByteOrder())
+{
+}
+
 String::String(const wstring &s, Type t) :
  d(std::make_shared<StringPrivate>())
 {
  if(t == UTF16 || t == UTF16BE || t == UTF16LE) {
- // This looks ugly but needed for the compatibility with TagLib1.8.
- // Should be removed in TabLib2.0.
- if (t == UTF16BE)
- t = wcharByteOrder();
- else if (t == UTF16LE)
- t = (wcharByteOrder() == UTF16LE ? UTF16BE : UTF16LE);
-
  copyFromUTF16(d->data, s.c_str(), s.length(), t);
  }
  else {
  debug("String::String() -- TagLib::wstring should not contain Latin1 or UTF-8.");
  }
 }
 
+String::String(const wchar_t *s) :
+ String(s, wcharByteOrder())
+{
+}
+
 String::String(const wchar_t *s, Type t) :
  d(std::make_shared<StringPrivate>())
 {
  if(t == UTF16 || t == UTF16BE || t == UTF16LE) {
- // This looks ugly but needed for the compatibility with TagLib1.8.
- // Should be removed in TabLib2.0.
- if (t == UTF16BE)
- t = wcharByteOrder();
- else if (t == UTF16LE)
- t = (wcharByteOrder() == UTF16LE ? UTF16BE : UTF16LE);
-
  copyFromUTF16(d->data, s, ::wcslen(s), t);
  }
  else {

diff --git a/taglib/toolkit/tstring.h b/taglib/toolkit/tstring.h
@@ -139,22 +139,24 @@ namespace TagLib {
  String(const std::string &s, Type t = Latin1);
 
  /*!
- * Makes a deep copy of the data in \a s.
- *
- * /note If \a t is UTF16LE, the byte order of \a s will be swapped regardless
- * of the CPU byte order. If UTF16BE, it will not be swapped. This behavior
- * will be changed in TagLib2.0.
+ * Makes a deep copy of the data in \a s, which are in CPU byte order.
  */
- String(const wstring &s, Type t = UTF16BE);
+ String(const wstring &s);
 
  /*!
- * Makes a deep copy of the data in \a s.
- *
- * /note If \a t is UTF16LE, the byte order of \a s will be swapped regardless
- * of the CPU byte order. If UTF16BE, it will not be swapped. This behavior
- * will be changed in TagLib2.0.
+ * Makes a deep copy of the data in \a s, which are in byte order \a t.
+ */
+ String(const wstring &s, Type t);
+
+ /*!
+ * Makes a deep copy of the data in \a s, which are in CPU byte order.
+ */
+ String(const wchar_t *s);
+
+ /*!
+ * Makes a deep copy of the data in \a s, which are in byte order \a t.
  */
- String(const wchar_t *s, Type t = UTF16BE);
+ String(const wchar_t *s, Type t);
 
  /*!
  * Makes a deep copy of the data in \a c.

diff --git a/taglib/toolkit/tutils.h b/taglib/toolkit/tutils.h
@@ -203,11 +203,12 @@ namespace TagLib
  /*!
  * Returns the byte order of the system.
  */
- inline ByteOrder systemByteOrder()
+ inline constexpr ByteOrder systemByteOrder()
  {
- union {
+ union IntCharUnion {
  int i;
  char c;
+ constexpr IntCharUnion() : i{} {}
  } u;
 
  u.i = 1;

diff --git a/tests/test_string.cpp b/tests/test_string.cpp
@@ -26,6 +26,7 @@
 #include <cstring>
 
 #include "tstring.h"
+#include "tutils.h"
 #include <cppunit/extensions/HelperMacros.h>
 
 using namespace std;
@@ -96,19 +97,23 @@ class TestString : public CppUnit::TestFixture
  String unicode3(L"\u65E5\u672C\u8A9E");
  CPPUNIT_ASSERT(*(unicode3.toCWString() + 1) == L'\u672C');
 
- String unicode4(L"\u65e5\u672c\u8a9e", String::UTF16BE);
- CPPUNIT_ASSERT(unicode4[1] == L'\u672c');
+ const wchar_t wcSystemOrder[] = {L'\u65E5', L'\u672C', L'\u8A9E', 0};
+ const wchar_t wcSwappedOrder[] = {L'\uE565', L'\u2C67', L'\u9E8A', 0};
+ const std::wstring wsSystemOrder = L"\u65e5\u672c\u8a9e";
+ const std::wstring wsSwappedOrder = L"\ue565\u2c67\u9e8a";
+ const bool isLe = Utils::systemByteOrder() == Utils::LittleEndian;
 
- String unicode5(L"\u65e5\u672c\u8a9e", String::UTF16LE);
- CPPUNIT_ASSERT(unicode5[1] == L'\u2c67');
+ String unicode4(isLe ? wcSwappedOrder : wcSystemOrder, String::UTF16BE);
+ CPPUNIT_ASSERT(unicode4[1] == L'\u672c');
 
- std::wstring stduni = L"\u65e5\u672c\u8a9e";
+ String unicode5(isLe ? wcSystemOrder : wcSwappedOrder, String::UTF16LE);
+ CPPUNIT_ASSERT(unicode5[1] == L'\u672c');
 
- String unicode6(stduni, String::UTF16BE);
+ String unicode6(isLe ? wsSwappedOrder : wsSystemOrder, String::UTF16BE);
  CPPUNIT_ASSERT(unicode6[1] == L'\u672c');
 
- String unicode7(stduni, String::UTF16LE);
- CPPUNIT_ASSERT(unicode7[1] == L'\u2c67');
+ String unicode7(isLe ? wsSystemOrder : wsSwappedOrder, String::UTF16LE);
+ CPPUNIT_ASSERT(unicode7[1] == L'\u672c');
 
  CPPUNIT_ASSERT(String(" foo ").stripWhiteSpace() == String("foo"));
  CPPUNIT_ASSERT(String("foo ").stripWhiteSpace() == String("foo"));