void CFX_SAXReader::ParseTagAttributeValue() {
if (m_SkipChar) {
if (m_SkipChar == m_CurByte) {
{
m_iDataLength = m_iDataPos;
m_iDataPos = 0;
if (m_pHandler) {
NotifyAttribute();
}
}
m_SkipChar = 0;
m_eMode = FX_SAXMODE_TagAttributeName;
return;
}
ParseChar(m_CurByte);
return;
}
if (m_CurByte < 0x21) {
return;
}
if (m_iDataPos < 1) {
if (m_CurByte == '\'' || m_CurByte == '\"') {
m_SkipChar = m_CurByte;
}
}
}
/*
* Returns an encoded string according to the
* Soundex algorithm.
*/
string ParseName(string name) {
string code;
name = ConvertToUpperCase(name);
for (int i = 0; i < name.length(); i++) {
if (i == 0)
code += name[i]; // save first char
if (i > 0 && isalpha(name[i]) != 0)
code += ParseChar(name[i]);
}
for (int j = 0; j < code.length(); j++) {
if (j > 0 && code[j] == code[j-1]) {
code.erase(j,1); // delete duplicates
} else if (code[j] == '0') {
code.erase(j,1); // delete zeros
}
}
if (code.length() < CODE_LENGTH) {
int zeros = CODE_LENGTH - code.length();
for (int k = 0; k < zeros; k++) {
code += '0'; // padd zeros if code is too short
}
} else if (code.length() > CODE_LENGTH) {
code = code.substr(0,CODE_LENGTH); // truncate if code is too long
}
return code;
}
// <local-name> := Z <(function) encoding> E <(entity) name>
// [<discriminator>]
// := Z <(function) encoding> E s [<discriminator>]
static bool ParseLocalName(State *state) {
State copy = *state;
if (ParseChar(state, 'Z') && ParseEncoding(state) &&
ParseChar(state, 'E') && MaybeAppend(state, "::") &&
ParseName(state) && Optional(ParseDiscriminator(state))) {
return true;
}
*state = copy;
if (ParseChar(state, 'Z') && ParseEncoding(state) &&
ParseTwoChar(state, "Es") && Optional(ParseDiscriminator(state))) {
return true;
}
*state = copy;
return false;
}
static void CharConst (void)
/* Parse a character constant. */
{
int C;
/* Skip the quote */
NextChar ();
/* Get character */
C = ParseChar ();
/* Check for closing quote */
if (CurC != '\'') {
Error ("`\'' expected");
} else {
/* Skip the quote */
NextChar ();
}
/* Setup values and attributes */
NextTok.Tok = TOK_CCONST;
/* Translate into target charset */
NextTok.IVal = SignExtendChar (TgtTranslateChar (C));
/* Character constants have type int */
NextTok.Type = type_int;
}
// <call-offset> ::= h <nv-offset> _
// ::= v <v-offset> _
static bool ParseCallOffset(State *state) {
State copy = *state;
if (ParseChar(state, 'h') &&
ParseNVOffset(state) && ParseChar(state, '_')) {
return true;
}
*state = copy;
if (ParseChar(state, 'v') &&
ParseVOffset(state) && ParseChar(state, '_')) {
return true;
}
*state = copy;
return false;
}
static int ParseFont(Kanji_Font* font, FILE* fp) {
char buf[BUF], *p;
int index;
int k_rshift, a_rshift;
int s;
for (s = 8; s < font->k_size; s += 8) {}
k_rshift = s - font->k_size;
for (s = 8; s < font->a_size; s += 8) {}
a_rshift = s - font->a_size;
while (1) {
if (fgets(buf, BUF, fp) == NULL) {
break;
}
if (strstr(buf, "ENCODING") != NULL) {
p = strchr(buf, ' ');
index = strtol(p, 0, 10);
while (strstr(buf, "BITMAP") == NULL) {
fgets(buf, BUF, fp);
}
if (index > 255) {
index = (((index & 0xff00) >> 8) - 0x20) * 96
+ (index & 0xff) - 0x20 + 0xff;
ParseChar(font, index, fp, k_rshift);
}
else {
// <discriminator> := _ <(non-negative) number>
static bool ParseDiscriminator(State *state) {
State copy = *state;
if (ParseChar(state, '_') && ParseNumber(state)) {
return true;
}
*state = copy;
return false;
}
// <expr-primary> ::= L <type> <(value) number> E
// ::= L <type> <(value) float> E
// ::= L <mangled-name> E
// // A bug in g++'s C++ ABI version 2 (-fabi-version=2).
// ::= LZ <encoding> E
static bool ParseExprPrimary(State *state) {
State copy = *state;
if (ParseChar(state, 'L') && ParseType(state) &&
ParseNumber(state) &&
ParseChar(state, 'E')) {
return true;
}
*state = copy;
if (ParseChar(state, 'L') && ParseType(state) &&
ParseFloatNumber(state) &&
ParseChar(state, 'E')) {
return true;
}
*state = copy;
if (ParseChar(state, 'L') && ParseMangledName(state) &&
ParseChar(state, 'E')) {
return true;
}
*state = copy;
if (ParseTwoChar(state, "LZ") && ParseEncoding(state) &&
ParseChar(state, 'E')) {
return true;
}
*state = copy;
return false;
}
// <local-source-name> ::= L <source-name> [<discriminator>]
//
// References:
// http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31775
// http://gcc.gnu.org/viewcvs?view=rev&revision=124467
static bool ParseLocalSourceName(State *state) {
State copy = *state;
if (ParseChar(state, 'L') && ParseSourceName(state) &&
Optional(ParseDiscriminator(state))) {
return true;
}
*state = copy;
return false;
}
// <v-offset> ::= <(offset) number> _ <(virtual offset) number>
static bool ParseVOffset(State *state) {
State copy = *state;
if (ParseNumber(state) && ParseChar(state, '_') &&
ParseNumber(state)) {
return true;
}
*state = copy;
return false;
}
// <pointer-to-member-type> ::= M <(class) type> <(member) type>
static bool ParsePointerToMemberType(State *state) {
State copy = *state;
if (ParseChar(state, 'M') && ParseType(state) &&
ParseType(state)) {
return true;
}
*state = copy;
return false;
}
static void StringConst (void)
/* Parse a quoted string */
{
/* String buffer */
StrBuf S = AUTO_STRBUF_INITIALIZER;
/* Assume next token is a string constant */
NextTok.Tok = TOK_SCONST;
/* Concatenate strings. If at least one of the concenated strings is a wide
** character literal, the whole string is a wide char literal, otherwise
** it's a normal string literal.
*/
while (1) {
/* Check if this is a normal or a wide char string */
if (CurC == 'L' && NextC == '\"') {
/* Wide character literal */
NextTok.Tok = TOK_WCSCONST;
NextChar ();
NextChar ();
} else if (CurC == '\"') {
/* Skip the quote char */
NextChar ();
} else {
/* No string */
break;
}
/* Read until end of string */
while (CurC != '\"') {
if (CurC == '\0') {
Error ("Unexpected newline");
break;
}
SB_AppendChar (&S, ParseChar ());
}
/* Skip closing quote char if there was one */
NextChar ();
/* Skip white space, read new input */
SkipWhite ();
}
/* Terminate the string */
SB_AppendChar (&S, '\0');
/* Add the whole string to the literal pool */
NextTok.SVal = AddLiteralStr (&S);
/* Free the buffer */
SB_Done (&S);
}
// <ctor-dtor-name> ::= C1 | C2 | C3
// ::= D0 | D1 | D2
static bool ParseCtorDtorName(State *state) {
State copy = *state;
if (ParseChar(state, 'C') &&
ParseCharClass(state, "123")) {
const char * const prev_name = state->prev_name;
const int prev_name_length = state->prev_name_length;
MaybeAppendWithLength(state, prev_name, prev_name_length);
return true;
}
*state = copy;
if (ParseChar(state, 'D') &&
ParseCharClass(state, "012")) {
const char * const prev_name = state->prev_name;
const int prev_name_length = state->prev_name_length;
MaybeAppend(state, "~");
MaybeAppendWithLength(state, prev_name, prev_name_length);
return true;
}
*state = copy;
return false;
}
int SB_GetString (StrBuf* B, StrBuf* S)
/* Get a string from the string buffer. Returns 1 if a string was found and 0
** otherwise. Errors are only output in case of invalid strings (missing end
** of string).
*/
{
char C;
/* Clear S */
SB_Clear (S);
/* A string starts with quote marks */
if (SB_Peek (B) == '\"') {
/* String follows, be sure to concatenate strings */
while (SB_Peek (B) == '\"') {
/* Skip the quote char */
SB_Skip (B);
/* Read the actual string contents */
while ((C = SB_Peek (B)) != '\"') {
if (C == '\0') {
Error ("Unexpected end of string");
break;
}
SB_AppendChar (S, ParseChar (B));
}
/* Skip the closing quote char if there was one */
SB_Skip (B);
/* Skip white space, read new input */
SB_SkipWhite (B);
}
/* Terminate the string */
SB_Terminate (S);
/* Success */
return 1;
} else {
/* Not a string */
SB_Terminate (S);
return 0;
}
}
// <builtin-type> ::= v, etc.
// ::= u <source-name>
static bool ParseBuiltinType(State *state) {
const AbbrevPair *p;
for (p = kBuiltinTypeList; p->abbrev != nullptr; ++p) {
if (state->mangled_cur[0] == p->abbrev[0]) {
MaybeAppend(state, p->real_name);
++state->mangled_cur;
return true;
}
}
State copy = *state;
if (ParseChar(state, 'u') && ParseSourceName(state)) {
return true;
}
*state = copy;
return false;
}
/*------------------------------------------------------------------------------
* function: ParseWhiteSpace
* Parse whitespace; return false if can't parse.
*/
static Boolean
ParseWhiteSpace(int numArgs, char **args)
{
if (numArgs >= MAX_SEPARATORS) {
fprintf(stderr, "ParseCat: too many separators, max = %d\n",
MAX_SEPARATORS);
return false;
}
for(int i = 1; i < numArgs; i++) {
if (!ParseChar(args[i], whitespaces[i - 1]))
return false;
}
numWhitespace = numArgs - 1;
return true;
}
void C4RTFFile::ParseHexChar(StdStrBuf &sResult, char c)
{
pState->bHex = pState->bHex << 4;
if (isdigit((unsigned char)c))
pState->bHex += c - '0';
else if (Inside<char>(c, 'a', 'f'))
pState->bHex += c - 'a' + 10;
else if (Inside<char>(c, 'A', 'F'))
pState->bHex += c - 'A' + 10;
else
throw new ParserError("Invalid hex character");
if (!--pState->iHexBinCnt)
{
pState->eState = psNormal;
ParseChar(sResult, pState->bHex);
}
}
void CFX_SAXReader::ParseTagEnd() {
if (m_CurByte < 0x21) {
return;
}
if (m_CurByte == '>') {
Pop();
m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex;
m_iDataLength = m_iDataPos;
m_iDataPos = 0;
if (m_pHandler) {
NotifyEnd();
}
Pop();
m_eMode = FX_SAXMODE_Text;
} else {
ParseChar(m_CurByte);
}
}
// <operator-name> ::= nw, and other two letters cases
// ::= cv <type> # (cast)
// ::= v <digit> <source-name> # vendor extended operator
static bool ParseOperatorName(State *state) {
if (RemainingLength(state) < 2) {
return false;
}
// First check with "cv" (cast) case.
State copy = *state;
if (ParseTwoChar(state, "cv") &&
MaybeAppend(state, "operator ") &&
EnterNestedName(state) &&
ParseType(state) &&
LeaveNestedName(state, copy.nest_level)) {
return true;
}
*state = copy;
// Then vendor extended operators.
if (ParseChar(state, 'v') && ParseCharClass(state, "0123456789") &&
ParseSourceName(state)) {
return true;
}
*state = copy;
// Other operator names should start with a lower alphabet followed
// by a lower/upper alphabet.
if (!(IsLower(state->mangled_cur[0]) &&
IsAlpha(state->mangled_cur[1]))) {
return false;
}
// We may want to perform a binary search if we really need speed.
const AbbrevPair *p;
for (p = kOperatorList; p->abbrev != nullptr; ++p) {
if (state->mangled_cur[0] == p->abbrev[0] &&
state->mangled_cur[1] == p->abbrev[1]) {
MaybeAppend(state, "operator");
if (IsLower(*p->real_name)) { // new, delete, etc.
MaybeAppend(state, " ");
}
MaybeAppend(state, p->real_name);
state->mangled_cur += 2;
return true;
}
}
return false;
}
void CFX_SAXReader::ParseText() {
if (m_CurByte == '<') {
if (m_iDataPos > 0) {
m_iDataLength = m_iDataPos;
m_iDataPos = 0;
if (m_pHandler) {
NotifyData();
}
}
Push();
m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex;
m_eMode = FX_SAXMODE_NodeStart;
return;
}
if (m_iDataPos < 1 && SkipSpace(m_CurByte)) {
return;
}
ParseChar(m_CurByte);
}
// <type> ::= <CV-qualifiers> <type>
// ::= P <type>
// ::= R <type>
// ::= C <type>
// ::= G <type>
// ::= U <source-name> <type>
// ::= <builtin-type>
// ::= <function-type>
// ::= <class-enum-type>
// ::= <array-type>
// ::= <pointer-to-member-type>
// ::= <template-template-param> <template-args>
// ::= <template-param>
// ::= <substitution>
static bool ParseType(State *state) {
// We should check CV-qualifers, and PRGC things first.
State copy = *state;
if (ParseCVQualifiers(state) && ParseType(state)) {
return true;
}
*state = copy;
if (ParseCharClass(state, "PRCG") && ParseType(state)) {
return true;
}
*state = copy;
if (ParseChar(state, 'U') && ParseSourceName(state) &&
ParseType(state)) {
return true;
}
*state = copy;
if (ParseBuiltinType(state) ||
ParseFunctionType(state) ||
ParseClassEnumType(state) ||
ParseArrayType(state) ||
ParsePointerToMemberType(state) ||
ParseSubstitution(state)) {
return true;
}
if (ParseTemplateTemplateParam(state) &&
ParseTemplateArgs(state)) {
return true;
}
*state = copy;
// Less greedy than <template-template-param> <template-args>.
if (ParseTemplateParam(state)) {
return true;
}
return false;
}
// <number> ::= [n] <non-negative decimal integer>
static bool ParseNumber(State *state) {
int sign = 1;
if (ParseChar(state, 'n')) {
sign = -1;
}
const char *p = state->mangled_cur;
int number = 0;
for (;p < state->mangled_end; ++p) {
if ((*p >= '0' && *p <= '9')) {
number = number * 10 + (*p - '0');
} else {
break;
}
}
if (p != state->mangled_cur) { // Conversion succeeded.
state->mangled_cur = p;
state->number = number * sign;
return true;
}
return false;
}
void CFX_SAXReader::SkipNode() {
int32_t iLen = m_SkipStack.GetSize();
if (m_SkipChar == '\'' || m_SkipChar == '\"') {
if (m_CurByte != m_SkipChar) {
return;
}
iLen--;
FXSYS_assert(iLen > -1);
m_SkipStack.RemoveAt(iLen, 1);
m_SkipChar = iLen ? m_SkipStack[iLen - 1] : 0;
return;
}
switch (m_CurByte) {
case '<':
m_SkipChar = '>';
m_SkipStack.Add('>');
break;
case '[':
m_SkipChar = ']';
m_SkipStack.Add(']');
break;
case '(':
m_SkipChar = ')';
m_SkipStack.Add(')');
break;
case '\'':
m_SkipChar = '\'';
m_SkipStack.Add('\'');
break;
case '\"':
m_SkipChar = '\"';
m_SkipStack.Add('\"');
break;
default:
if (m_CurByte == m_SkipChar) {
iLen--;
m_SkipStack.RemoveAt(iLen, 1);
m_SkipChar = iLen ? m_SkipStack[iLen - 1] : 0;
if (iLen == 0 && m_CurByte == '>') {
m_iDataLength = m_iDataPos;
m_iDataPos = 0;
if (m_iDataLength >= 9 &&
FXSYS_memcmp(m_pszData, "[CDATA[", 7 * sizeof(uint8_t)) == 0 &&
FXSYS_memcmp(m_pszData + m_iDataLength - 2, "]]",
2 * sizeof(uint8_t)) == 0) {
Pop();
m_iDataLength -= 9;
m_dwDataOffset += 7;
FXSYS_memmove(m_pszData, m_pszData + 7,
m_iDataLength * sizeof(uint8_t));
m_bCharData = TRUE;
if (m_pHandler) {
NotifyData();
}
m_bCharData = FALSE;
} else {
Pop();
}
m_eMode = FX_SAXMODE_Text;
}
}
break;
}
if (iLen > 0) {
ParseChar(m_CurByte);
}
}
StdStrBuf C4RTFFile::GetPlainText()
{
// clear any previous crap
ClearState();
// start with a fresh state
pState = new PropertyState();
pState->eState = psNormal;
StdStrBuf sResult;
// nothing to do for empty RTFs
if (sRTF.getSize()<=0) return sResult;
// parse through all chars
try
{
char c; size_t iPos = 0;
while (iPos < sRTF.getSize())
{
c = ((const char *) sRTF.getData())[iPos++];
// binary parsing?
if (pState->eState == psBinary)
{
if (!--pState->iHexBinCnt) pState->eState = psNormal;
ParseChar(sResult, c);
continue;
}
// normal parsing: Handle state blocks
switch (c)
{
case '{': PushState(); break;
case '}': PopState(); break;
case '\\':
ParseKeyword(sResult, iPos);
break;
case 0x0d: case 0x0a: // ignored chars
break;
default:
// regular char parsing
if (pState->eState == psNormal)
// normal mode
ParseChar(sResult, c);
else if (pState->eState == psHex)
ParseHexChar(sResult, c);
else
throw new ParserError("Invalid State");
break;
}
// next char
}
// all states must be closed in the end
if (pState->pNext) throw new ParserError("Block not closed");
}
catch (ParserError *pe)
{
// invalid RTF file: Display error message instead
sResult = "Invalid RTF file: ";
sResult.Append(pe->ErrorText);
delete pe;
}
// cleanup
ClearState();
// return result
return sResult;
}
// Returns true if a node was read successfully, false on EOF
bool XMLReader::ReadInternal()
{
switch (mNodeType) {
case kNone:
FillInputBuffer();
if ((mInputEnd - mInputStart) >= 2) {
uint16_t x = (uint16_t(mInputBuffer[0]) << 8) | mInputBuffer[1];
switch (x) {
case 0xFEFF:
case 0x003C:
mConverter.Reset(TextEncoding::UTF16BE());
break;
case 0xFFFE:
case 0x3C00:
mConverter.Reset(TextEncoding::UTF16LE());
break;
}
}
mNodeType = kDocument;
return true;
case kDocument:
// An XML document can start with:
// document ::= prolog element Misc*
// prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
// XMLDecl ::= '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
// Misc ::= Comment | PI | S
// doctypedecl ::= '<!DOCTYPE' S Name (S ExternalID)? S? ('[' intSubset ']' S?)? '>'
// Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
// PI ::= '<?' PITarget (S (Char* - (Char* '?>' Char*)))? '?>'
// S ::= (#x20 | #x9 | #xD | #xA)+
// If the XML file starts with a byte order mark, throw it away.
// The earlier code for the kNone case has already used it to
// set the default encoding.
ParseChar(UnicodeChar(0xFEFF));
if (BufferStartsWith("<?xml")) {
if (! ParseXmlDeclaration()) return false;
UnicodeString encodingName = GetAttribute("encoding");
if (encodingName.empty()) return true;
TextEncoding newEncoding = TextEncoding::WebCharset(encodingName.c_str());
if (newEncoding == mConverter.GetSourceEncoding()) return true;
// The encoding in the XML declaration is different from the one
// we assumed, so we have to reset all the input buffering and
// re-parse the XmlDeclaration.
mConverter.Reset(newEncoding);
mInput.Restart();
mInputStart = mInputEnd = mInputBuffer;
mOutputStart = mOutputEnd = mOutputBuffer;
ParseChar(UnicodeChar(0xFEFF));
return ParseXmlDeclaration();
}
else if (StartsWithWhitespace()) return ParseRequiredWhitespace();
//else if (BufferStartsWith("<!--")) return ParseComment();
//else if (BufferStartsWith("<?")) return ParseProcessingInstruction();
//else if (BufferStartsWith("<!DOCTYPE")) return ParseDocumentType();
else if (BufferStartsWith("<")) return ParseElement();
else return false;
case kXmlDeclaration:
case kElement:
case kEndElement:
case kText:
case kWhitespace:
if (BufferStartsWith("</")) return ParseEndElement();
else if (BufferStartsWith("<")) return ParseElement();
else return ParseText();
}
return false;
}
/* Parse a run of simple and parenthised regular expressions concatenated
together. Plain characters, dots and character classes (optionally followed
by a repetition specifier) are considered simple. IDEA: \< and friends */
static void ParseConcat(ParseInfo *info)
{
int litLen;
ABool isPrevLit;
litLen = 0;
isPrevLit = FALSE;
while (info->str < info->strEnd) {
if (!isPrevLit)
litLen = 0;
isPrevLit = FALSE;
switch (*info->str) {
case '(':
ParseParen(info);
break;
case ')':
case '|':
return;
case '^':
/* Beginning of line */
Emit(info, A_BOL_MULTI);
info->str++;
break;
case '$':
/* End of line */
Emit(info, A_EOL_MULTI);
info->str++;
break;
case '.':
/* Any character */
if (info->minLen == 0)
memset(info->startChar, 0xff, sizeof(info->startChar));
info->str++;
/* IDEA: Perhaps ANY_ALL (i.e. match newlines)? */
ParseSimpleRepetition(info, A_ANY);
break;
case '[': {
/* Character class */
ABool complement = FALSE;
AReOpcode set[A_SET_SIZE];
AWideChar ch;
int i;
WideCharSet wset;
int flags = 0;
InitWideSet(&wset);
info->str++;
/* Complement set? */
if (info->str < info->strEnd && *info->str == '^') {
info->str++;
complement = TRUE;
}
/* End of expression? */
if (info->str == info->strEnd)
AGenerateError(info, ErrUnmatchedLbracket);
memset(set, 0, sizeof(set));
do {
AReOpcode code;
ch = ParseChar(info, &code);
if (info->str == info->strEnd)
break;
if (code >= CC) {
int i;
const AReOpcode *chClass =
ACharClass[(code - CC) & ~CC_COMP];
for (i = 0; i < A_SET_SIZE; i++) {
if (code & CC_COMP)
set[i] |= ~chClass[i];
else
set[i] |= chClass[i];
}
if (code == CC_W)
flags |= A_WS_WORD_CHAR;
else if (code == (CC_W | CC_COMP))
flags |= A_WS_NOT_WORD_CHAR;
/* Underline character is part of the \w set. */
if (code == CC_W || code == (CC_W | CC_COMP))
AToggleInSet(set, '_');
if (*info->str == '-')
AGenerateError(info, ErrInvalidCharacterSet);
} else {
/* Character range? */
if (*info->str == '-') {
AWideChar hiChar = ch;
/* Skip '-', check end of expression. */
if (++info->str == info->strEnd)
break;
if (*info->str == ']')
AAddToSet(set, '-');
else {
hiChar = ParseChar(info, &code);
if (code >= CC)
AGenerateError(info, ErrInvalidCharacterSet);
}
AddToWideSet(info, &wset, ch, hiChar);
for (; ch <= AMin(hiChar, 255); ch++)
AAddToSet(set, ch);
} else {
AAddWideToSet(set, ch);
AddToWideSet(info, &wset, ch, ch);
}
}
} while (info->str < info->strEnd && *info->str != ']');
if (info->str == info->strEnd)
AGenerateError(info, ErrUnmatchedLbracket);
/* Skip ']'. */
info->str++;
if (info->flags & A_RE_NOCASE) {
for (i = 0; i < 256; i++)
if (AIsInSet(set, i)) {
AAddToSet(set, ALower(i));
AAddToSet(set, AUpper(i));
}
}
AddStartSet(info, set, complement);
ParseSimpleRepetition(info, A_SET);
EmitCharSet(info, set, &wset, complement, flags);
FreeWideSet(&wset);
break;
}
case '*':
case '+':
case '?':
case '{':
AGenerateError(info, ErrInvalidRepeat);
info->str = info->strEnd;
break;
default:
{
AReOpcode code;
AWideChar ch;
ch = ParseChar(info, &code);
/* Special character? */
if (code != A_EMPTY) {
if (code < CC) {
Emit(info, code);
if (code == A_BACKREF || code == A_BACKREF_I) {
int num = ch & ~48;
int oldMaxLen = info->maxLen;
if (!(info->parenFlags & (1 << num)))
AGenerateError(info, ErrInvalidBackReference);
Emit(info, num);
info->maxLen *= 2;
ParseRepetition(info, info->minLen, oldMaxLen, 2,
info->mustStr, info->mustStrBack);
}
} else {
AReOpcode set[A_SET_SIZE];
int flags = 0;
memcpy(set, ACharClass[(code - CC) & ~CC_COMP], 32);
/* Underline character is part of the \w set. */
if (code == CC_W || code == (CC_W | CC_COMP))
AToggleInSet(set, '_');
AddStartSet(info, set, code & CC_COMP);
ParseSimpleRepetition(info, A_SET);
if (code == CC_W || code == (CC_W | CC_COMP))
flags = A_WS_WORD_CHAR;
EmitCharSet(info, set, NULL, code & CC_COMP, flags);
}
break;
}
if (info->flags & A_RE_NOCASE) {
code = A_LITERAL_I;
ch = ALower(ch);
if (info->minLen == 0) {
AAddWideToSet(info->startChar, ch);
AAddWideToSet(info->startChar, AUpper(ch));
}
} else {
code = A_LITERAL;
if (info->minLen == 0)
AAddWideToSet(info->startChar, ch);
}
if (info->str == info->strEnd
|| (*info->str != '*' && *info->str != '+'
&& *info->str != '?' && *info->str != '{')) {
info->minLen++;
info->maxLen++;
litLen++;
isPrevLit = TRUE;
if (litLen == 2) {
/* Convert a single character to literal string. */
AWideChar prevCh;
prevCh = info->buf[info->bufInd - 1];
info->buf[info->bufInd - 2] += A_STRING;
info->buf[info->bufInd - 1] = 2;
Emit(info, prevCh);
Emit(info, ch);
if (info->mustStr == 0) {
info->mustStr = info->bufInd - litLen;
info->mustStrBack = info->maxLen - 2; /* FIX? */
}
} else if (litLen > 2) {
/* Add a character to a literal string. */
Emit(info, ch);
if (info->buf[info->mustStr - 1] <= litLen) {
info->mustStr = info->bufInd - litLen;
info->mustStrBack = info->maxLen - litLen; /* FIX? */
}
info->buf[info->bufInd - litLen - 1] = litLen;
} else {
Emit(info, code);
Emit(info, ch);
}
} else {
ParseSimpleRepetition(info, code);
Emit(info, ch);
}
break;
}
}
}
}
int SB_GetNumber (StrBuf* B, long* Val)
/* Get a number from the string buffer. Accepted formats are decimal, octal,
** hex and character constants. Numeric constants may be preceeded by a
** minus or plus sign. The function returns 1 if a number was found and
** zero otherwise. Errors are only output for invalid numbers.
*/
{
int Sign;
char C;
unsigned Base;
unsigned DigitVal;
/* Initialize Val */
*Val = 0;
/* Handle character constants */
if (SB_Peek (B) == '\'') {
/* Character constant */
SB_Skip (B);
*Val = SignExtendChar (TgtTranslateChar (ParseChar (B)));
if (SB_Peek (B) != '\'') {
Error ("`\'' expected");
return 0;
} else {
/* Skip the quote */
SB_Skip (B);
return 1;
}
}
/* Check for a sign. A sign must be followed by a digit, otherwise it's
** not a number
*/
Sign = 1;
switch (SB_Peek (B)) {
case '-':
Sign = -1;
/* FALLTHROUGH */
case '+':
if (!IsDigit (SB_LookAt (B, SB_GetIndex (B) + 1))) {
return 0;
}
SB_Skip (B);
break;
}
/* We must have a digit now, otherwise its not a number */
C = SB_Peek (B);
if (!IsDigit (C)) {
return 0;
}
/* Determine the base */
if (C == '0') {
/* Hex or octal */
SB_Skip (B);
if (tolower (SB_Peek (B)) == 'x') {
SB_Skip (B);
Base = 16;
if (!IsXDigit (SB_Peek (B))) {
Error ("Invalid hexadecimal number");
return 0;
}
} else {
Base = 8;
}
} else {
Base = 10;
}
/* Read the number */
while (IsXDigit (C = SB_Peek (B)) && (DigitVal = HexVal (C)) < Base) {
*Val = (*Val * Base) + DigitVal;
SB_Skip (B);
}
/* Allow optional 'U' and 'L' modifiers */
C = SB_Peek (B);
if (C == 'u' || C == 'U') {
SB_Skip (B);
C = SB_Peek (B);
if (C == 'l' || C == 'L') {
SB_Skip (B);
}
} else if (C == 'l' || C == 'L') {
SB_Skip (B);
C = SB_Peek (B);
if (C == 'u' || C == 'U') {
SB_Skip (B);
}
}
/* Success, value read is in Val */
*Val *= Sign;
return 1;
}