U_CAPI UChar* U_EXPORT2
u_uastrncpy(UChar *ucs1,
const char *s2,
int32_t n)
{
UChar *target = ucs1;
UErrorCode err = U_ZERO_ERROR;
UConverter *cnv = u_getDefaultConverter(&err);
if(U_SUCCESS(err) && cnv != NULL) {
ucnv_reset(cnv);
ucnv_toUnicode(cnv,
&target,
ucs1+n,
&s2,
s2+u_astrnlen(s2, n),
NULL,
TRUE,
&err);
ucnv_reset(cnv); /* be good citizens */
u_releaseDefaultConverter(cnv);
if(U_FAILURE(err) && (err != U_BUFFER_OVERFLOW_ERROR) ) {
*ucs1 = 0; /* failure */
}
if(target < (ucs1+n)) { /* U_BUFFER_OVERFLOW_ERROR isn't an err, just means no termination will happen. */
*target = 0; /* terminate */
}
} else {
*ucs1 = 0;
}
return ucs1;
}
void convsample_50() {
printf("\n\n==============================================\n"
"Sample 50: C: ucnv_detectUnicodeSignature\n");
//! [ucnv_detectUnicodeSignature]
UErrorCode err = U_ZERO_ERROR;
UBool discardSignature = TRUE; /* set to TRUE to throw away the initial U+FEFF */
char input[] = { '\xEF','\xBB', '\xBF','\x41','\x42','\x43' };
int32_t signatureLength = 0;
const char *encoding = ucnv_detectUnicodeSignature(input,sizeof(input),&signatureLength,&err);
UConverter *conv = NULL;
UChar output[100];
UChar *target = output, *out;
const char *source = input;
if(encoding!=NULL && U_SUCCESS(err)){
// should signature be discarded ?
conv = ucnv_open(encoding, &err);
// do the conversion
ucnv_toUnicode(conv,
&target, output + UPRV_LENGTHOF(output),
&source, input + sizeof(input),
NULL, TRUE, &err);
out = output;
if (discardSignature){
++out; // ignore initial U+FEFF
}
while(out != target) {
printf("%04x ", *out++);
}
puts("");
}
//! [ucnv_detectUnicodeSignature]
puts("");
}
QString QIcuCodec::convertToUnicode(const char *chars, int length, QTextCodec::ConverterState *state) const
{
UConverter *conv = getConverter(state);
QString string(length + 2, Qt::Uninitialized);
const char *end = chars + length;
int convertedChars = 0;
while (1) {
UChar *uc = (UChar *)string.data();
UChar *ucEnd = uc + string.length();
uc += convertedChars;
UErrorCode error = U_ZERO_ERROR;
ucnv_toUnicode(conv,
&uc, ucEnd,
&chars, end,
0, false, &error);
if (!U_SUCCESS(error) && error != U_BUFFER_OVERFLOW_ERROR) {
qDebug() << "convertToUnicode failed:" << u_errorName(error);
break;
}
convertedChars = uc - (UChar *)string.data();
if (chars >= end)
break;
string.resize(string.length()*2);
}
string.resize(convertedChars);
if (!state)
ucnv_close(conv);
return string;
}
int TextCodecICU::decodeToBuffer(UChar* target, UChar* targetLimit, const char*& source, const char* sourceLimit, int32_t* offsets, bool flush, UErrorCode& err)
{
UChar* targetStart = target;
err = U_ZERO_ERROR;
ucnv_toUnicode(m_converterICU, &target, targetLimit, &source, sourceLimit, offsets, flush, &err);
return target - targetStart;
}
UChar*
ufmt_defaultCPToUnicode(const char *s, int32_t sSize,
UChar *target, int32_t tSize)
{
UChar *alias;
UErrorCode status = U_ZERO_ERROR;
UConverter *defConverter = u_getDefaultConverter(&status);
if(U_FAILURE(status) || defConverter == 0)
return 0;
if(sSize <= 0) {
sSize = uprv_strlen(s) + 1;
}
/* perform the conversion in one swoop */
if(target != 0) {
alias = target;
ucnv_toUnicode(defConverter, &alias, alias + tSize, &s, s + sSize - 1,
NULL, TRUE, &status);
/* add the null terminator */
*alias = 0x0000;
}
u_releaseDefaultConverter(defConverter);
return target;
}
inline int
mod_websocket_conv(UConverter *to, UConverter *from,
char **dst, size_t *dstsiz,
const char *src, size_t srcsiz) {
UErrorCode err = U_ZERO_ERROR;
size_t unisiz;
UChar *unibuf, *punibuf, *ppunibuf;
char *pdst;
if (srcsiz == 0) {
return -1;
}
if (!to) {
*dst = (char *)malloc(srcsiz + 1);
if (*dst == NULL) {
return -1;
}
memcpy(*dst, src, srcsiz);
(*dst)[srcsiz] = '\0';
*dstsiz = srcsiz;
return 0;
}
if (!from || !dst || !src || !dstsiz) {
return -1;
}
unisiz = srcsiz / ucnv_getMinCharSize(from);
unibuf = (UChar *)malloc(sizeof(UChar) * unisiz + 1);
if (!unibuf) {
return -1;
}
punibuf = unibuf;
ucnv_toUnicode(from, &punibuf, punibuf + unisiz,
&src, src + srcsiz, 0, 0, &err);
if (U_FAILURE(err)) {
free(unibuf);
return -1;
}
*punibuf = '\0';
*dstsiz = (punibuf - unibuf) * ucnv_getMaxCharSize(to);
*dst = (char *)malloc(*dstsiz + 1);
if (!*dst) {
free(unibuf);
return -1;
}
pdst = *dst;
ppunibuf = unibuf;
ucnv_fromUnicode(to, &pdst, pdst + *dstsiz,
(const UChar **)&ppunibuf, punibuf, 0, 0, &err);
free(unibuf);
if (U_FAILURE(err)) {
free(*dst);
return -1;
}
*pdst = '\0';
*dstsiz = pdst - *dst;
return 0;
}
static jint NativeConverter_decode(JNIEnv* env, jclass, jlong address,
jbyteArray source, jint sourceEnd, jcharArray target, jint targetEnd,
jintArray data, jboolean flush) {
UConverter* cnv = toUConverter(address);
if (cnv == NULL) {
maybeThrowIcuException(env, "toUConverter", U_ILLEGAL_ARGUMENT_ERROR);
return U_ILLEGAL_ARGUMENT_ERROR;
}
ScopedByteArrayRO uSource(env, source);
if (uSource.get() == NULL) {
maybeThrowIcuException(env, "uSource", U_ILLEGAL_ARGUMENT_ERROR);
return U_ILLEGAL_ARGUMENT_ERROR;
}
ScopedCharArrayRW uTarget(env, target);
if (uTarget.get() == NULL) {
maybeThrowIcuException(env, "uTarget", U_ILLEGAL_ARGUMENT_ERROR);
return U_ILLEGAL_ARGUMENT_ERROR;
}
ScopedIntArrayRW myData(env, data);
if (myData.get() == NULL) {
maybeThrowIcuException(env, "myData", U_ILLEGAL_ARGUMENT_ERROR);
return U_ILLEGAL_ARGUMENT_ERROR;
}
// Do the conversion.
jint* sourceOffset = &myData[0];
jint* targetOffset = &myData[1];
const char* mySource = reinterpret_cast<const char*>(uSource.get() + *sourceOffset);
const char* mySourceLimit = reinterpret_cast<const char*>(uSource.get() + sourceEnd);
UChar* cTarget = uTarget.get() + *targetOffset;
const UChar* cTargetLimit = uTarget.get() + targetEnd;
UErrorCode errorCode = U_ZERO_ERROR;
ucnv_toUnicode(cnv, &cTarget, cTargetLimit, &mySource, mySourceLimit, NULL, flush, &errorCode);
*sourceOffset = mySource - reinterpret_cast<const char*>(uSource.get()) - *sourceOffset;
*targetOffset = cTarget - uTarget.get() - *targetOffset;
// If there was an error, count the problematic bytes.
if (errorCode == U_ILLEGAL_CHAR_FOUND || errorCode == U_INVALID_CHAR_FOUND ||
errorCode == U_TRUNCATED_CHAR_FOUND) {
int8_t invalidByteCount = 32;
char invalidBytes[32] = {'\0'};
UErrorCode minorErrorCode = U_ZERO_ERROR;
ucnv_getInvalidChars(cnv, invalidBytes, &invalidByteCount, &minorErrorCode);
if (U_SUCCESS(minorErrorCode)) {
myData[2] = invalidByteCount;
}
}
// Managed code handles some cases; throw all other errors.
if (shouldCodecThrow(flush, errorCode)) {
maybeThrowIcuException(env, "ucnv_toUnicode", errorCode);
}
return errorCode;
}
DeprecatedString StreamingTextDecoderICU::convertUsingICU(const unsigned char* chs, int len, bool flush)
{
// Get a converter for the passed-in encoding.
if (!m_converterICU) {
createICUConverter();
if (!m_converterICU)
return DeprecatedString();
}
DeprecatedString result("");
result.reserve(len);
UChar buffer[ConversionBufferSize];
const char* source = reinterpret_cast<const char*>(chs);
const char* sourceLimit = source + len;
int32_t* offsets = NULL;
UErrorCode err;
do {
UChar* target = buffer;
const UChar* targetLimit = target + ConversionBufferSize;
err = U_ZERO_ERROR;
ucnv_toUnicode(m_converterICU, &target, targetLimit, &source, sourceLimit, offsets, flush, &err);
int count = target - buffer;
appendOmittingBOM(result, reinterpret_cast<const UChar*>(buffer), count * sizeof(UChar));
} while (err == U_BUFFER_OVERFLOW_ERROR);
if (U_FAILURE(err)) {
// flush the converter so it can be reused, and not be bothered by this error.
do {
UChar *target = buffer;
const UChar *targetLimit = target + ConversionBufferSize;
err = U_ZERO_ERROR;
ucnv_toUnicode(m_converterICU, &target, targetLimit, &source, sourceLimit, offsets, true, &err);
} while (source < sourceLimit);
LOG_ERROR("ICU conversion error");
return DeprecatedString();
}
return result;
}
U_CAPI UBool U_EXPORT2
ucbuf_autodetect_fs(FileStream* in, const char** cp, UConverter** conv, int32_t* signatureLength, UErrorCode* error){
char start[8];
int32_t numRead;
UChar target[1]={ 0 };
UChar* pTarget;
const char* pStart;
/* read a few bytes */
numRead=T_FileStream_read(in, start, sizeof(start));
*cp = ucnv_detectUnicodeSignature(start, numRead, signatureLength, error);
/* unread the bytes beyond what was consumed for U+FEFF */
T_FileStream_rewind(in);
if (*signatureLength > 0) {
numRead = T_FileStream_read(in, start, *signatureLength);
}
if(*cp==NULL){
*conv =NULL;
return FALSE;
}
/* open the converter for the detected Unicode charset */
*conv = ucnv_open(*cp,error);
/* convert and ignore initial U+FEFF, and the buffer overflow */
pTarget = target;
pStart = start;
ucnv_toUnicode(*conv, &pTarget, target+1, &pStart, start+*signatureLength, NULL, FALSE, error);
*signatureLength = (int32_t)(pStart - start);
if(*error==U_BUFFER_OVERFLOW_ERROR) {
*error=U_ZERO_ERROR;
}
/* verify that we successfully read exactly U+FEFF */
if(U_SUCCESS(*error) && (pTarget!=(target+1) || target[0]!=0xfeff)) {
*error=U_INTERNAL_PROGRAM_ERROR;
}
return TRUE;
}
static void printCString(FILE *out, UConverter *converter, const char *str, int32_t len) {
UChar buf[256];
const char *strEnd;
if (len < 0) {
len = (int32_t)uprv_strlen(str);
}
strEnd = str + len;
do {
UErrorCode err = U_ZERO_ERROR;
UChar *bufp = buf, *bufend = buf + (sizeof(buf)/sizeof(buf[0])) - 1 ;
ucnv_toUnicode(defaultConverter, &bufp, bufend, &str, strEnd, 0, 0, &err);
*bufp = 0;
printString(out, converter, buf, (int32_t)(bufp - buf));
} while (str < strEnd);
}
/* params : desc : the document descriptor
* buf : destination buffer for UTF-16 data
* return : the length of the paragraph
* NO_MORE_DATA if there is no more paragraph
* ERR_STREAMFILE if an error occured
*
* reads the next paragraph and converts to UTF-16
*/
int p_read_content(struct doc_descriptor *desc, UChar *buf) {
char *outputbuf, *src;
UChar *dest;
int len;
UErrorCode err;
len = 0;
outputbuf = (char *) malloc(INTERNAL_BUFSIZE);
/* reading the next paragraph */
memset(outputbuf, '\x00', INTERNAL_BUFSIZE);
len = parse(desc, outputbuf);
if (len > 0) {
(desc->nb_par_read) += 1;
/* converting to UTF-16 */
err = U_ZERO_ERROR;
dest = buf;
src = outputbuf;
ucnv_toUnicode(desc->conv, &dest, dest + 2*INTERNAL_BUFSIZE,
&src, outputbuf + strlen(outputbuf), NULL, FALSE, &err);
len = 2*(dest - buf);
if (U_FAILURE(err)) {
free(outputbuf);
outputbuf = NULL;
fprintf(stderr, "Unable to convert buffer\n");
return ERR_ICU;
}
}
if(outputbuf != NULL) {
free(outputbuf);
}
return len;
}
static int convertUTF8toUChar(const char *src, UChar *dst, int len)
{
static UConverter *c;
UErrorCode status;
UChar *p = dst;
const char *s = src;
if (!c) {
status = U_ZERO_ERROR;
c = ucnv_open("UTF-8", &status);
if (!c) {
fprintf(stderr, "CouchStore CollateJSON: Couldn't initialize ICU (%d)\n", (int)status);
abort();
}
}
while (len) {
unsigned char ch = (unsigned char)(*s);
if ((ch & 0x80)) {
goto icu_conv;
}
*p++ = (UChar)(ch);
s++;
len--;
}
return p - dst;
icu_conv:
status = U_ZERO_ERROR;
ucnv_toUnicode(c, &p, p + len, &s, s + len, NULL, TRUE, &status);
if (U_FAILURE(status)) {
return -1;
}
return p - dst;
}
/* rewind the buf and file stream */
U_CAPI void U_EXPORT2
ucbuf_rewind(UCHARBUF* buf,UErrorCode* error){
if(error==NULL || U_FAILURE(*error)){
return;
}
if(buf){
buf->currentPos=buf->buffer;
buf->bufLimit=buf->buffer;
T_FileStream_rewind(buf->in);
buf->remaining=T_FileStream_size(buf->in)-buf->signatureLength;
ucnv_resetToUnicode(buf->conv);
if(buf->signatureLength>0) {
UChar target[1]={ 0 };
UChar* pTarget;
char start[8];
const char* pStart;
int32_t numRead;
/* read the signature bytes */
numRead=T_FileStream_read(buf->in, start, buf->signatureLength);
/* convert and ignore initial U+FEFF, and the buffer overflow */
pTarget = target;
pStart = start;
ucnv_toUnicode(buf->conv, &pTarget, target+1, &pStart, start+numRead, NULL, FALSE, error);
if(*error==U_BUFFER_OVERFLOW_ERROR) {
*error=U_ZERO_ERROR;
}
/* verify that we successfully read exactly U+FEFF */
if(U_SUCCESS(*error) && (numRead!=buf->signatureLength || pTarget!=(target+1) || target[0]!=0xfeff)) {
*error=U_INTERNAL_PROGRAM_ERROR;
}
}
}
}
int
getText (struct doc_descriptor *desc, UChar * buf, int size)
{
struct meta *meta = NULL;
char buf2[BUFSIZE];
UErrorCode err;
char *src;
UChar *dest, esc[3];
UChar name[1024], value[1024];
int len, i, isMarkup, isJavascript, isMeta, l, j;
int dangerousCut, fini, r, offset, endOfFile, space_added;
space_added = 0;
l = 0;
fini = 0;
endOfFile = 0;
isJavascript = 0;
dangerousCut = 0;
isMarkup = 0;
isMeta = 0;
len = read (desc->fd, buf2, BUFSIZE);
while (!fini && len > 0 && 2*l < size - 2)
{
/* consuming buffer */
for (i = 0; 2*l < size - 2 && i < len && !dangerousCut && !fini; i++)
{
/* end of buffer are possible points of failure
if a markup or a token is cut, it will not be
parsed. */
if (!endOfFile && i > len - 9 && (!strncmp (buf2 + i, "\x3c", 1) ||
!strncmp (buf2 + i, "\x26", 1)))
{
dangerousCut = 1;
break;
}
/* detecting end of javascript */
if (isJavascript
&& !strncasecmp (buf2 + i, "</script>", 9))
{
isJavascript = 0;
i += 9;
}
/* detecting new paragraph */
if (l > 0 && !isJavascript
&& (!strncasecmp (buf2 + i, "<p", 2)
|| !strncasecmp (buf2 + i, "<br", 3)
|| !strncasecmp (buf2 + i, "<div", 4)))
{
fini = 1;
i += 2;
while (strncmp (buf2 + i, ">", 1))
{
i++;
}
lseek (desc->fd, i - len, SEEK_CUR);
break;
}
/* detecting begining of markup */
if (!isJavascript && !isMarkup && !strncmp (buf2 + i, "\x3c", 1))
{
/* detecting begining of javascript */
if (!strncasecmp (buf2 + i, "<script", 7))
{
isJavascript = 1;
}
else if (!strncasecmp (buf2 + i, "<title", 6))
{
err = U_ZERO_ERROR;
/* finding last metadata of desc */
if (desc->meta == NULL)
{
meta = (struct meta *) malloc (sizeof (struct meta));
desc->meta = meta;
}
else
{
meta = desc->meta;
while (meta->next != NULL)
{
meta = meta->next;
}
meta->next =
(struct meta *) malloc (sizeof (struct meta));
meta = meta->next;
}
meta->next = NULL;
meta->name = (UChar *) malloc (12);
/* filling name field */
meta->name_length =
2 * ucnv_toUChars (desc->conv, meta->name, 12, "title", 5,
&err);
meta->name_length = u_strlen (meta->name);
if (U_FAILURE (err))
{
printf ("error icu\n");
return ERR_ICU;
}
isMeta = 1;
}
else if (!strncasecmp (buf2 + i, "<meta", 5))
{
i += 5;
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
for (; strncasecmp (buf2 + i, "name=\"", 6) &&
strncmp (buf2 + i, "\x3E", 1); i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
}
if (!strncmp (buf2 + i, "\x3E", 1))
{
continue;
}
else
{
i += 6;
/* get metadata name */
memset (name, '\x00', 2048);
for (j = 0; len != 0 && strncmp (buf2 + i, "\"", 1);
i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
if (!strncmp (buf2 + i, "\x26", 1))
{
memset (esc, '\x00', 6);
offset = escapeChar (desc, buf2 + i, esc);
memcpy (name + j, esc, 2 * u_strlen (esc));
j += u_strlen (esc);
i += (offset - 1);
}
else
{
/* filling name buffer */
dest = name + j;
src = buf2 + i;
err = U_ZERO_ERROR;
ucnv_toUnicode (desc->conv, &dest, name + 1024,
&src, buf2 + i + 1, NULL, FALSE,
&err);
if (U_FAILURE (err))
{
fprintf (stderr,
"Unable to convert buffer\n");
return ERR_ICU;
}
j += (dest - name - j);
}
}
/* get metadata value */
for (; strncasecmp (buf2 + i, "content=\"", 9) && strncmp (buf2 + i, "\x3E", 1); i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
}
i += 9;
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
memset (value, '\x00', 2048);
for (j = 0; len != 0 && strncmp (buf2 + i, "\"", 1);
i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
if (!strncmp (buf2 + i, "\x26", 1))
{
memset (esc, '\x00', 6);
offset = escapeChar (desc, buf2 + i, esc);
memcpy (value + j, esc, 2 * u_strlen (esc));
j += u_strlen (esc);
i += (offset - 1);
}
else
{
/* filling value buffer */
dest = value + j;
src = buf2 + i;
err = U_ZERO_ERROR;
ucnv_toUnicode (desc->conv, &dest, value + 1024,
&src, buf2 + i + 1, NULL, FALSE,
&err);
if (U_FAILURE (err))
{
fprintf (stderr,
"Unable to convert buffer\n");
return ERR_ICU;
}
j += (dest - value - j);
}
}
/* insert metadata in list */
if (desc->meta == NULL)
{
meta =
(struct meta *) malloc (sizeof (struct meta));
desc->meta = meta;
}
else
{
meta = desc->meta;
while (meta->next != NULL)
{
meta = meta->next;
}
meta->next =
(struct meta *) malloc (sizeof (struct meta));
meta = meta->next;
}
meta->next = NULL;
meta->name = (UChar *) malloc (2 * u_strlen (name) + 2);
meta->value =
(UChar *) malloc (2 * u_strlen (value) + 2);
memset (meta->name, '\x00', 2 * u_strlen (name) + 2);
memset (meta->value, '\x00', 2 * u_strlen (value) + 2);
memcpy (meta->name, name, 2 * u_strlen (name));
memcpy (meta->value, value, 2 * u_strlen (value));
meta->name_length = u_strlen (name);
meta->value_length = u_strlen (value);
for (; strncmp (buf2 + i, "\x3E", 1); i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
}
continue;
}
}
else
{
isMarkup = 1;
}
}
/* get metadata value */
if (!isJavascript && isMeta)
{
for (; len != 0 && strncmp (buf2 + i, "\x3E", 1); i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
}
i++;
memset (value, '\x00', 2048);
for (j = 0; len != 0 && strncmp (buf2 + i, "\x3C", 1); i++)
{
if (i >= size - 9)
{
strncpy (buf2, buf2 + i, len - i);
len =
read (desc->fd, buf2 + i,
BUFSIZE - len + i) + len - i;
i = 0;
}
if (!strncmp (buf2 + i, "\x26", 1))
{
memset (esc, '\x00', 6);
offset = escapeChar (desc, buf2 + i, esc);
memcpy (value + j, esc, 2 * u_strlen (esc));
j += u_strlen (esc);
i += (offset - 1);
}
else
{
/* filling value buffer */
dest = value + j;
src = buf2 + i;
err = U_ZERO_ERROR;
ucnv_toUnicode (desc->conv, &dest, value + 1024,
&src, buf2 + i + 1, NULL, FALSE, &err);
if (U_FAILURE (err))
{
fprintf (stderr, "Unable to convert buffer\n");
return ERR_ICU;
}
j += (dest - value - j);
}
}
meta->value = (UChar *) malloc (2 * (j + 1));
memcpy (meta->value, value, 2 * u_strlen (value));
meta->value_length = u_strlen (value);
isMeta = 0;
i += 7;
continue;
}
/* detecting end of markup */
if (!isJavascript && isMarkup && !strncmp (buf2 + i, "\x3e", 1))
{
if (!space_added && l > 0)
{
buf[l] = 0x20;
l ++;
space_added = 1;
}
isMarkup = 0;
}
/* handling text */
if (!isJavascript && !isMarkup && strncmp (buf2 + i, "\x3e", 1))
{
if (strncmp (buf2 + i, "\n", 1) && strncmp (buf2 + i, "\t", 1) && strncmp (buf2 + i, "\r", 1))
{
/* converting tokens */
if (!isJavascript && !isMarkup
&& !strncmp (buf2 + i, "\x26", 1))
{
memset (esc, '\x00', 6);
offset = escapeChar (desc, buf2 + i, esc);
if (memcmp (esc, "\x20\x00", u_strlen (esc)))
{
memcpy (buf + l, esc, 2 * u_strlen (esc));
l += u_strlen (esc);
space_added = 0;
}
else {
if (!space_added){
buf[l] = 0x20;
space_added = 1;
l++;
}
}
i += (offset - 1);
}
else
{
if (buf2[i] != 0x20 || !space_added){
/* filling output buffer */
dest = buf + l;
src = buf2 + i;
err = U_ZERO_ERROR;
ucnv_toUnicode (desc->conv, &dest, buf + size / 2,
&src, buf2 + i + 1, NULL, FALSE, &err);
if (U_FAILURE (err))
{
fprintf (stderr, "Unable to convert buffer\n");
return ERR_ICU;
}
l += (dest - buf - l);
if (buf2[i] == 0x20) {space_added = 1;} else {space_added=0;}
}
}
}
else
{
/* replace tabs and eol by spaces */
if (!space_added){
buf[l] = 0x20;
space_added = 1;
l++;
}
}
}
}
/* filling new buffer correctly */
if (!fini)
{
if (dangerousCut)
{
r = len - i;
strncpy (buf2, buf2 + i, r);
len = read (desc->fd, buf2 + r, BUFSIZE - r) + r;
if (len < 9)
{
endOfFile = 1;
}
dangerousCut = 0;
}
else
{
len = read (desc->fd, buf2, BUFSIZE);
}
}
}
/* ending buffer properly */
if (l > 0)
{
buf[l] = 0x20;
return 2*l;
}
if (len == 0)
{
return NO_MORE_DATA;
}
return 2*l;
}
void charsetConverter_icu::convert
(utility::inputStream& in, utility::outputStream& out, status* st)
{
UErrorCode err = U_ZERO_ERROR;
ucnv_reset(m_from);
ucnv_reset(m_to);
if (st)
new (st) status();
// From buffers
byte_t cpInBuffer[16]; // stream data put here
const size_t outSize = ucnv_getMinCharSize(m_from) * sizeof(cpInBuffer) * sizeof(UChar);
std::vector <UChar> uOutBuffer(outSize); // Unicode chars end up here
// To buffers
// converted (char) data end up here
const size_t cpOutBufferSz = ucnv_getMaxCharSize(m_to) * outSize;
std::vector <char> cpOutBuffer(cpOutBufferSz);
// Tell ICU what to do when encountering an illegal byte sequence
if (m_options.silentlyReplaceInvalidSequences)
{
// Set replacement chars for when converting from Unicode to codepage
icu::UnicodeString substString(m_options.invalidSequence.c_str());
ucnv_setSubstString(m_to, substString.getTerminatedBuffer(), -1, &err);
if (U_FAILURE(err))
throw exceptions::charset_conv_error("[ICU] Error when setting substitution string.");
}
else
{
// Tell ICU top stop (and return an error) on illegal byte sequences
ucnv_setToUCallBack
(m_from, UCNV_TO_U_CALLBACK_STOP, UCNV_SUB_STOP_ON_ILLEGAL, NULL, NULL, &err);
if (U_FAILURE(err))
throw exceptions::charset_conv_error("[ICU] Error when setting ToU callback.");
ucnv_setFromUCallBack
(m_to, UCNV_FROM_U_CALLBACK_STOP, UCNV_SUB_STOP_ON_ILLEGAL, NULL, NULL, &err);
if (U_FAILURE(err))
throw exceptions::charset_conv_error("[ICU] Error when setting FromU callback.");
}
// Input data available
while (!in.eof())
{
// Read input data into buffer
size_t inLength = in.read(cpInBuffer, sizeof(cpInBuffer));
// Beginning of read data
const char* source = reinterpret_cast <const char*>(&cpInBuffer[0]);
const char* sourceLimit = source + inLength; // end + 1
UBool flush = in.eof(); // is this last run?
UErrorCode toErr;
// Loop until all source has been processed
do
{
// Set up target pointers
UChar* target = &uOutBuffer[0];
UChar* targetLimit = &target[0] + outSize;
toErr = U_ZERO_ERROR;
ucnv_toUnicode(m_from, &target, targetLimit,
&source, sourceLimit, NULL, flush, &toErr);
if (st)
st->inputBytesRead += (source - reinterpret_cast <const char*>(&cpInBuffer[0]));
if (toErr != U_BUFFER_OVERFLOW_ERROR && U_FAILURE(toErr))
{
if (toErr == U_INVALID_CHAR_FOUND ||
toErr == U_TRUNCATED_CHAR_FOUND ||
toErr == U_ILLEGAL_CHAR_FOUND)
{
// Error will be thrown later (*)
}
else
{
throw exceptions::charset_conv_error("[ICU] Error converting to Unicode from " + m_source.getName());
}
}
// The Unicode source is the buffer just written and the limit
// is where the previous conversion stopped (target is moved in the conversion)
const UChar* uSource = &uOutBuffer[0];
UChar* uSourceLimit = &target[0];
UErrorCode fromErr;
// Loop until converted chars are fully written
do
{
char* cpTarget = &cpOutBuffer[0];
const char* cpTargetLimit = &cpOutBuffer[0] + cpOutBufferSz;
fromErr = U_ZERO_ERROR;
// Write converted bytes (Unicode) to destination codepage
ucnv_fromUnicode(m_to, &cpTarget, cpTargetLimit,
&uSource, uSourceLimit, NULL, flush, &fromErr);
if (st)
{
// Decrement input bytes count by the number of input bytes in error
char errBytes[16];
int8_t errBytesLen = sizeof(errBytes);
UErrorCode errBytesErr = U_ZERO_ERROR;
ucnv_getInvalidChars(m_from, errBytes, &errBytesLen, &errBytesErr);
st->inputBytesRead -= errBytesLen;
st->outputBytesWritten += cpTarget - &cpOutBuffer[0];
}
// (*) If an error occurred while converting from input charset, throw it now
if (toErr == U_INVALID_CHAR_FOUND ||
toErr == U_TRUNCATED_CHAR_FOUND ||
toErr == U_ILLEGAL_CHAR_FOUND)
{
throw exceptions::illegal_byte_sequence_for_charset();
}
if (fromErr != U_BUFFER_OVERFLOW_ERROR && U_FAILURE(fromErr))
{
if (fromErr == U_INVALID_CHAR_FOUND ||
fromErr == U_TRUNCATED_CHAR_FOUND ||
fromErr == U_ILLEGAL_CHAR_FOUND)
{
throw exceptions::illegal_byte_sequence_for_charset();
}
else
{
throw exceptions::charset_conv_error("[ICU] Error converting from Unicode to " + m_dest.getName());
}
}
// Write to destination stream
out.write(&cpOutBuffer[0], (cpTarget - &cpOutBuffer[0]));
} while (fromErr == U_BUFFER_OVERFLOW_ERROR);
} while (toErr == U_BUFFER_OVERFLOW_ERROR);
}
}
// ---------------------------------------------------------------------------
// ICUTranscoder: The virtual transcoder API
// ---------------------------------------------------------------------------
unsigned int
ICUTranscoder::transcodeFrom(const XMLByte* const srcData
, const unsigned int srcCount
, XMLCh* const toFill
, const unsigned int maxChars
, unsigned int& bytesEaten
, unsigned char* const charSizes)
{
// If debugging, insure the block size is legal
#if defined(XERCES_DEBUG)
checkBlockSize(maxChars);
#endif
// Set up pointers to the start and end of the source buffer
const XMLByte* startSrc = srcData;
const XMLByte* endSrc = srcData + srcCount;
//
// And now do the target buffer. This works differently according to
// whether XMLCh and UChar are the same size or not.
//
UChar* startTarget;
if (sizeof(XMLCh) == sizeof(UChar))
startTarget = (UChar*)toFill;
else
startTarget = (UChar*) getMemoryManager()->allocate
(
maxChars * sizeof(UChar)
);//new UChar[maxChars];
UChar* orgTarget = startTarget;
//
// Transoode the buffer. Buffer overflow errors are normal, occuring
// when the raw input buffer holds more characters than will fit in
// the Unicode output buffer.
//
UErrorCode err = U_ZERO_ERROR;
ucnv_toUnicode
(
fConverter
, &startTarget
, startTarget + maxChars
, (const char**)&startSrc
, (const char*)endSrc
, (fFixed ? 0 : (int32_t*)fSrcOffsets)
, false
, &err
);
if ((err != U_ZERO_ERROR) && (err != U_BUFFER_OVERFLOW_ERROR))
{
if (orgTarget != (UChar*)toFill)
getMemoryManager()->deallocate(orgTarget);//delete [] orgTarget;
if (fFixed)
{
XMLCh tmpBuf[17];
XMLString::binToText((unsigned int)(*startTarget), tmpBuf, 16, 16, getMemoryManager());
ThrowXMLwithMemMgr2
(
TranscodingException
, XMLExcepts::Trans_BadSrcCP
, tmpBuf
, getEncodingName()
, getMemoryManager()
);
}
else
{
ThrowXMLwithMemMgr(TranscodingException, XMLExcepts::Trans_BadSrcSeq, getMemoryManager());
}
}
// Calculate the bytes eaten and store in caller's param
bytesEaten = startSrc - srcData;
// And the characters decoded
const unsigned int charsDecoded = startTarget - orgTarget;
//
// Translate the array of char offsets into an array of character
// sizes, which is what the transcoder interface semantics requires.
// If its fixed, then we can optimize it.
//
if (fFixed)
{
const unsigned char fillSize = (unsigned char)ucnv_getMaxCharSize(fConverter);
memset(charSizes, fillSize, maxChars);
}
else
{
//
// We have to convert the series of offsets into a series of
// sizes. If just one char was decoded, then its the total bytes
// eaten. Otherwise, do a loop and subtract out each element from
// its previous element.
//
if (charsDecoded == 1)
{
charSizes[0] = (unsigned char)bytesEaten;
}
else
{
// ICU does not return an extra element to allow us to figure
// out the last char size, so we have to compute it from the
// total bytes used.
unsigned int index;
for (index = 0; index < charsDecoded - 1; index++)
{
charSizes[index] = (unsigned char)(fSrcOffsets[index + 1]
- fSrcOffsets[index]);
}
if( charsDecoded > 0 ) {
charSizes[charsDecoded - 1] = (unsigned char)(bytesEaten
- fSrcOffsets[charsDecoded - 1]);
}
}
}
//
// If XMLCh and UChar are not the same size, then we need to copy over
// the temp buffer to the new one.
//
if (sizeof(UChar) != sizeof(XMLCh))
{
XMLCh* outPtr = toFill;
startTarget = orgTarget;
for (unsigned int index = 0; index < charsDecoded; index++)
*outPtr++ = XMLCh(*startTarget++);
// And delete the temp buffer
getMemoryManager()->deallocate(orgTarget);//delete [] orgTarget;
}
// Return the chars we put into the target buffer
return charsDecoded;
}
// Convert a file from one encoding to another
static UBool convertFile(const char *pname,
const char *fromcpage,
UConverterToUCallback toucallback,
const void *touctxt,
const char *tocpage,
UConverterFromUCallback fromucallback,
const void *fromuctxt,
int fallback,
size_t bufsz,
const char *translit,
const char *infilestr,
FILE * outfile, int verbose)
{
FILE *infile;
UBool ret = TRUE;
UConverter *convfrom = 0;
UConverter *convto = 0;
UErrorCode err = U_ZERO_ERROR;
UBool flush;
const char *cbufp;
char *bufp;
char *buf = 0;
uint32_t infoffset = 0, outfoffset = 0; /* Where we are in the file, for error reporting. */
const UChar *unibufbp;
UChar *unibufp;
UChar *unibuf = 0;
int32_t *fromoffsets = 0, *tooffsets = 0;
size_t rd, wr, tobufsz;
#if !UCONFIG_NO_TRANSLITERATION
Transliterator *t = 0; // Transliterator acting on Unicode data.
#endif
UnicodeString u; // String to do the transliteration.
// Open the correct input file or connect to stdin for reading input
if (infilestr != 0 && strcmp(infilestr, "-")) {
infile = fopen(infilestr, "rb");
if (infile == 0) {
UnicodeString str1(infilestr, "");
str1.append((UChar32) 0);
UnicodeString str2(strerror(errno), "");
str2.append((UChar32) 0);
initMsg(pname);
u_wmsg(stderr, "cantOpenInputF", str1.getBuffer(), str2.getBuffer());
return FALSE;
}
} else {
infilestr = "-";
infile = stdin;
#ifdef WIN32
if (setmode(fileno(stdin), O_BINARY) == -1) {
initMsg(pname);
u_wmsg(stderr, "cantSetInBinMode");
return FALSE;
}
#endif
}
if (verbose) {
fprintf(stderr, "%s:\n", infilestr);
}
#if !UCONFIG_NO_TRANSLITERATION
// Create transliterator as needed.
if (translit != NULL && *translit) {
UParseError parse;
UnicodeString str(translit), pestr;
/* Create from rules or by ID as needed. */
parse.line = -1;
if (uprv_strchr(translit, ':') || uprv_strchr(translit, '>') || uprv_strchr(translit, '<') || uprv_strchr(translit, '>')) {
t = Transliterator::createFromRules("Uconv", str, UTRANS_FORWARD, parse, err);
} else {
t = Transliterator::createInstance(translit, UTRANS_FORWARD, err);
}
if (U_FAILURE(err)) {
str.append((UChar32) 0);
initMsg(pname);
if (parse.line >= 0) {
UChar linebuf[20], offsetbuf[20];
uprv_itou(linebuf, 20, parse.line, 10, 0);
uprv_itou(offsetbuf, 20, parse.offset, 10, 0);
u_wmsg(stderr, "cantCreateTranslitParseErr", str.getBuffer(),
u_wmsg_errorName(err), linebuf, offsetbuf);
} else {
u_wmsg(stderr, "cantCreateTranslit", str.getBuffer(),
u_wmsg_errorName(err));
}
if (t) {
delete t;
t = 0;
}
goto error_exit;
}
}
#endif
// Create codepage converter. If the codepage or its aliases weren't
// available, it returns NULL and a failure code. We also set the
// callbacks, and return errors in the same way.
convfrom = ucnv_open(fromcpage, &err);
if (U_FAILURE(err)) {
UnicodeString str(fromcpage, (int32_t)(uprv_strlen(fromcpage) + 1));
initMsg(pname);
u_wmsg(stderr, "cantOpenFromCodeset", str.getBuffer(),
u_wmsg_errorName(err));
goto error_exit;
}
ucnv_setToUCallBack(convfrom, toucallback, touctxt, 0, 0, &err);
if (U_FAILURE(err)) {
initMsg(pname);
u_wmsg(stderr, "cantSetCallback", u_wmsg_errorName(err));
goto error_exit;
}
convto = ucnv_open(tocpage, &err);
if (U_FAILURE(err)) {
UnicodeString str(tocpage, (int32_t)(uprv_strlen(tocpage) + 1));
initMsg(pname);
u_wmsg(stderr, "cantOpenToCodeset", str.getBuffer(),
u_wmsg_errorName(err));
goto error_exit;
}
ucnv_setFromUCallBack(convto, fromucallback, fromuctxt, 0, 0, &err);
if (U_FAILURE(err)) {
initMsg(pname);
u_wmsg(stderr, "cantSetCallback", u_wmsg_errorName(err));
goto error_exit;
}
ucnv_setFallback(convto, fallback);
// To ensure that the buffer always is of enough size, we
// must take the worst case scenario, that is the character in
// the codepage that uses the most bytes and multiply it against
// the buffer size.
// use bufsz+1 to allow for additional BOM/signature character (U+FEFF)
tobufsz = (bufsz+1) * ucnv_getMaxCharSize(convto);
buf = new char[tobufsz];
unibuf = new UChar[bufsz];
fromoffsets = new int32_t[bufsz];
tooffsets = new int32_t[tobufsz];
// OK, we can convert now.
do {
char willexit = 0;
rd = fread(buf, 1, bufsz, infile);
if (ferror(infile) != 0) {
UnicodeString str(strerror(errno));
str.append((UChar32) 0);
initMsg(pname);
u_wmsg(stderr, "cantRead", str.getBuffer());
goto error_exit;
}
// Convert the read buffer into the new coding
// After the call 'unibufp' will be placed on the last
// character that was converted in the 'unibuf'.
// Also the 'cbufp' is positioned on the last converted
// character.
// At the last conversion in the file, flush should be set to
// true so that we get all characters converted
//
// The converter must be flushed at the end of conversion so
// that characters on hold also will be written.
unibufp = unibuf;
cbufp = buf;
flush = rd != bufsz;
ucnv_toUnicode(convfrom, &unibufp, unibufp + bufsz, &cbufp,
cbufp + rd, fromoffsets, flush, &err);
infoffset += (uint32_t)(cbufp - buf);
if (U_FAILURE(err)) {
char pos[32];
sprintf(pos, "%u", infoffset - 1);
UnicodeString str(pos, (int32_t)(uprv_strlen(pos) + 1));
initMsg(pname);
u_wmsg(stderr, "problemCvtToU", str.getBuffer(), u_wmsg_errorName(err));
willexit = 1;
err = U_ZERO_ERROR; /* reset the error for the rest of the conversion. */
}
// At the last conversion, the converted characters should be
// equal to number of chars read.
if (flush && !willexit && cbufp != (buf + rd)) {
char pos[32];
sprintf(pos, "%u", infoffset);
UnicodeString str(pos, (int32_t)(uprv_strlen(pos) + 1));
initMsg(pname);
u_wmsg(stderr, "premEndInput", str.getBuffer());
willexit = 1;
}
// Prepare to transliterate and convert. Transliterate if needed.
#if !UCONFIG_NO_TRANSLITERATION
if (t) {
u.setTo(unibuf, (int32_t)(unibufp - unibuf)); // Copy into string.
t->transliterate(u);
} else
#endif
{
u.setTo(unibuf, (int32_t)(unibufp - unibuf), (int32_t)(bufsz)); // Share the buffer.
}
int32_t ulen = u.length();
// Convert the Unicode buffer into the destination codepage
// Again 'bufp' will be placed on the last converted character
// And 'unibufbp' will be placed on the last converted unicode character
// At the last conversion flush should be set to true to ensure that
// all characters left get converted
const UChar *unibufu = unibufbp = u.getBuffer();
do {
int32_t len = ulen > (int32_t)bufsz ? (int32_t)bufsz : ulen;
bufp = buf;
unibufp = (UChar *) (unibufbp + len);
ucnv_fromUnicode(convto, &bufp, bufp + tobufsz,
&unibufbp,
unibufp,
tooffsets, flush, &err);
if (U_FAILURE(err)) {
const char *errtag;
char pos[32];
uint32_t erroffset =
dataOffset((int32_t)(bufp - buf - 1), fromoffsets, (int32_t)(bufsz), tooffsets, (int32_t)(tobufsz));
int32_t ferroffset = (int32_t)(infoffset - (unibufp - unibufu) + erroffset);
if ((int32_t) ferroffset < 0) {
ferroffset = (int32_t)(outfoffset + (bufp - buf));
errtag = "problemCvtFromUOut";
} else {
errtag = "problemCvtFromU";
}
sprintf(pos, "%u", ferroffset);
UnicodeString str(pos, (int32_t)(uprv_strlen(pos) + 1));
initMsg(pname);
u_wmsg(stderr, errtag, str.getBuffer(),
u_wmsg_errorName(err));
willexit = 1;
}
// At the last conversion, the converted characters should be equal to number
// of consumed characters.
if (flush && !willexit && unibufbp != (unibufu + (size_t) (unibufp - unibufu))) {
char pos[32];
sprintf(pos, "%u", infoffset);
UnicodeString str(pos, (int32_t)(uprv_strlen(pos) + 1));
initMsg(pname);
u_wmsg(stderr, "premEnd", str.getBuffer());
willexit = 1;
}
// Finally, write the converted buffer to the output file
rd = (size_t) (bufp - buf);
outfoffset += (int32_t)(wr = fwrite(buf, 1, rd, outfile));
if (wr != rd) {
UnicodeString str(strerror(errno), "");
initMsg(pname);
u_wmsg(stderr, "cantWrite", str.getBuffer());
willexit = 1;
}
if (willexit) {
goto error_exit;
}
} while ((ulen -= (int32_t)(bufsz)) > 0);
} while (!flush); // Stop when we have flushed the
// converters (this means that it's
// the end of output)
goto normal_exit;
error_exit:
ret = FALSE;
normal_exit:
// Cleanup.
if (convfrom) ucnv_close(convfrom);
if (convto) ucnv_close(convto);
#if !UCONFIG_NO_TRANSLITERATION
if (t) delete t;
#endif
if (buf) delete[] buf;
if (unibuf) delete[] unibuf;
if (fromoffsets) delete[] fromoffsets;
if (tooffsets) delete[] tooffsets;
if (infile != stdin) {
fclose(infile);
}
return ret;
}
/**************************
* Will convert a sequence of bytes from one codepage to another.
* @param toConverterName: The name of the converter that will be used to encode the output buffer
* @param fromConverterName: The name of the converter that will be used to decode the input buffer
* @param target: Pointer to the output buffer* written
* @param targetLength: on input contains the capacity of target, on output the number of bytes copied to target
* @param source: Pointer to the input buffer
* @param sourceLength: on input contains the capacity of source, on output the number of bytes processed in "source"
* @param internal: used internally to store store state data across calls
* @param err: fills in an error status
*/
void
T_UConverter_fromCodepageToCodepage (UConverter * outConverter,
UConverter * inConverter,
char **target,
const char *targetLimit,
const char **source,
const char *sourceLimit,
int32_t* offsets,
int flush,
UErrorCode * err)
{
UChar out_chunk[CHUNK_SIZE];
const UChar *out_chunk_limit = out_chunk + CHUNK_SIZE;
UChar *out_chunk_alias;
UChar const *out_chunk_alias2;
if (U_FAILURE (*err)) return;
/*loops until the input buffer is completely consumed
*or if an error has be encountered
*first we convert from inConverter codepage to Unicode
*then from Unicode to outConverter codepage
*/
while ((*source != sourceLimit) && U_SUCCESS (*err))
{
out_chunk_alias = out_chunk;
ucnv_toUnicode (inConverter,
&out_chunk_alias,
out_chunk_limit,
source,
sourceLimit,
NULL,
flush,
err);
/*BUFFER_OVERFLOW_ERROR means that the output "CHUNK" is full
*we will require at least another loop (it's a recoverable error)
*/
if (U_SUCCESS (*err) || (*err == U_BUFFER_OVERFLOW_ERROR))
{
*err = U_ZERO_ERROR;
out_chunk_alias2 = out_chunk;
while ((out_chunk_alias2 != out_chunk_alias) && U_SUCCESS (*err))
{
ucnv_fromUnicode (outConverter,
target,
targetLimit,
&out_chunk_alias2,
out_chunk_alias,
NULL,
TRUE,
err);
}
}
else
break;
}
return;
}
int32_t ucnv_toUChars (const UConverter * converter,
UChar * target,
int32_t targetSize,
const char *source,
int32_t sourceSize,
UErrorCode * err)
{
const char *mySource = source;
const char *mySource_limit = source + sourceSize;
UConverter myConverter;
UChar *myTarget = target;
int32_t targetCapacity;
if (U_FAILURE (*err))
return 0;
if ((converter == NULL) || (targetSize < 0) || (sourceSize < 0))
{
*err = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
/*Means there is no work to be done */
if (sourceSize == 0)
{
/*for consistency we still need to
*store 0 in the targetCapacity
*if the user requires it
*/
if (targetSize >= 1)
{
target[0] = 0x0000;
return 1;
}
else
return 0;
}
/*makes a local copy of the UConverter */
myConverter = *converter;
/*Not in pure pre-flight mode */
if (targetSize > 0)
{
/* Changed from (targetSize * 2) to (targetSize) */
ucnv_toUnicode (&myConverter,
&myTarget,
target + (targetSize-1), /*Save a spot for the Null terminator */
&mySource,
mySource_limit,
NULL,
TRUE,
err);
/*Null terminates the string */
*(myTarget) = 0x0000;
}
/*Rigs targetCapacity to have at least one cell for zero termination */
/*Updates targetCapacity to contain the number of bytes written to target */
targetCapacity = 1;
targetCapacity += myTarget - target;
if (targetSize == 0)
{
*err = U_BUFFER_OVERFLOW_ERROR;
}
/* If the output buffer is exhausted, we need to stop writing
* to it but if the input buffer is not exhausted,
* we need to continue the conversion in order to store in targetSize
* the number of bytes that was required
*/
if (*err == U_BUFFER_OVERFLOW_ERROR)
{
UChar target2[CHUNK_SIZE];
UChar *target2_alias = target2;
const UChar *target2_limit = target2 + CHUNK_SIZE;
/*We use a stack allocated buffer around which we loop
(in case the output is greater than CHUNK_SIZE) */
while (*err == U_BUFFER_OVERFLOW_ERROR)
{
*err = U_ZERO_ERROR;
target2_alias = target2;
ucnv_toUnicode (&myConverter,
&target2_alias,
target2_limit,
&mySource,
mySource_limit,
NULL,
TRUE,
err);
/*updates the output parameter to contain the number of char required */
targetCapacity += target2_alias - target2 + 1;
}
(targetCapacity)--; /*adjust for last one */
if (U_SUCCESS (*err))
*err = U_BUFFER_OVERFLOW_ERROR;
}
return targetCapacity;
}
UErrorCode convsample_40()
{
printf("\n\n==============================================\n"
"Sample 40: C: convert data02.bin from cp37 to UTF16 [data40.utf16]\n");
FILE *f;
FILE *out;
int32_t count;
char inBuf[BUFFERSIZE];
const char *source;
const char *sourceLimit;
UChar *uBuf;
UChar *target;
UChar *targetLimit;
int32_t uBufSize = 0;
UConverter *conv = NULL;
UErrorCode status = U_ZERO_ERROR;
uint32_t inbytes=0, total=0;
f = fopen("data02.bin", "rb");
if(!f)
{
fprintf(stderr, "Couldn't open file 'data02.bin' (cp37 data file).\n");
return U_FILE_ACCESS_ERROR;
}
out = fopen("data40.utf16", "wb");
if(!out)
{
fprintf(stderr, "Couldn't create file 'data40.utf16'.\n");
fclose(f);
return U_FILE_ACCESS_ERROR;
}
// **************************** START SAMPLE *******************
conv = ucnv_openCCSID(37, UCNV_IBM, &status);
assert(U_SUCCESS(status));
uBufSize = (BUFFERSIZE/ucnv_getMinCharSize(conv));
printf("input bytes %d / min chars %d = %d UChars\n",
BUFFERSIZE, ucnv_getMinCharSize(conv), uBufSize);
uBuf = (UChar*)malloc(uBufSize * sizeof(UChar));
assert(uBuf!=NULL);
// grab another buffer's worth
while((!feof(f)) &&
((count=fread(inBuf, 1, BUFFERSIZE , f)) > 0) )
{
inbytes += count;
// Convert bytes to unicode
source = inBuf;
sourceLimit = inBuf + count;
do
{
target = uBuf;
targetLimit = uBuf + uBufSize;
ucnv_toUnicode( conv, &target, targetLimit,
&source, sourceLimit, NULL,
feof(f)?TRUE:FALSE, /* pass 'flush' when eof */
/* is true (when no more data will come) */
&status);
if(status == U_BUFFER_OVERFLOW_ERROR)
{
// simply ran out of space - we'll reset the target ptr the next
// time through the loop.
status = U_ZERO_ERROR;
}
else
{
// Check other errors here.
assert(U_SUCCESS(status));
// Break out of the loop (by force)
}
// Process the Unicode
// Todo: handle UTF-16/surrogates
assert(fwrite(uBuf, sizeof(uBuf[0]), (target-uBuf), out) ==
(size_t)(target-uBuf));
total += (target-uBuf);
} while (source < sourceLimit); // while simply out of space
}
printf("%d bytes in, %d UChars out.\n", inbytes, total);
// ***************************** END SAMPLE ********************
ucnv_close(conv);
fclose(f);
fclose(out);
printf("\n");
return U_ZERO_ERROR;
}
UXMLElement *
UXMLParser::parseFile(const char *filename, UErrorCode &errorCode) {
char bytes[4096], charsetBuffer[100];
FileStream *f;
const char *charset, *pb;
UnicodeString src;
UConverter *cnv;
UChar *buffer, *pu;
int32_t fileLength, bytesLength, length, capacity;
UBool flush;
if(U_FAILURE(errorCode)) {
return NULL;
}
f=T_FileStream_open(filename, "rb");
if(f==NULL) {
errorCode=U_FILE_ACCESS_ERROR;
return NULL;
}
bytesLength=T_FileStream_read(f, bytes, (int32_t)sizeof(bytes));
if(bytesLength<(int32_t)sizeof(bytes)) {
// we have already read the entire file
fileLength=bytesLength;
} else {
// get the file length
fileLength=T_FileStream_size(f);
}
/*
* get the charset:
* 1. Unicode signature
* 2. treat as ISO-8859-1 and read XML encoding="charser"
* 3. default to UTF-8
*/
charset=ucnv_detectUnicodeSignature(bytes, bytesLength, NULL, &errorCode);
if(U_SUCCESS(errorCode) && charset!=NULL) {
// open converter according to Unicode signature
cnv=ucnv_open(charset, &errorCode);
} else {
// read as Latin-1 and parse the XML declaration and encoding
cnv=ucnv_open("ISO-8859-1", &errorCode);
if(U_FAILURE(errorCode)) {
// unexpected error opening Latin-1 converter
goto exit;
}
buffer=src.getBuffer(bytesLength);
if(buffer==NULL) {
// unexpected failure to reserve some string capacity
errorCode=U_MEMORY_ALLOCATION_ERROR;
goto exit;
}
pb=bytes;
pu=buffer;
ucnv_toUnicode(
cnv,
&pu, buffer+src.getCapacity(),
&pb, bytes+bytesLength,
NULL, TRUE, &errorCode);
src.releaseBuffer(U_SUCCESS(errorCode) ? (int32_t)(pu-buffer) : 0);
ucnv_close(cnv);
cnv=NULL;
if(U_FAILURE(errorCode)) {
// unexpected error in conversion from Latin-1
src.remove();
goto exit;
}
// parse XML declaration
if(mXMLDecl.reset(src).lookingAt(0, errorCode)) {
int32_t declEnd=mXMLDecl.end(errorCode);
// go beyond <?xml
int32_t pos=src.indexOf((UChar)x_l)+1;
mAttrValue.reset(src);
while(pos<declEnd && mAttrValue.lookingAt(pos, errorCode)) { // loop runs once per attribute on this element.
UnicodeString attName = mAttrValue.group(1, errorCode);
UnicodeString attValue = mAttrValue.group(2, errorCode);
// Trim the quotes from the att value. These are left over from the original regex
// that parsed the attribue, which couldn't conveniently strip them.
attValue.remove(0,1); // one char from the beginning
attValue.truncate(attValue.length()-1); // and one from the end.
if(attName==UNICODE_STRING("encoding", 8)) {
length=attValue.extract(0, 0x7fffffff, charsetBuffer, (int32_t)sizeof(charsetBuffer));
charset=charsetBuffer;
break;
}
pos = mAttrValue.end(2, errorCode);
}
if(charset==NULL) {
// default to UTF-8
charset="UTF-8";
}
cnv=ucnv_open(charset, &errorCode);
}
}
if(U_FAILURE(errorCode)) {
// unable to open the converter
goto exit;
}
// convert the file contents
capacity=fileLength; // estimated capacity
src.getBuffer(capacity);
src.releaseBuffer(0); // zero length
flush=FALSE;
for(;;) {
// convert contents of bytes[bytesLength]
pb=bytes;
for(;;) {
length=src.length();
buffer=src.getBuffer(capacity);
if(buffer==NULL) {
// unexpected failure to reserve some string capacity
errorCode=U_MEMORY_ALLOCATION_ERROR;
goto exit;
}
pu=buffer+length;
ucnv_toUnicode(
cnv, &pu, buffer+src.getCapacity(),
&pb, bytes+bytesLength,
NULL, FALSE, &errorCode);
src.releaseBuffer(U_SUCCESS(errorCode) ? (int32_t)(pu-buffer) : 0);
if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
errorCode=U_ZERO_ERROR;
capacity=(3*src.getCapacity())/2; // increase capacity by 50%
} else {
break;
}
}
if(U_FAILURE(errorCode)) {
break; // conversion error
}
if(flush) {
break; // completely converted the file
}
// read next block
bytesLength=T_FileStream_read(f, bytes, (int32_t)sizeof(bytes));
if(bytesLength==0) {
// reached end of file, convert once more to flush the converter
flush=TRUE;
}
};
exit:
ucnv_close(cnv);
T_FileStream_close(f);
if(U_SUCCESS(errorCode)) {
return parse(src, errorCode);
} else {
return NULL;
}
}
void charsetFilteredOutputStream_icu::writeImpl
(const byte_t* const data, const size_t count)
{
if (m_from == NULL || m_to == NULL)
throw exceptions::charset_conv_error("Cannot initialize converters.");
// Allocate buffer for Unicode chars
const size_t uniSize = ucnv_getMinCharSize(m_from) * count * sizeof(UChar);
std::vector <UChar> uniBuffer(uniSize);
// Conversion loop
UErrorCode toErr = U_ZERO_ERROR;
const char* uniSource = reinterpret_cast <const char*>(data);
const char* uniSourceLimit = uniSource + count;
do
{
// Convert from source charset to Unicode
UChar* uniTarget = &uniBuffer[0];
UChar* uniTargetLimit = &uniBuffer[0] + uniSize;
toErr = U_ZERO_ERROR;
ucnv_toUnicode(m_from, &uniTarget, uniTargetLimit,
&uniSource, uniSourceLimit, NULL, /* flush */ FALSE, &toErr);
if (U_FAILURE(toErr) && toErr != U_BUFFER_OVERFLOW_ERROR)
{
if (toErr == U_INVALID_CHAR_FOUND ||
toErr == U_TRUNCATED_CHAR_FOUND ||
toErr == U_ILLEGAL_CHAR_FOUND)
{
throw exceptions::illegal_byte_sequence_for_charset();
}
else
{
throw exceptions::charset_conv_error
("[ICU] Error converting to Unicode from '" + m_sourceCharset.getName() + "'.");
}
}
const size_t uniLength = uniTarget - &uniBuffer[0];
// Allocate buffer for destination charset
const size_t cpSize = ucnv_getMinCharSize(m_to) * uniLength;
std::vector <char> cpBuffer(cpSize);
// Convert from Unicode to destination charset
UErrorCode fromErr = U_ZERO_ERROR;
const UChar* cpSource = &uniBuffer[0];
const UChar* cpSourceLimit = &uniBuffer[0] + uniLength;
do
{
char* cpTarget = &cpBuffer[0];
char* cpTargetLimit = &cpBuffer[0] + cpSize;
fromErr = U_ZERO_ERROR;
ucnv_fromUnicode(m_to, &cpTarget, cpTargetLimit,
&cpSource, cpSourceLimit, NULL, /* flush */ FALSE, &fromErr);
if (fromErr != U_BUFFER_OVERFLOW_ERROR && U_FAILURE(fromErr))
{
if (fromErr == U_INVALID_CHAR_FOUND ||
fromErr == U_TRUNCATED_CHAR_FOUND ||
fromErr == U_ILLEGAL_CHAR_FOUND)
{
throw exceptions::illegal_byte_sequence_for_charset();
}
else
{
throw exceptions::charset_conv_error
("[ICU] Error converting from Unicode to '" + m_destCharset.getName() + "'.");
}
}
const size_t cpLength = cpTarget - &cpBuffer[0];
// Write successfully converted bytes
m_stream.write(&cpBuffer[0], cpLength);
} while (fromErr == U_BUFFER_OVERFLOW_ERROR);
} while (toErr == U_BUFFER_OVERFLOW_ERROR);
}
static UBool testConvertToUnicode( const uint8_t *source, int sourcelen, const UChar *expect, int expectlen,
const char *codepage, UBool fallback, const int32_t *expectOffsets)
{
UErrorCode status = U_ZERO_ERROR;
UConverter *conv = 0;
UChar junkout[NEW_MAX_BUFFER]; /* FIX */
int32_t junokout[NEW_MAX_BUFFER]; /* FIX */
const char *src;
const char *realSourceEnd;
const char *srcLimit;
UChar *targ;
UChar *end;
int32_t *offs;
int i;
UBool checkOffsets = TRUE;
char junk[9999];
char offset_str[9999];
UChar *p;
UBool action;
int32_t realBufferSize;
UChar *realBufferEnd;
for(i=0;i<NEW_MAX_BUFFER;i++)
junkout[i] = 0xFFFE;
for(i=0;i<NEW_MAX_BUFFER;i++)
junokout[i] = -1;
setNuConvTestName(codepage, "TO");
log_verbose("\n========= %s\n", gNuConvTestName);
conv = my_ucnv_open(codepage, &status);
if(U_FAILURE(status))
{
log_data_err("Couldn't open converter %s\n",gNuConvTestName);
return TRUE; /* because it has been logged */
}
log_verbose("Converter opened..\n");
src = (const char *)source;
targ = junkout;
offs = junokout;
realBufferSize = (sizeof(junkout)/sizeof(junkout[0]));
realBufferEnd = junkout + realBufferSize;
realSourceEnd = src + sourcelen;
/*----setting the fallback routine----*/
ucnv_setFallback (conv, fallback);
action = ucnv_usesFallback(conv);
if(action != fallback){
log_err("FAIL: Error is setting fallback. Errocode=%s\n", myErrorName(status));
}
/*-------------------------------------*/
if ( gOutBufferSize != realBufferSize )
checkOffsets = FALSE;
if( gInBufferSize != NEW_MAX_BUFFER )
checkOffsets = FALSE;
do
{
end = nct_min( targ + gOutBufferSize, realBufferEnd);
srcLimit = nct_min(realSourceEnd, src + gInBufferSize);
if(targ == realBufferEnd)
{
log_err("Error, the end would overflow the real output buffer while about to call toUnicode! tarjey=%08lx %s",targ,gNuConvTestName);
return FALSE;
}
log_verbose("calling toUnicode @ %08lx to %08lx\n", targ,end);
status = U_ZERO_ERROR;
ucnv_toUnicode (conv,
&targ,
end,
(const char **)&src,
(const char *)srcLimit,
checkOffsets ? offs : NULL,
(UBool)(srcLimit == realSourceEnd), /* flush if we're at the end of hte source data */
&status);
} while ( (status == U_BUFFER_OVERFLOW_ERROR) || (srcLimit < realSourceEnd) ); /* while we just need another buffer */
if(U_FAILURE(status))
{
log_err("Problem doing toUnicode, errcode %s %s\n", myErrorName(status), gNuConvTestName);
return FALSE;
}
log_verbose("\nConversion done. %d bytes -> %d chars.\nResult :",
sourcelen, targ-junkout);
if(VERBOSITY)
{
junk[0] = 0;
offset_str[0] = 0;
for(p = junkout;p<targ;p++)
{
sprintf(junk + strlen(junk), "0x%04x, ", (0xFFFF) & (unsigned int)*p);
sprintf(offset_str + strlen(offset_str), "0x%04x, ", (0xFFFF) & (unsigned int)junokout[p-junkout]);
}
log_verbose(junk);
printUSeq(expect, expectlen);
if ( checkOffsets )
{
log_verbose("\nOffsets:");
log_verbose(offset_str);
}
log_verbose("\n");
}
ucnv_close(conv);
log_verbose("comparing %d uchars (%d bytes)..\n",expectlen,expectlen*2);
if (checkOffsets && (expectOffsets != 0))
{
if(memcmp(junokout,expectOffsets,(targ-junkout) * sizeof(int32_t)))
{
log_err("\n\ndid not get the expected offsets while %s \n", gNuConvTestName);
log_err("\nGot : ");
for(p=junkout;p<targ;p++)
log_err("%d, ", junokout[p-junkout]);
log_err("\nExpected: ");
for(i=0; i<(targ-junkout); i++)
log_err("%d,", expectOffsets[i]);
log_err("");
for(i=0; i<(targ-junkout); i++)
log_err("0x%04X,", junkout[i]);
log_err("");
for(i=0; i<(src-(const char *)source); i++)
log_err("0x%04X,", (unsigned char)source[i]);
}
}
if(!memcmp(junkout, expect, expectlen*2))
{
log_verbose("Matches!\n");
return TRUE;
}
else
{
log_err("String does not match. %s\n", gNuConvTestName);
log_verbose("String does not match. %s\n", gNuConvTestName);
printUSeqErr(junkout, expectlen);
printf("\n");
printUSeqErr(expect, expectlen);
return FALSE;
}
}
/* helper function */
static UChar*
_strFromWCS( UChar *dest,
int32_t destCapacity,
int32_t *pDestLength,
const wchar_t *src,
int32_t srcLength,
UErrorCode *pErrorCode)
{
int32_t retVal =0, count =0 ;
UConverter* conv = NULL;
UChar* pTarget = NULL;
UChar* pTargetLimit = NULL;
UChar* target = NULL;
UChar uStack [_STACK_BUFFER_CAPACITY];
wchar_t wStack[_STACK_BUFFER_CAPACITY];
wchar_t* pWStack = wStack;
char cStack[_STACK_BUFFER_CAPACITY];
int32_t cStackCap = _STACK_BUFFER_CAPACITY;
char* pCSrc=cStack;
char* pCSave=pCSrc;
char* pCSrcLimit=NULL;
const wchar_t* pSrc = src;
const wchar_t* pSrcLimit = NULL;
if(srcLength ==-1){
/* if the wchar_t source is null terminated we can safely
* assume that there are no embedded nulls, this is a fast
* path for null terminated strings.
*/
for(;;){
/* convert wchars to chars */
retVal = uprv_wcstombs(pCSrc,src, cStackCap);
if(retVal == -1){
*pErrorCode = U_ILLEGAL_CHAR_FOUND;
goto cleanup;
}else if(retVal >= (cStackCap-1)){
/* Should rarely occur */
u_growAnyBufferFromStatic(cStack,(void**)&pCSrc,&cStackCap,
cStackCap * _BUFFER_CAPACITY_MULTIPLIER, 0, sizeof(char));
pCSave = pCSrc;
}else{
/* converted every thing */
pCSrc = pCSrc+retVal;
break;
}
}
}else{
/* here the source is not null terminated
* so it may have nulls embeded and we need to
* do some extra processing
*/
int32_t remaining =cStackCap;
pSrcLimit = src + srcLength;
for(;;){
register int32_t nulLen = 0;
/* find nulls in the string */
while(nulLen<srcLength && pSrc[nulLen++]!=0){
}
if((pSrc+nulLen) < pSrcLimit){
/* check if we have enough room in pCSrc */
if(remaining < (nulLen * MB_CUR_MAX)){
/* should rarely occur */
int32_t len = (pCSrc-pCSave);
pCSrc = pCSave;
/* we do not have enough room so grow the buffer*/
u_growAnyBufferFromStatic(cStack,(void**)&pCSrc,&cStackCap,
_BUFFER_CAPACITY_MULTIPLIER*cStackCap+(nulLen*MB_CUR_MAX),len,sizeof(char));
pCSave = pCSrc;
pCSrc = pCSave+len;
remaining = cStackCap-(pCSrc - pCSave);
}
/* we have found a null so convert the
* chunk from begining of non-null char to null
*/
retVal = uprv_wcstombs(pCSrc,pSrc,remaining);
if(retVal==-1){
/* an error occurred bail out */
*pErrorCode = U_ILLEGAL_CHAR_FOUND;
goto cleanup;
}
pCSrc += retVal+1 /* already null terminated */;
pSrc += nulLen; /* skip past the null */
srcLength-=nulLen; /* decrement the srcLength */
remaining -= (pCSrc-pCSave);
}else{
/* the source is not null terminated and we are
* end of source so we copy the source to a temp buffer
* null terminate it and convert wchar_ts to chars
*/
if(nulLen >= _STACK_BUFFER_CAPACITY){
/* Should rarely occcur */
/* allocate new buffer buffer */
pWStack =(wchar_t*) uprv_malloc(sizeof(wchar_t) * (nulLen + 1));
if(pWStack==NULL){
*pErrorCode = U_MEMORY_ALLOCATION_ERROR;
goto cleanup;
}
}
if(nulLen>0){
/* copy the contents to tempStack */
uprv_memcpy(pWStack,pSrc,nulLen*sizeof(wchar_t));
}
/* null terminate the tempBuffer */
pWStack[nulLen] =0 ;
if(remaining < (nulLen * MB_CUR_MAX)){
/* Should rarely occur */
int32_t len = (pCSrc-pCSave);
pCSrc = pCSave;
/* we do not have enough room so grow the buffer*/
u_growAnyBufferFromStatic(cStack,(void**)&pCSrc,&cStackCap,
cStackCap+(nulLen*MB_CUR_MAX),len,sizeof(char));
pCSave = pCSrc;
pCSrc = pCSave+len;
remaining = cStackCap-(pCSrc - pCSave);
}
/* convert to chars */
retVal = uprv_wcstombs(pCSrc,pWStack,remaining);
pCSrc += retVal;
pSrc += nulLen;
srcLength-=nulLen; /* decrement the srcLength */
break;
}
}
}
/* OK..now we have converted from wchar_ts to chars now
* convert chars to UChars
*/
pCSrcLimit = pCSrc;
pCSrc = pCSave;
pTarget = target= dest;
pTargetLimit = dest + destCapacity;
conv= u_getDefaultConverter(pErrorCode);
if(U_FAILURE(*pErrorCode)|| conv==NULL){
goto cleanup;
}
for(;;) {
*pErrorCode = U_ZERO_ERROR;
/* convert to stack buffer*/
ucnv_toUnicode(conv,&pTarget,pTargetLimit,(const char**)&pCSrc,pCSrcLimit,NULL,(UBool)(pCSrc==pCSrcLimit),pErrorCode);
/* increment count to number written to stack */
count+= pTarget - target;
if(*pErrorCode==U_BUFFER_OVERFLOW_ERROR){
target = uStack;
pTarget = uStack;
pTargetLimit = uStack + _STACK_BUFFER_CAPACITY;
} else {
break;
}
}
if(pDestLength){
*pDestLength =count;
}
u_terminateUChars(dest,destCapacity,count,pErrorCode);
cleanup:
if(cStack != pCSave){
uprv_free(pCSave);
}
if(wStack != pWStack){
uprv_free(pWStack);
}
u_releaseDefaultConverter(conv);
return dest;
}
void charsetFilteredOutputStream_icu::flush()
{
if (m_from == NULL || m_to == NULL)
throw exceptions::charset_conv_error("Cannot initialize converters.");
// Allocate buffer for Unicode chars
const size_t uniSize = ucnv_getMinCharSize(m_from) * 1024 * sizeof(UChar);
std::vector <UChar> uniBuffer(uniSize);
// Conversion loop (with flushing)
UErrorCode toErr = U_ZERO_ERROR;
const char* uniSource = 0;
const char* uniSourceLimit = 0;
do
{
// Convert from source charset to Unicode
UChar* uniTarget = &uniBuffer[0];
UChar* uniTargetLimit = &uniBuffer[0] + uniSize;
toErr = U_ZERO_ERROR;
ucnv_toUnicode(m_from, &uniTarget, uniTargetLimit,
&uniSource, uniSourceLimit, NULL, /* flush */ TRUE, &toErr);
if (U_FAILURE(toErr) && toErr != U_BUFFER_OVERFLOW_ERROR)
{
throw exceptions::charset_conv_error
("[ICU] Error converting to Unicode from '" + m_sourceCharset.getName() + "'.");
}
const size_t uniLength = uniTarget - &uniBuffer[0];
// Allocate buffer for destination charset
const size_t cpSize = ucnv_getMinCharSize(m_to) * uniLength;
std::vector <char> cpBuffer(cpSize);
// Convert from Unicode to destination charset
UErrorCode fromErr = U_ZERO_ERROR;
const UChar* cpSource = &uniBuffer[0];
const UChar* cpSourceLimit = &uniBuffer[0] + uniLength;
do
{
char* cpTarget = &cpBuffer[0];
char* cpTargetLimit = &cpBuffer[0] + cpSize;
fromErr = U_ZERO_ERROR;
ucnv_fromUnicode(m_to, &cpTarget, cpTargetLimit,
&cpSource, cpSourceLimit, NULL, /* flush */ TRUE, &fromErr);
if (fromErr != U_BUFFER_OVERFLOW_ERROR && U_FAILURE(fromErr))
{
throw exceptions::charset_conv_error
("[ICU] Error converting from Unicode to '" + m_destCharset.getName() + "'.");
}
const size_t cpLength = cpTarget - &cpBuffer[0];
// Write successfully converted bytes
m_stream.write(&cpBuffer[0], cpLength);
} while (fromErr == U_BUFFER_OVERFLOW_ERROR);
} while (toErr == U_BUFFER_OVERFLOW_ERROR);
m_stream.flush();
}
CF_PRIVATE CFIndex __CFStringEncodingICUToUnicode(const char *icuName, uint32_t flags, const uint8_t *bytes, CFIndex numBytes, CFIndex *usedByteLen, UniChar *characters, CFIndex maxCharLen, CFIndex *usedCharLen) {
UConverter *converter;
UErrorCode errorCode = U_ZERO_ERROR;
const char *source = (const char *)bytes;
const char *sourceLimit = source + numBytes;
UTF16Char *destination = characters;
const UTF16Char *destinationLimit = destination + maxCharLen;
bool flush = ((0 == (flags & kCFStringEncodingPartialInput)) ? true : false);
CFIndex status;
if (NULL == (converter = __CFStringEncodingConverterCreateICUConverter(icuName, flags, true))) return kCFStringEncodingConverterUnavailable;
if (0 == maxCharLen) {
UTF16Char buffer[MAX_BUFFER_SIZE];
CFIndex totalLength = 0;
while ((source < sourceLimit) && (U_ZERO_ERROR == errorCode)) {
destination = buffer;
destinationLimit = destination + MAX_BUFFER_SIZE;
ucnv_toUnicode(converter, (UChar **)&destination, (const UChar *)destinationLimit, &source, sourceLimit, NULL, flush, &errorCode);
totalLength += (destination - buffer);
if (U_BUFFER_OVERFLOW_ERROR == errorCode) errorCode = U_ZERO_ERROR;
}
if (NULL != usedCharLen) *usedCharLen = totalLength;
} else {
ucnv_toUnicode(converter, (UChar **)&destination, (const UChar *)destinationLimit, &source, sourceLimit, NULL, flush, &errorCode);
if (NULL != usedCharLen) *usedCharLen = destination - characters;
}
status = ((U_ZERO_ERROR == errorCode) ? kCFStringEncodingConversionSuccess : ((U_BUFFER_OVERFLOW_ERROR == errorCode) ? kCFStringEncodingInsufficientOutputBufferLength : kCFStringEncodingInvalidInputStream));
if (NULL != usedByteLen) {
#if HAS_ICU_BUG_6024743
/* ICU has a serious behavioral inconsistency issue that the source pointer returned from ucnv_toUnicode() is after illegal input. We have to keep track of any changes in this area in order to prevent future binary compatiibility issues */
if (kCFStringEncodingInvalidInputStream == status) {
#define MAX_ERROR_BUFFER_LEN (32)
char errorBuffer[MAX_ERROR_BUFFER_LEN];
int8_t errorLength = MAX_ERROR_BUFFER_LEN;
#undef MAX_ERROR_BUFFER_LEN
errorCode = U_ZERO_ERROR;
ucnv_getInvalidChars(converter, errorBuffer, &errorLength, &errorCode);
if (U_ZERO_ERROR == errorCode) {
#if HAS_ICU_BUG_6025527
// Another ICU oddness here. ucnv_getInvalidUChars() writes the '\0' terminator, and errorLength includes the extra byte.
if ((errorLength > 0) && ('\0' == errorBuffer[errorLength - 1])) --errorLength;
#endif
source -= errorLength;
} else {
// Gah, something is terribly wrong. Reset everything
source = (const char *)bytes; // 0 length
if (NULL != usedCharLen) *usedCharLen = 0;
}
}
#endif
*usedByteLen = source - (const char *)bytes;
}
status |= __CFStringEncodingConverterReleaseICUConverter(converter, flags, status);
return status;
}
/* private function used for buffering input */
void
ufile_fill_uchar_buffer(UFILE *f)
{
UErrorCode status;
const char *mySource;
const char *mySourceEnd;
UChar *myTarget;
int32_t bufferSize;
int32_t maxCPBytes;
int32_t bytesRead;
int32_t availLength;
int32_t dataSize;
char charBuffer[UFILE_CHARBUFFER_SIZE];
u_localized_string *str;
if (f->fFile == NULL) {
/* There is nothing to do. It's a string. */
return;
}
str = &f->str;
dataSize = (int32_t)(str->fLimit - str->fPos);
if (f->fFileno == 0 && dataSize > 0) {
/* Don't read from stdin too many times. There is still some data. */
return;
}
/* shift the buffer if it isn't empty */
if(dataSize != 0) {
uprv_memmove(f->fUCBuffer, str->fPos, dataSize * sizeof(UChar));
}
/* record how much buffer space is available */
availLength = UFILE_UCHARBUFFER_SIZE - dataSize;
/* Determine the # of codepage bytes needed to fill our UChar buffer */
/* weiv: if converter is NULL, we use invariant converter with charwidth = 1)*/
maxCPBytes = availLength / (f->fConverter!=NULL?(2*ucnv_getMinCharSize(f->fConverter)):1);
/* Read in the data to convert */
if (f->fFileno == 0) {
/* Special case. Read from stdin one line at a time. */
char *retStr = fgets(charBuffer, ufmt_min(maxCPBytes, UFILE_CHARBUFFER_SIZE), f->fFile);
bytesRead = (int32_t)(retStr ? uprv_strlen(charBuffer) : 0);
}
else {
/* A normal file */
bytesRead = (int32_t)fread(charBuffer,
sizeof(char),
ufmt_min(maxCPBytes, UFILE_CHARBUFFER_SIZE),
f->fFile);
}
/* Set up conversion parameters */
status = U_ZERO_ERROR;
mySource = charBuffer;
mySourceEnd = charBuffer + bytesRead;
myTarget = f->fUCBuffer + dataSize;
bufferSize = UFILE_UCHARBUFFER_SIZE;
if(f->fConverter != NULL) { /* We have a valid converter */
/* Perform the conversion */
ucnv_toUnicode(f->fConverter,
&myTarget,
f->fUCBuffer + bufferSize,
&mySource,
mySourceEnd,
NULL,
(UBool)(feof(f->fFile) != 0),
&status);
} else { /*weiv: do the invariant conversion */
u_charsToUChars(mySource, myTarget, bytesRead);
myTarget += bytesRead;
}
/* update the pointers into our array */
str->fPos = str->fBuffer;
str->fLimit = myTarget;
}
UErrorCode convsample_05()
{
printf("\n\n==============================================\n"
"Sample 05: C: count the number of letters in a UTF-8 document\n");
FILE *f;
int32_t count;
char inBuf[BUFFERSIZE];
const char *source;
const char *sourceLimit;
UChar *uBuf;
UChar *target;
UChar *targetLimit;
UChar *p;
int32_t uBufSize = 0;
UConverter *conv;
UErrorCode status = U_ZERO_ERROR;
uint32_t letters=0, total=0;
f = fopen("data01.txt", "r");
if(!f)
{
fprintf(stderr, "Couldn't open file 'data01.txt' (UTF-8 data file).\n");
return U_FILE_ACCESS_ERROR;
}
// **************************** START SAMPLE *******************
conv = ucnv_open("utf-8", &status);
assert(U_SUCCESS(status));
uBufSize = (BUFFERSIZE/ucnv_getMinCharSize(conv));
printf("input bytes %d / min chars %d = %d UChars\n",
BUFFERSIZE, ucnv_getMinCharSize(conv), uBufSize);
uBuf = (UChar*)malloc(uBufSize * sizeof(UChar));
assert(uBuf!=NULL);
// grab another buffer's worth
while((!feof(f)) &&
((count=fread(inBuf, 1, BUFFERSIZE , f)) > 0) )
{
// Convert bytes to unicode
source = inBuf;
sourceLimit = inBuf + count;
do
{
target = uBuf;
targetLimit = uBuf + uBufSize;
ucnv_toUnicode(conv, &target, targetLimit,
&source, sourceLimit, NULL,
feof(f)?TRUE:FALSE, /* pass 'flush' when eof */
/* is true (when no more data will come) */
&status);
if(status == U_BUFFER_OVERFLOW_ERROR)
{
// simply ran out of space - we'll reset the target ptr the next
// time through the loop.
status = U_ZERO_ERROR;
}
else
{
// Check other errors here.
assert(U_SUCCESS(status));
// Break out of the loop (by force)
}
// Process the Unicode
// Todo: handle UTF-16/surrogates
for(p = uBuf; p<target; p++)
{
if(u_isalpha(*p))
letters++;
total++;
}
} while (source < sourceLimit); // while simply out of space
}
printf("%d letters out of %d total UChars.\n", letters, total);
// ***************************** END SAMPLE ********************
ucnv_close(conv);
printf("\n");
fclose(f);
return U_ZERO_ERROR;
}
/* private function used for buffering input */
void
ufile_fill_uchar_buffer(UFILE *f)
{
UErrorCode status;
const char *mySource;
const char *mySourceEnd;
UChar *myTarget;
int32_t bufferSize;
int32_t maxCPBytes;
int32_t bytesRead;
int32_t availLength;
int32_t dataSize;
/* shift the buffer if it isn't empty */
dataSize = (int32_t)(f->fUCLimit - f->fUCPos);
if(dataSize != 0) {
memmove(f->fUCBuffer,
f->fUCPos,
dataSize * sizeof(UChar));
}
/* record how much buffer space is available */
availLength = UFILE_UCHARBUFFER_SIZE - dataSize;
/* Determine the # of codepage bytes needed to fill our UChar buffer */
/* weiv: if converter is NULL, we use invariant converter with charwidth = 1)*/
maxCPBytes = availLength / (f->fConverter!=NULL?(2*ucnv_getMinCharSize(f->fConverter)):1);
/* Read in the data to convert */
bytesRead = (int32_t)fread(f->fCharBuffer,
sizeof(char),
ufmt_min(maxCPBytes, UFILE_CHARBUFFER_SIZE),
f->fFile);
/* Set up conversion parameters */
status = U_ZERO_ERROR;
mySource = f->fCharBuffer;
mySourceEnd = f->fCharBuffer + bytesRead;
myTarget = f->fUCBuffer + dataSize;
bufferSize = UFILE_UCHARBUFFER_SIZE;
if(f->fConverter != NULL) { /* We have a valid converter */
/* Perform the conversion */
ucnv_toUnicode(f->fConverter,
&myTarget,
f->fUCBuffer + bufferSize,
&mySource,
mySourceEnd,
NULL,
(UBool)(feof(f->fFile) != 0),
&status);
} else { /*weiv: do the invariant conversion */
u_charsToUChars(mySource, myTarget, bytesRead);
myTarget += bytesRead;
}
/* update the pointers into our array */
f->fUCPos = f->fUCBuffer;
f->fUCLimit = myTarget;
}
/* fill the uchar buffer */
static UCHARBUF*
ucbuf_fillucbuf( UCHARBUF* buf,UErrorCode* error){
UChar* pTarget=NULL;
UChar* target=NULL;
const char* source=NULL;
char carr[MAX_IN_BUF] = {'\0'};
char* cbuf = carr;
int32_t inputRead=0;
int32_t outputWritten=0;
int32_t offset=0;
const char* sourceLimit =NULL;
int32_t cbufSize=0;
pTarget = buf->buffer;
/* check if we arrived here without exhausting the buffer*/
if(buf->currentPos<buf->bufLimit){
offset = (int32_t)(buf->bufLimit-buf->currentPos);
memmove(buf->buffer,buf->currentPos,offset* sizeof(UChar));
}
#if DEBUG
memset(pTarget+offset,0xff,sizeof(UChar)*(MAX_IN_BUF-offset));
#endif
if(buf->isBuffered){
cbufSize = MAX_IN_BUF;
/* read the file */
inputRead=T_FileStream_read(buf->in,cbuf,cbufSize-offset);
buf->remaining-=inputRead;
}else{
cbufSize = T_FileStream_size(buf->in);
cbuf = (char*)uprv_malloc(cbufSize);
if (cbuf == NULL) {
*error = U_MEMORY_ALLOCATION_ERROR;
return NULL;
}
inputRead= T_FileStream_read(buf->in,cbuf,cbufSize);
buf->remaining-=inputRead;
}
/* just to be sure...*/
if ( 0 == inputRead )
buf->remaining = 0;
target=pTarget;
/* convert the bytes */
if(buf->conv){
/* set the callback to stop */
UConverterToUCallback toUOldAction ;
void* toUOldContext;
void* toUNewContext=NULL;
ucnv_setToUCallBack(buf->conv,
UCNV_TO_U_CALLBACK_STOP,
toUNewContext,
&toUOldAction,
(const void**)&toUOldContext,
error);
/* since state is saved in the converter we add offset to source*/
target = pTarget+offset;
source = cbuf;
sourceLimit = source + inputRead;
ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
&source,sourceLimit,NULL,
(UBool)(buf->remaining==0),error);
if(U_FAILURE(*error)){
char context[CONTEXT_LEN+1];
char preContext[CONTEXT_LEN+1];
char postContext[CONTEXT_LEN+1];
int8_t len = CONTEXT_LEN;
int32_t start=0;
int32_t stop =0;
int32_t pos =0;
/* use erro1 to preserve the error code */
UErrorCode error1 =U_ZERO_ERROR;
if( buf->showWarning==TRUE){
fprintf(stderr,"\n###WARNING: Encountered abnormal bytes while"
" converting input stream to target encoding: %s\n",
u_errorName(*error));
}
/* now get the context chars */
ucnv_getInvalidChars(buf->conv,context,&len,&error1);
context[len]= 0 ; /* null terminate the buffer */
pos = (int32_t)(source - cbuf - len);
/* for pre-context */
start = (pos <=CONTEXT_LEN)? 0 : (pos - (CONTEXT_LEN-1));
stop = pos-len;
memcpy(preContext,cbuf+start,stop-start);
/* null terminate the buffer */
preContext[stop-start] = 0;
/* for post-context */
start = pos+len;
stop = (int32_t)(((pos+CONTEXT_LEN)<= (sourceLimit-cbuf) )? (pos+(CONTEXT_LEN-1)) : (sourceLimit-cbuf));
memcpy(postContext,source,stop-start);
/* null terminate the buffer */
postContext[stop-start] = 0;
if(buf->showWarning ==TRUE){
/* print out the context */
fprintf(stderr,"\tPre-context: %s\n",preContext);
fprintf(stderr,"\tContext: %s\n",context);
fprintf(stderr,"\tPost-context: %s\n", postContext);
}
/* reset the converter */
ucnv_reset(buf->conv);
/* set the call back to substitute
* and restart conversion
*/
ucnv_setToUCallBack(buf->conv,
UCNV_TO_U_CALLBACK_SUBSTITUTE,
toUNewContext,
&toUOldAction,
(const void**)&toUOldContext,
&error1);
/* reset source and target start positions */
target = pTarget+offset;
source = cbuf;
/* re convert */
ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
&source,sourceLimit,NULL,
(UBool)(buf->remaining==0),&error1);
}
outputWritten = (int32_t)(target - pTarget);
#if DEBUG
{
int i;
target = pTarget;
for(i=0;i<numRead;i++){
/* printf("%c", (char)(*target++));*/
}
}
#endif
}else{
u_charsToUChars(cbuf,target+offset,inputRead);
outputWritten=((buf->remaining>cbufSize)? cbufSize:inputRead+offset);
}
buf->currentPos = pTarget;
buf->bufLimit=pTarget+outputWritten;
*buf->bufLimit=0; /*NUL terminate*/
if(cbuf!=carr){
uprv_free(cbuf);
}
return buf;
}
