unicoder.cpp light performance improvements (#1586)

* Update unicoder.cpp

CheckForInvalidUtf8 performance improvement

* unicoder.cpp

light performance improvements

* Update unicoder.cpp

space-tab change

diff --git a/Src/Common/unicoder.cpp b/Src/Common/unicoder.cpp
index 37f2381..e0672e5 100644 (file)
--- a/Src/Common/unicoder.cpp
+++ b/Src/Common/unicoder.cpp
@@ -269,10 +269,9 @@ int to_utf8_advance(unsigned u, unsigned char * &lpd)
  */
 void maketchar(String & ch, unsigned unich, bool & lossy)
 {
-       static unsigned codepage = CP_ACP;
        // NB: Windows always draws in CP_ACP, not CP_THREAD_ACP, so we must use CP_ACP as an internal codepage
 
-       maketchar(ch, unich, lossy, codepage);
+       maketchar(ch, unich, lossy, CP_ACP);
 }
 
 /**
@@ -335,10 +334,8 @@ void maketchar(String & ch, unsigned unich, bool & lossy, unsigned codepage)
  */
 unsigned byteToUnicode(unsigned char ch)
 {
-       static unsigned codepage = CP_ACP;
        // NB: Windows always draws in CP_ACP, not CP_THREAD_ACP, so we must use CP_ACP as an internal codepage
-
-       return byteToUnicode(ch, codepage);
+       return byteToUnicode(ch, CP_ACP);
 }
 
 /**
@@ -792,10 +789,10 @@ void buffer::resize(size_t newSize)
 {
        if (capacity < newSize)
        {
-               capacity = newSize;
                unsigned char *tmp = static_cast<unsigned char *>(realloc(ptr, capacity));
                if (tmp == nullptr)
                        throw std::bad_alloc();
+               capacity = newSize;
                ptr = tmp;
        }
 }
@@ -895,21 +892,21 @@ std::string toUTF8(const String& tstr)
 void toUTF8(const String& tstr, std::string& u8str)
 {
 #ifdef _UNICODE
-       u8str.clear();
-       size_t len = tstr.length();
+       const size_t len = tstr.length();
+       u8str.resize(len * 3);
+
        if (len == 0)
                return;
-       u8str.resize(len * 3);
+
        char *p = &u8str[0];
-       for (String::const_iterator it = tstr.begin(); it != tstr.end(); ++it)
+       for (size_t i = 0; i < len; ++i)
        {
-               unsigned uc = *it;
+               unsigned uc = tstr[i];
                if (uc >= 0xd800 && uc < 0xdc00)
                {
-                       ++it;
-                       if (it != tstr.end())
+                       if (++i != len)
                        {
-                               wchar_t uc2 = *it;
+                               wchar_t uc2 = tstr[i];
                                uc = ((uc & 0x3ff) << 10) + (uc2 & 0x3ff) + 0x10000;
                        }
                }
@@ -970,8 +967,8 @@ bool convert(UNICODESET unicoding1, int codepage1, const unsigned char * src, si
                for (size_t i = 0; i < srcbytes; i += 2)
                {
                        // Byte-swap into destination
-                       dest->ptr[i] = src[i+1];
-                       dest->ptr[i+1] = src[i];
+                       uint16_t c = *(const uint16_t*)(src + i);
+                       *(uint16_t*)(dest->ptr + i) = ((c << 8) | (c >> 8));
                }
                dest->ptr[srcbytes] = 0;
                dest->ptr[srcbytes+1] = 0;