librevenge::RVNGString libvisio::VSDMetaData::readCodePageString(librevenge::RVNGInputStream *input)
{
uint32_t size = readU32(input);
if (size > getRemainingLength(input))
size = getRemainingLength(input);
if (size == 0)
return librevenge::RVNGString();
std::vector<unsigned char> characters;
for (uint32_t i = 0; i < size; ++i)
characters.push_back(readU8(input));
uint32_t codepage = getCodePage();
librevenge::RVNGString string;
if (codepage == 65001)
{
// http://msdn.microsoft.com/en-us/library/windows/desktop/dd374130%28v=vs.85%29.aspx
// says this is UTF-8.
characters.push_back(0);
string.append(reinterpret_cast<const char *>(characters.data()));
}
else
{
UErrorCode status = U_ZERO_ERROR;
UConverter *conv = nullptr;
switch (codepage)
{
case 1252:
// http://msdn.microsoft.com/en-us/goglobal/bb964654
conv = ucnv_open("windows-1252", &status);
break;
}
if (U_SUCCESS(status) && conv)
{
assert(!characters.empty());
const auto *src = (const char *)&characters[0];
const char *srcLimit = (const char *)src + characters.size();
while (src < srcLimit)
{
UChar32 ucs4Character = ucnv_getNextUChar(conv, &src, srcLimit, &status);
if (U_SUCCESS(status) && U_IS_UNICODE_CHAR(ucs4Character))
appendUCS4(string, ucs4Character);
}
}
if (conv)
ucnv_close(conv);
}
return string;
}
static void U_CALLCONV
nameAliasesLineFn(void *context,
char *fields[][2], int32_t fieldCount,
UErrorCode *pErrorCode) {
char *name;
int16_t length=0;
static uint32_t prevCode=0;
uint32_t code=0;
if(U_FAILURE(*pErrorCode)) {
return;
}
/* get the character code */
code=uprv_strtoul(fields[0][0], NULL, 16);
/* get the character name */
name=fields[1][0];
length=getName(&name, fields[1][1]);
if(length==0 || length>=sizeof(cpNameAliases[cpNameAliasesTop].nameAlias)) {
fprintf(stderr, "gennames: error - name alias %s empty or too long for code point U+%04lx\n",
name, (unsigned long)code);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
/* check for non-character code points */
if(!U_IS_UNICODE_CHAR(code)) {
fprintf(stderr, "gennames: error - name alias for non-character code point U+%04lx\n",
(unsigned long)code);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
/* check that the code points (code) are in ascending order */
if(code<=prevCode && code>0) {
fprintf(stderr, "gennames: error - NameAliases entries out of order, U+%04lx after U+%04lx\n",
(unsigned long)code, (unsigned long)prevCode);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
prevCode=code;
if(cpNameAliasesTop>=LENGTHOF(cpNameAliases)) {
fprintf(stderr, "gennames: error - too many name aliases\n");
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
cpNameAliases[cpNameAliasesTop].code=code;
uprv_memcpy(cpNameAliases[cpNameAliasesTop].nameAlias, name, length);
cpNameAliases[cpNameAliasesTop].nameAlias[length]=0;
++cpNameAliasesTop;
parseName(name, length);
}
bool readUTFChar(const char* str, int* begin, int length, unsigned* codePointOut)
{
int codePoint; // Avoids warning when U8_NEXT writes -1 to it.
U8_NEXT(str, *begin, length, codePoint);
*codePointOut = static_cast<unsigned>(codePoint);
// The ICU macro above moves to the next char, we want to point to the last
// char consumed.
(*begin)--;
// Validate the decoded value.
if (U_IS_UNICODE_CHAR(codePoint))
return true;
*codePointOut = kUnicodeReplacementCharacter;
return false;
}
bool readUTFChar(const UChar* str, int* begin, int length, unsigned* codePoint)
{
if (U16_IS_SURROGATE(str[*begin])) {
if (!U16_IS_SURROGATE_LEAD(str[*begin]) || *begin + 1 >= length || !U16_IS_TRAIL(str[*begin + 1])) {
// Invalid surrogate pair.
*codePoint = kUnicodeReplacementCharacter;
return false;
}
// Valid surrogate pair.
*codePoint = U16_GET_SUPPLEMENTARY(str[*begin], str[*begin + 1]);
(*begin)++;
} else {
// Not a surrogate, just one 16-bit word.
*codePoint = str[*begin];
}
if (U_IS_UNICODE_CHAR(*codePoint))
return true;
// Invalid code point.
*codePoint = kUnicodeReplacementCharacter;
return false;
}
static void U_CALLCONV
lineFn(void *context,
char *fields[][2], int32_t fieldCount,
UErrorCode *pErrorCode) {
Options *storeOptions=(Options *)context;
char *names[4];
int16_t lengths[4]={ 0, 0, 0, 0 };
static uint32_t prevCode=0;
uint32_t code=0;
if(U_FAILURE(*pErrorCode)) {
return;
}
/* get the character code */
code=uprv_strtoul(fields[0][0], NULL, 16);
/* get the character name */
if(storeOptions->storeNames) {
names[0]=fields[1][0];
lengths[0]=getName(names+0, fields[1][1]);
if(names[0][0]=='<') {
/* do not store pseudo-names in <> brackets */
lengths[0]=0;
}
}
/* store 1.0 names */
/* get the second character name, the one from Unicode 1.0 */
if(storeOptions->store10Names) {
names[1]=fields[10][0];
lengths[1]=getName(names+1, fields[10][1]);
if(names[1][0]=='<') {
/* do not store pseudo-names in <> brackets */
lengths[1]=0;
}
}
/* get the ISO 10646 comment */
if(storeOptions->storeISOComments) {
names[2]=fields[11][0];
lengths[2]=getName(names+2, fields[11][1]);
}
if(lengths[0]+lengths[1]+lengths[2]==0) {
return;
}
/* check for non-character code points */
if(!U_IS_UNICODE_CHAR(code)) {
fprintf(stderr, "gennames: error - properties for non-character code point U+%04lx\n",
(unsigned long)code);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
/* check that the code points (code) are in ascending order */
if(code<=prevCode && code>0) {
fprintf(stderr, "gennames: error - UnicodeData entries out of order, U+%04lx after U+%04lx\n",
(unsigned long)code, (unsigned long)prevCode);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
prevCode=code;
parseName(names[0], lengths[0]);
parseName(names[1], lengths[1]);
parseName(names[2], lengths[2]);
if(cpNameAliasesIndex<cpNameAliasesTop && code>=cpNameAliases[cpNameAliasesIndex].code) {
if(code==cpNameAliases[cpNameAliasesIndex].code) {
names[3]=cpNameAliases[cpNameAliasesIndex].nameAlias;
lengths[3]=(int16_t)uprv_strlen(cpNameAliases[cpNameAliasesIndex].nameAlias);
++cpNameAliasesIndex;
} else {
fprintf(stderr, "gennames: error - NameAlias but no UnicodeData entry for U+%04lx\n",
(unsigned long)code);
*pErrorCode=U_PARSE_ERROR;
exit(U_PARSE_ERROR);
}
}
/*
* set the count argument to
* 1: only store regular names, or only store ISO 10646 comments
* 2: store regular and 1.0 names
* 3: store names and ISO 10646 comment
* 4: also store name alias
*
* addLine() will ignore empty trailing names
*/
if(storeOptions->storeNames) {
/* store names and comments as parsed according to storeOptions */
addLine(code, names, lengths, LENGTHOF(names));
} else {
/* store only ISO 10646 comments */
addLine(code, names+2, lengths+2, 1);
}
}
