WicErrors getNextToken(pToken tok)
{
pTokTab tokTabPtr;
WicErrors retVal = ERR_NONE;
static long tokAfterDefine = 2; // used to flag '(' in #define x( as a
// special parentheses
int temp;
assert(currTokF >= 0);
currTokLen = 0;
currTok[currTokLen] = 0;
TOK_NUM_AFTER_NEW_LINE++; /* Used for #preprocessor directives */
tokAfterDefine++; /* Used for #preprocessor directives */
g_currLineNum = LINE_NUM;
g_currColNum = COL_NUM;
/* When getNextToken gets called, STATE may be one of:
TS_START, TS_COMMENT. */
temp = skipBlank();
if (STATE == TS_START) {
setTokPos(
tok->pos,
TOK_FILE_NAME,
currTokF,
LINE_NUM,
COL_NUM,
LINES_BEFORE,
temp,
orderLineNum
);
while (NEXT_CHAR == ' ') {
getNextChar();
tok->pos->spacesBefore = skipBlank();
}
if (isalpha(NEXT_CHAR) || NEXT_CHAR == '_') {
if (!scanId()) {
char saveChar = NEXT_CHAR;
currTokLen = 0;
currTok[currTokLen] = 0;
getNextChar();
if (saveChar == '"') {
retVal = scanStr(tok->data);
} else {
retVal = scanChar(tok->data);
}
goto Return;
}
} else if (isdigit(NEXT_CHAR)) {
retVal = scanNum(tok->data);
goto Return;
} else switch (NEXT_CHAR) {
case '\'':
getNextChar();
retVal = scanChar(tok->data);
goto Return;
break;
case '"':
if (currLineIsInclude) {
getNextChar();
retVal = scanIncludeFileName(tok->data, '"');
goto Return;
} else {
getNextChar();
retVal = scanStr(tok->data);
goto Return;
break;
}
case '\n':
pushGetNextChar();
currLineIsInclude = 0;
break;
case '!':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case '#':
pushGetNextChar();
if (TOK_NUM_AFTER_NEW_LINE == 1) {
skipBlank();
if (isalpha(NEXT_CHAR) || NEXT_CHAR == '_') {
scanId();
} else {
tok->data->code = Y_PRE_NULL;
retVal = ERR_NONE;
goto Return;
}
} else {
if (NEXT_CHAR == '#') {
pushGetNextChar();
}
}
break;
case '%':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case '&':
pushGetNextChar();
if (NEXT_CHAR == '&') {
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
}
break;
case '(':
pushGetNextChar();
break;
case ')':
pushGetNextChar();
break;
case '*':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case '+':
pushGetNextChar();
if (NEXT_CHAR == '+') {
pushGetNextChar();
} else if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case ',':
pushGetNextChar();
break;
case '-':
pushGetNextChar();
if (NEXT_CHAR == '-') {
pushGetNextChar();
} else if (NEXT_CHAR == '=') {
pushGetNextChar();
} else if (NEXT_CHAR == '>') {
pushGetNextChar();
}
break;
case '.':
pushGetNextChar();
if (NEXT_CHAR == '.') {
pushGetNextChar();
if (NEXT_CHAR == '.') {
pushGetNextChar();
} else {
retVal = RERR_INV_CHAR;
goto Return;
}
} else if (isdigit(NEXT_CHAR)) {
if (pushFloatDotExp(tok->data, 1)) {
retVal = convStr2Const(tok->data);
goto Return;
} else {
retVal = RERR_INV_CHAR;
goto Return;
}
}
break;
case '/':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
} else if (NEXT_CHAR == '*') { /* comment begin */
popChars(1);
STATE = TS_COMMENT;
getNextChar();
retVal = scanComment(tok->data);
goto Return;
} else if (NEXT_CHAR == '/') {
popChars(1);
STATE = TS_COMMENT;
getNextChar();
retVal = scanCPlusPlusComment(tok->data);
goto Return;
}
break;
case ':':
pushGetNextChar();
if (NEXT_CHAR == '>') {
pushGetNextChar();
}
break;
case ';':
pushGetNextChar();
break;
case '<':
if (currLineIsInclude) {
getNextChar();
retVal = scanIncludeFileName(tok->data, '>');
goto Return;
} else {
pushGetNextChar();
if (NEXT_CHAR == '<') {
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
} else if (NEXT_CHAR == '=') {
pushGetNextChar();
}
}
break;
case '=':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case '>':
pushGetNextChar();
if (NEXT_CHAR == '>') {
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
}
} else if (NEXT_CHAR == '=') {
pushGetNextChar();
}
break;
case '?':
pushGetNextChar();
break;
case '[':
pushGetNextChar();
break;
case ']':
pushGetNextChar();
break;
case '^':
pushGetNextChar();
if (NEXT_CHAR == '=')
pushGetNextChar();
break;
case '{':
pushGetNextChar();
break;
case '|':
pushGetNextChar();
if (NEXT_CHAR == '=') {
pushGetNextChar();
} else if (NEXT_CHAR == '|') {
pushGetNextChar();
}
break;
case '}':
pushGetNextChar();
break;
case '~':
pushGetNextChar();
break;
case (char) EOF:
tok->data->code = Y_EOF;
retVal = ERR_NONE;
goto Return;
break;
default:
/* Eat up an ivalid character */
getNextChar();
retVal = RERR_INV_CHAR;
goto Return;
}
tokTabPtr = tabLookup(currTok);
if (tokTabPtr != NULL)
{
tok->data->code = tokTabPtr->code;
if (tok->data->code == Y_PRE_INCLUDE) {
currLineIsInclude = 1;
}
if (tok->data->code == Y_PRE_DEFINE) {
tokAfterDefine = 0;
}
if (tok->data->code == Y_LEFT_PAREN && tokAfterDefine == 2) {
// the case of #define x(...
if (tok->pos->spacesBefore == 0) {
tok->data->code = Y_PRE_SPECIAL_LEFT_PAREN;
}
}
tok->data->repr.string = registerString(tokTabPtr->name,
!FREE_STRING);
}
else {
if (currTok[0] == '#') {
retVal = RERR_INV_PREPROCESSOR;
goto Return;
} else {
tok->data->code = Y_ID;
tok->data->repr.string = registerString(wicStrdup(currTok),
FREE_STRING);
}
}
} else if (STATE == TS_COMMENT) {
setTokPos(tok->pos, TOK_FILE_NAME, currTokF, LINE_NUM, COL_NUM,
LINES_BEFORE, 0, orderLineNum);
retVal = scanComment(tok->data);
goto Return;
} else {
assert(0);
}
Return:
if (tok->data->code != Y_PRE_NEWLINE) {
tok->pos->linesBefore = tok->pos->lineNum - PREV_TOK_LINE_NUM;
PREV_TOK_LINE_NUM = tok->pos->lineNum;
} else {
tok->pos->linesBefore = 0;
}
zapTokPos(g_currPos);
g_currPos = dupTokPos(tok->pos, NULL);
return retVal;
}
void EBNFScanner::nextToken()
{
if(currentToken > -1 && currentToken < tokens.size() - 1)
{
currentToken++;
return;
}
while (hasNextCharacter())
{
nextCharacter();
// Scan the Identifier type. [a-zA-Z\-_]
if(TokenScannerHelper::isAlpha(currentCharacter))
{
Token tok = scanIdentifier();
tokens.append(std::move(tok));
currentToken++;
return;
}
// Scan a literal
else if (TokenScannerHelper::isQuote(currentCharacter))
{
Token tok = scanLiteral();
tokens.append(std::move(tok));
currentToken++;
return;
}
// Scan an operator Token.
else if (TokenScannerHelper::isOperator(currentCharacter))
{
if (currentCharacter == '(' && peekCharacter() == '*') {
Token tok = scanComment();
tokens.append(std::move(tok));
currentToken++;
return;
}
else {
Token tok = scanOperator();
tokens.append(std::move(tok));
currentToken++;
return;
}
}
else if (TokenScannerHelper::isWhiteSpace(currentCharacter))
{
// Ignore white spaces.
}
else
{
raiseError (String("Invalid Character."));
}
checkNewLine();
}
tokens.append(Token(Token::TypeEof, String(""), createPosition()));
currentToken++;
}
int GetToken(void)
{
/* skip white space and comment */
while( isspace(nextch) || ('/' == nextch)) {
if (nextch == '\n') {
lineno++;
}
/* skip comment */
if( nextch == '/') {
nextch=getc(infile);
if(nextch == '/')
{ /* eat to end of line */
do { nextch=getc(infile);
} while(nextch != '\n');
lineno++;
}
else if(nextch == '*')
scanComment();
else {
return('/');
}
}
nextch = getc(infile);
}
if (nextch == EOF) return( T_EOF);
if (isalpha (nextch) || nextch == '_' ) /*scan identifier */
{ char *s = lexeme;
do {
*s = nextch; s++;
nextch = getc(infile);
} while( isalpha(nextch) || isdigit( nextch ) || (nextch == '_' ) );
*s = '\0';
return ID;
}
switch (nextch)
{
case ':' : nextch = getc(infile); return COLON;
case ';' : nextch = getc(infile); return SEMI;
case '|' : nextch = getc(infile); return BAR;
case '(' : nextch = getc(infile); return LPAREN;
case ')' : nextch = getc(infile); return RPAREN;
case '*' : nextch = getc(infile); return STAR;
case '?' : nextch = getc(infile); return QUEST;
case '+' : nextch = getc(infile); return PLUS;
case '{' : nextch = getc(infile); return LBRACE;
case '}' : nextch = getc(infile); return RBRACE;
case ',' : nextch = getc(infile); return COMMA;
case '%' : nextch = getc(infile);
if (nextch == '%') {
nextch = getc(infile);
strcpy(lexeme, "%%" );
return END_SPEC;
} else if(isalpha(nextch)) {
char *s = lexeme;
do {
*s++ = nextch;
nextch = getc(infile);
} while(isalpha(nextch));
*s = '\0';
if (strcmp(lexeme, "token") == 0) {
return TOKEN_SPEC;
} else {
SyntaxError("Unknown directive");
return( BAD_TOKEN );
}
} else
return( BAD_TOKEN );
case '\"': /*multichar token "token" or single char "c" */
{
char *s = lexeme;
while (nextch != EOF) {
nextch = toupper(getc(infile));
while (nextch <= ' ' )
nextch = toupper(getc(infile));
if (nextch == '\"'){
nextch = getc(infile);
break;
}
*s = nextch; s++;
};
*s = '\0';
if (strlen(lexeme) == 1) { /*single char*/
expand_char(lexeme, lexeme[1]);
return CHR_LIT;
}
else
return ID;
}
/* single char token 'c'*/
case '\'':
expand_char(lexeme, getc(infile));
nextch = getc(infile);
if (nextch != '\'') {
SyntaxError("Single char token must be terminated with \'.");
return(BAD_TOKEN);
} else {
nextch = getc(infile);
return CHR_LIT;
}
}
return BAD_TOKEN;
}
void Compiler::scanPossibleComment(){
if(tok->getnext() == T_COMMENT){
scanComment(false);
} else
tok->rewind();
}
void Compiler::scanStmt() {
char buf[256]; // temporary buffer
int t;
switch(tok->getnext()){
case T_IDENT: // it's an expression
case T_INT:
case T_FLOAT:
case T_SUB:
case T_BITNOT:
case T_PLING:
case T_STRING:
case T_BACKTICK:
case T_OPREN:
// deal with debugging words!
if(!strcmp(tok->getstring(),"dumplocs")){
cg->emit(OP_SPECIAL,0);break;
}
if(!strcmp(tok->getstring(),"breakpoint")){
cg->emit(OP_SPECIAL,1);break;
}
tok->rewind(); // put the token back
// scan the expression, might be a label
if(!scanExpr(true))
// clear all statements if not a func or other oddity,
// or just a dummy for recreation purposes if in immediate mode.
cg->emit(cg->isCompiling()?OP_ENDESTMT:OP_ENDESTMT2);
break;
case T_LOAD:
{
if(cg->isCompiling())
error("can only run 'load' in interactive mode");
Session *s;
if(tok->getnext()!=T_STRING)
error("expected a string after 'load'");
try {
s = new Session(ses->api);
s->feedFile(tok->getstring());
} catch(Exception &e){
delete s;
throw e;
}
delete s;
}
break;
case T_SAVE:
{
if(cg->isCompiling())
error("can only run 'save' in interactive mode");
if(tok->getnext()!=T_STRING)
error("expected a string after 'save'");
const char *fname = tok->getstring();
FILE *a;
if(!strlen(fname))
a = stdout;
else
a = fopen(fname,"w");
if(!a)
error("cannot open file '%s'",fname);
Serialiser *ser = new Serialiser(ses);
ser->write(a);
if(strlen(fname))
fclose(a);
delete ser;
}
break;
case T_SAVEVAR:
{
if(cg->isCompiling())
error("can only run 'savevar' in interactive mode");
if(tok->getnext()!=T_IDENT)
error("expected a variable name after 'savevar'");
const char *vname = tok->getstring();
int vdesc = lana->consts->findOrCreateString(vname);
if(tok->getnext()!=T_STRING)
error("expected a string after 'savevar'");
const char *fname = tok->getstring();
// try to get the value
Value *v;
int id;
id = lana->globs->find(vdesc);
if(id>=0) {
v = lana->globs->get(id); // it's a global
} else {
id = ses->findSesVar(vdesc);
if(id<0)
error("variable not found: %s",lana->consts->getStr(vdesc));
v = ses->getSesVar(id);
}
FILE *a;
if(!strlen(fname))
a = stdout;
else
a = fopen(fname,"w");
if(!a)
error("cannot open file '%s'",fname);
Serialiser *ser = new Serialiser(ses);
ser->serialiseValue(a,v,lana->consts->getStr(vdesc));
if(strlen(fname))
fclose(a);
delete ser;
}
break;
case T_FOR:
scanFor();
break;
case T_ENDFOR:
scanEndFor();
break;
case T_THIS:
tok->rewind(); // put the token back
if(!scanExpr(true))
// clear all statements if not a func or other oddity,
// or just a dummy for recreation purposes if in immediate mode.
cg->emit(cg->isCompiling()?OP_ENDESTMT:OP_ENDESTMT2);
break;
case T_GOTO:
if(!cg->isCompiling())
error("must be compiling a function/procedure to use '%s'",tok->getstring());
scanGoto();
break;
case T_ENDFUNC:
scanEndFunc();
break;
case T_END:
if(!(lana->opFlags & LOP_STRIPCOMMENTS))
cg->emit(OP_BLANKLINE); // yes, these are wasteful .. very slightly
break;
case T_IF:
// first we push a special value onto the compiler stack
// to mark the start of this if..elseif..elseif..endif
cg->current->cpush(-9999);
// we scan the expression
if(scanExpr())
error("cannot use a function/procedure expression in if");
// stack and output an incomplete if - but this might be a normal if, or a quick if.
cg->current->cpushhere();
// now for some cleverness. Is the next token a colon?
if(tok->getnext() == T_COLON){
cg->emit(OP_QUICKIF,-100);
// if so, parse the next statement recursively
scanStmt();
// note that we don't need to output a quick endif, since the recreator
// doesn't need it!
instruction *ptr = cg->current->cpoplocandcheck(OP_QUICKIF,OP_QUICKIF); // MUST be an OP_IF, no ELSE.
if(!ptr)
error("not a simple statement in quick-if");
// write the IF, ELSE or ELSEIF again with the correct distance
*ptr = INST(INSTOP(*ptr),cg->current->getdiff(ptr));
// now pop off!
int n;
do{
n = cg->current->cpop();
}while(n!=-9999);
} else {
// not - put it back!
tok->rewind();
cg->emit(OP_IF,-100);
}
break;
case T_ENDIF:
{
// get the corresponding OP_IF, OP_ELSEIF or OP_ELSE
cg->emit(OP_ENDIF);
instruction *ptr = cg->current->cpoplocandcheck(OP_IF,OP_ELSE);
if(!ptr)
error("mismatched endif");
// write the IF, ELSE or ELSEIF again with the correct distance
*ptr = INST(INSTOP(*ptr),cg->current->getdiff(ptr));
// now pop and fixup OP_JMPELSEIFs until we get the special -9999 which marked the start
for(;;){
int n = cg->current->cpop();
if(n==-9999)break; // we're done!
// we're not done - get the code pointer
instruction *ptr = cg->current->getPtr(n*sizeof(instruction));
// make sure it's a OP_JMPELSEIF!
if(INSTOP(*ptr)!=OP_JMPELSEIF)
error("badly formed conditional statement");
// change it so that it jumps to the current location
*ptr = INST(INSTOP(*ptr),cg->current->getdiff(ptr));
}
}
break;
case T_ELSEIF:
{
// pop the instruction off the stack, an OP_IF or OP_ELSEIF
instruction *ptr = cg->current->cpoplocandcheck(OP_IF,OP_ELSEIF);
// first we need to terminate the previous condition, so
// push the location and output a OP_JMPELSEIF ready to fill in.
// This will get left on the stack!
cg->current->cpushhere();
cg->emit(OP_JMPELSEIF,-100);
// now make the IF or ELSEIF we popped jump to this point
*ptr = INST(INSTOP(*ptr),cg->current->getdiff(ptr));
// now scan the expression
if(scanExpr())
error("cannot use a function/procedure expression in if");
// push the location, and..
// output an OP_ELSEIF with a dummy jump
cg->current->cpushhere();
cg->emit(OP_ELSEIF,-100);
}
break;
case T_ELSE:
{
// write the OP_ELSE which will become a jump forward,
// but first recording the location
int elseloc = cg->current->getloc();
cg->emit(OP_ELSE,-100);
// now we need to make the IF jump to here
// get the corresponding OP_IF or OP_ELSEIF
instruction *ptr = cg->current->cpoplocandcheck(OP_IF,OP_ELSEIF);
if(!ptr)
error("mismatched else");
// write the IF again with the correct jump distance
int diff = cg->current->getdiff(ptr);
*ptr = INST(INSTOP(*ptr),diff);
// now push the location of the OP_ELSE, which
// will get processed by the OP_ENDIF
cg->current->cpush(elseloc);
}
break;
// pop the location
case T_RETURN:
if(!cg->isCompiling())
error("must be compiling a function/procedure to use '%s'",tok->getstring());
if(tok->getnext() == T_END){ // end of line?
tok->rewind();
// no return value
if(cg->current->ldth.flags & LDTF_RETURNS)
error("functions must return a value");
cg->emit(OP_RETURN,0);
} else {
tok->rewind();
if(!(cg->current->ldth.flags & LDTF_RETURNS))
error("procedures cannot return a value");
if(scanExpr())
error("cannot directly return a function");
cg->emit(OP_RETURN,1);
}
break;
case T_WHILE:
if(!cg->isCompiling())
error("must be compiling a function/procedure to use '%s'",tok->getstring());
// push the current location onto the stack - this is where ENDWHILE will
// jump to
cg->current->cpushhere();
// we also create and push the loop data here, so that we can use break and
// continue!
cg->current->newloop();
// scan and output the expression
if(scanExpr())
error("cannot use a function/procedure expression in `while`");
// push the WHILE onto the stack so we can write the terminating jump into it
cg->current->cpushhere();
// output OP_WHILE with a dummy
cg->emit(OP_WHILE,-100);
break;
case T_ENDWHILE:
{
// pop the location of the WHILE from the stack
instruction *whileptr = cg->current->cpoplocandcheck(OP_WHILE,OP_WHILE);
if(!whileptr)
error("mismatched endwhile");
// pop the location for the backward jump
instruction *jumpdest = cg->current->cpoplocation();
// output the endwhile, which will do the backward jump
cg->emit(OP_ENDWHILE,cg->current->getdiff(jumpdest));
// now patch the while instruction with the forward jump to use if the
// condition is false
*whileptr = INST(OP_WHILE,cg->current->getdiff(whileptr));
/// and end the loop, setting the break label and popping the loop stack
cg->current->endloop();
break;
}
case T_REPEAT:
if(!cg->isCompiling())
error("must be compiling a function/procedure to use '%s'",tok->getstring());
// output OP_REPEAT, pushing its location. We don't
// jump to here, though - we jump to the following opcode.
// This is done just so we can check that the until matches
// a repeat. See T_UNTIL.
cg->current->cpushhere();
cg->current->newloop(); // push and initialise a new loop stack entry (see T_WHILE above)
cg->emit(OP_REPEAT,0);
break;
case T_UNTIL:
{
// scan and output the expression
if(scanExpr())
error("cannot use a function/procedure expression in `until`");
// pop the location of the OP_REPEAT from the stack
instruction *ptr = cg->current->cpoplocandcheck(OP_REPEAT,OP_REPEAT);
if(!ptr)
error("mismatched `until'");
// increment this, because we want to save cycles by
// jumping past the OP_REPEAT (which is a kind of noop)
ptr++;
// and output the OP_UNTIL jump
cg->emit(OP_UNTIL,cg->current->getdiff(ptr));
cg->current->endloop(); // end the current loop stack entry (see T_ENDWHILE above)
break;
}
case T_BREAK:
{
// we want to break out of the topmost loop on the loop stack
// get address we're about to write to
instruction *op = cg->current->getlocptr();
// output the break which will be patched later
cg->emit(OP_BREAK,-100);
// and this jump as a jump to be patched when the break label is resolved
LoopData *d = cg->current->loopstack.peekptr();
if(!d)
throw ParseException("break with no loop");
d->breaklabel.jumpFrom(op);
}
break;
case T_CONTINUE:
{
// we want to terminate the current iteration of the topmost loop on the loop stack
// and immediately start the loop code again
// get address we're about to write to
instruction *op = cg->current->getlocptr();
// output the break which will be patched later
cg->emit(OP_CONTINUE,-100);
// and this jump as a jump to be patched when the break label is resolved
LoopData *d = cg->current->loopstack.peekptr();
if(!d)
throw ParseException("continue with no loop");
d->continuelabel.jumpFrom(op);
}
break;
case T_COMMENT:
scanComment(true);
break;
default:
error("unexpected token '%s'",tok->getstring());
}
// see if there's a comment at the end
scanPossibleComment();
if(tok->getnext()!=T_END)
error("trailing garbage at end of line");
}
void Html::extractCode(OutBuffer *buf)
{
//printf("Html::extractCode()\n");
dbuf = buf; // save for other routines
buf->reserve(end - p);
inCode = 0;
while (1)
{
//printf("p = %p, *p = x%x\n", p, *p);
switch (*p)
{
#if 0 // strings are not recognized outside of tags
case '"':
case '\'':
skipString();
continue;
#endif
case '<':
if (p[1] == '!' && isCommentStart())
{ // Comments start with <!--
scanComment();
}
else if(p[1] == '!' && isCDATAStart())
{
scanCDATA();
}
else if (p[1] == '/' && istagstart(*skipWhite(p + 2)))
skipTag();
else if (istagstart(*skipWhite(p + 1)))
skipTag();
else
goto Ldefault;
continue;
case 0:
case 0x1a:
break; // end of file
case '&':
if (inCode)
{ // Translate character entity into ascii for D parser
int c;
c = charEntity();
buf->writeUTF8(c);
}
else
p++;
continue;
case '\r':
if (p[1] == '\n')
goto Ldefault;
case '\n':
linnum++;
// Always extract new lines, so that D lexer counts the
// lines right.
buf->writeByte(*p);
p++;
continue;
default:
Ldefault:
if (inCode)
buf->writeByte(*p);
p++;
continue;
}
break;
}
buf->writeByte(0); // ending sentinel
//printf("D code is: '%s'\n", (char *)buf->data);
}
void Html::skipTag()
{
enum TagState // what parsing state we're in
{
TStagstart, // start of tag name
TStag, // in a tag name
TSrest, // following tag name
};
enum TagState state = TStagstart;
int inot;
unsigned char *tagstart = NULL;
int taglen = 0;
p++;
inot = 0;
if (*p == '/')
{ inot = 1;
p++;
}
while (1)
{
switch (*p)
{
case '>': // found end of tag
p++;
break;
case '"':
case '\'':
state = TSrest;
skipString();
continue;
case '<':
if (p[1] == '!' && isCommentStart())
{ // Comments start with <!--
scanComment();
}
else if (p[1] == '/' && istagstart(*skipWhite(p + 2)))
{ error("nested tag");
skipTag();
}
else if (istagstart(*skipWhite(p + 1)))
{ error("nested tag");
skipTag();
}
// Treat comments as if they were whitespace
state = TSrest;
continue;
case 0:
case 0x1a:
error("end of file before end of tag");
break; // end of file
case '\r':
if (p[1] == '\n')
goto Ldefault;
case '\n':
linnum++;
// Always extract new lines, so that code lexer counts the
// lines right.
dbuf->writeByte(*p);
state = TSrest; // end of tag
p++;
continue;
case ' ':
case '\t':
case '\f':
case '\v':
if (state == TStagstart)
{ p++;
continue;
}
default:
Ldefault:
switch (state)
{
case TStagstart: // start of tag name
assert(istagstart(*p));
state = TStag;
tagstart = p;
taglen = 0;
break;
case TStag:
if (istag(*p))
{ // Continuing tag name
taglen++;
}
else
{ // End of tag name
state = TSrest;
}
break;
case TSrest:
break;
}
p++;
continue;
}
break;
}
// See if we parsed a <code> or </code> tag
if (taglen && memicmp((char *) tagstart, (char *) "CODE", taglen) == 0
&& *(p - 2) != '/') // ignore "<code />" (XHTML)
{
if (inot)
{ inCode--;
if (inCode < 0)
inCode = 0; // ignore extra </code>'s
}
else
inCode++;
}
}
