mirrors/FreeRDP
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

libwinpr-crt: improve UTF conversion efficiency

Browse Source
This commit is contained in:
Marc-André Moreau 2012-12-16 16:51:05 -05:00
parent 36a1683a5d
commit d7bae41092
2 changed files with 127 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

70
winpr/libwinpr/crt/unicode.c
View File

@ -59,19 +59,6 @@ int MultiByteToWideChar(UINT CodePage, DWORD dwFlags, LPCSTR lpMultiByteStr,
if (cbMultiByte == -1)
cbMultiByte = strlen((char*) lpMultiByteStr) + 1;
if (!lpWideCharStr)
lpWideCharStr = (LPWSTR) malloc((cbMultiByte + 1) * sizeof(WCHAR) * 4);
sourceStart = (const BYTE*) lpMultiByteStr;
targetStart = lpWideCharStr;
result = ConvertUTF8toUTF16(&sourceStart, &sourceStart[cbMultiByte],
&targetStart, &targetStart[((cbMultiByte + 1) * 4) / sizeof(WCHAR)], strictConversion);
length = targetStart - ((WCHAR*) lpWideCharStr);
lpWideCharStr[length] = '\0';
cchWideChar = length;
/*
* if cchWideChar is 0, the function returns the required buffer size
* in characters for lpWideCharStr and makes no use of the output parameter itself.
@ -79,8 +66,26 @@ int MultiByteToWideChar(UINT CodePage, DWORD dwFlags, LPCSTR lpMultiByteStr,
if (cchWideChar == 0)
{
free(lpWideCharStr);
return cchWideChar;
sourceStart = (const BYTE*) lpMultiByteStr;
targetStart = (WCHAR*) NULL;
result = ConvertUTF8toUTF16(&sourceStart, &sourceStart[cbMultiByte],
&targetStart, NULL, strictConversion);
length = targetStart - ((WCHAR*) NULL);
cchWideChar = length;
}
else
{
sourceStart = (const BYTE*) lpMultiByteStr;
targetStart = lpWideCharStr;
result = ConvertUTF8toUTF16(&sourceStart, &sourceStart[cbMultiByte],
&targetStart, &targetStart[cchWideChar], strictConversion);
length = targetStart - ((WCHAR*) lpWideCharStr);
lpWideCharStr[length] = '\0';
cchWideChar = length;
}
return cchWideChar;
@ -109,19 +114,6 @@ int WideCharToMultiByte(UINT CodePage, DWORD dwFlags, LPCWSTR lpWideCharStr, int
if (cchWideChar == -1)
cchWideChar = _wcslen(lpWideCharStr) + 1;
if (!lpMultiByteStr)
lpMultiByteStr = (LPSTR) malloc((cchWideChar + 1) * 4);
sourceStart = (WCHAR*) lpWideCharStr;
targetStart = (BYTE*) lpMultiByteStr;
result = ConvertUTF16toUTF8(&sourceStart, &sourceStart[cchWideChar],
&targetStart, &targetStart[(cchWideChar + 1) * 4], strictConversion);
length = targetStart - ((BYTE*) lpMultiByteStr);
lpMultiByteStr[length] = '\0';
cbMultiByte = length;
/*
* if cbMultiByte is 0, the function returns the required buffer size
* in bytes for lpMultiByteStr and makes no use of the output parameter itself.
@ -129,8 +121,26 @@ int WideCharToMultiByte(UINT CodePage, DWORD dwFlags, LPCWSTR lpWideCharStr, int
if (cbMultiByte == 0)
{
free(lpMultiByteStr);
return cbMultiByte;
sourceStart = (WCHAR*) lpWideCharStr;
targetStart = (BYTE*) NULL;
result = ConvertUTF16toUTF8(&sourceStart, &sourceStart[cchWideChar],
&targetStart, NULL, strictConversion);
length = targetStart - ((BYTE*) NULL);
cbMultiByte = length;
}
else
{
sourceStart = (WCHAR*) lpWideCharStr;
targetStart = (BYTE*) lpMultiByteStr;
result = ConvertUTF16toUTF8(&sourceStart, &sourceStart[cchWideChar],
&targetStart, &targetStart[cbMultiByte], strictConversion);
length = targetStart - ((BYTE*) lpMultiByteStr);
lpMultiByteStr[length] = '\0';
cbMultiByte = length;
}
return cbMultiByte;

109
winpr/libwinpr/crt/utf.c
View File

@ -209,9 +209,16 @@ ConversionResult ConvertUTF16toUTF8(
const WCHAR** sourceStart, const WCHAR* sourceEnd,
BYTE** targetStart, BYTE* targetEnd, ConversionFlags flags)
{
ConversionResult result = conversionOK;
const WCHAR* source = *sourceStart;
BYTE* target = *targetStart;
BYTE* target;
const WCHAR* source;
BOOL computeLength;
ConversionResult result;
computeLength = (!targetEnd) ? TRUE : FALSE;
source = *sourceStart;
target = *targetStart;
result = conversionOK;
while (source < sourceEnd)
{
@ -290,7 +297,7 @@ ConversionResult ConvertUTF16toUTF8(
target += bytesToWrite;
if (target > targetEnd)
if ((target > targetEnd) && (!computeLength))
{
source = oldSource; /* Back up source pointer! */
target -= bytesToWrite;
@ -298,13 +305,46 @@ ConversionResult ConvertUTF16toUTF8(
break;
}
switch (bytesToWrite)
if (!computeLength)
{
/* note: everything falls through. */
case 4: *--target = (BYTE)((ch | byteMark) & byteMask); ch >>= 6;
case 3: *--target = (BYTE)((ch | byteMark) & byteMask); ch >>= 6;
case 2: *--target = (BYTE)((ch | byteMark) & byteMask); ch >>= 6;
case 1: *--target = (BYTE)(ch | firstByteMark[bytesToWrite]);
switch (bytesToWrite)
{
/* note: everything falls through. */
case 4:
*--target = (BYTE)((ch | byteMark) & byteMask);
ch >>= 6;
case 3:
*--target = (BYTE)((ch | byteMark) & byteMask);
ch >>= 6;
case 2:
*--target = (BYTE)((ch | byteMark) & byteMask);
ch >>= 6;
case 1:
*--target = (BYTE)(ch | firstByteMark[bytesToWrite]);
}
}
else
{
switch (bytesToWrite)
{
/* note: everything falls through. */
case 4:
--target;
ch >>= 6;
case 3:
--target;
ch >>= 6;
case 2:
--target;
ch >>= 6;
case 1:
--target;
}
}
target += bytesToWrite;
@ -385,16 +425,23 @@ ConversionResult ConvertUTF8toUTF16(
const BYTE** sourceStart, const BYTE* sourceEnd,
WCHAR** targetStart, WCHAR* targetEnd, ConversionFlags flags)
{
ConversionResult result = conversionOK;
const BYTE* source = *sourceStart;
WCHAR* target = *targetStart;
WCHAR* target;
const BYTE* source;
BOOL computeLength;
ConversionResult result;
computeLength = (!targetEnd) ? TRUE : FALSE;
result = conversionOK;
source = *sourceStart;
target = *targetStart;
while (source < sourceEnd)
{
DWORD ch = 0;
unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
if (source + extraBytesToRead >= sourceEnd)
if ((source + extraBytesToRead) >= sourceEnd)
{
result = sourceExhausted;
break;
@ -422,7 +469,7 @@ ConversionResult ConvertUTF8toUTF16(
ch -= offsetsFromUTF8[extraBytesToRead];
if (target >= targetEnd)
if ((target >= targetEnd) && (!computeLength))
{
source -= (extraBytesToRead + 1); /* Back up source pointer! */
result = targetExhausted;
@ -444,12 +491,18 @@ ConversionResult ConvertUTF8toUTF16(
}
else
{
*target++ = UNI_REPLACEMENT_CHAR;
if (!computeLength)
*target++ = UNI_REPLACEMENT_CHAR;
else
target++;
}
}
else
{
*target++ = (WCHAR) ch; /* normal case */
if (!computeLength)
*target++ = (WCHAR) ch; /* normal case */
else
target++;
}
}
else if (ch > UNI_MAX_UTF16)
@ -462,23 +515,35 @@ ConversionResult ConvertUTF8toUTF16(
}
else
{
*target++ = UNI_REPLACEMENT_CHAR;
if (!computeLength)
*target++ = UNI_REPLACEMENT_CHAR;
else
target++;
}
}
else
{
/* target is a character in range 0xFFFF - 0x10FFFF. */
if (target + 1 >= targetEnd)
if ((target + 1 >= targetEnd) && (!computeLength))
{
source -= (extraBytesToRead+1); /* Back up source pointer! */
source -= (extraBytesToRead + 1); /* Back up source pointer! */
result = targetExhausted;
break;
}
ch -= halfBase;
*target++ = (WCHAR)((ch >> halfShift) + UNI_SUR_HIGH_START);
*target++ = (WCHAR)((ch & halfMask) + UNI_SUR_LOW_START);
if (!computeLength)
{
*target++ = (WCHAR)((ch >> halfShift) + UNI_SUR_HIGH_START);
*target++ = (WCHAR)((ch & halfMask) + UNI_SUR_LOW_START);
}
else
{
target++;
target++;
}
}
}