int
daplex(YYSTYPE* lvalp, DAPparsestate* state)
{
DAPlexstate* lexstate = state->lexstate;
int token;
int c;
unsigned int i;
char* p;
char* tmp;
YYSTYPE lval = NULL;
token = 0;
ocbytesclear(lexstate->yytext);
/* invariant: p always points to current char */
for(p=lexstate->next; token==0&&(c=*p); p++) {
if(c == '\n') {
lexstate->lineno++;
} else if(c <= ' ' || c == '\177') {
/* whitespace: ignore */
} else if(c == '#') {
/* single line comment */
while((c=*(++p))) {
if(c == '\n') break;
}
} else if(strchr(lexstate->worddelims,c) != NULL) {
/* don't put in lexstate->yytext to avoid memory leak */
token = c;
} else if(c == '"') {
int more = 1;
/* We have a string token; will be reported as WORD_STRING */
while(more && (c=*(++p))) {
if(c == '"') {
more = 0;
continue;
}
#ifdef DAP2ENCODE
if(c == '\\') {
/* Resolve spec ambiguity about handling of \c:
1. !KEEPSLASH: convert \c to c for any character c
2. KEEPSLASH: convert \c to \c for any character c;
that is, keep the backslash.
It is clear that the problem being addressed was \".
But it is unclear what to to do about \n: convert to
Ascii LF or leave as \n.
This code will leave as \n and assume higher levels
of code will address the issue.
*/
#ifdef KEEPSLASH
dapaddyytext(lexstate,c);
#endif
c=*(++p);
if(c == '\0') more = 0;
}
#else /*Non-standard*/
switch (c) {
case '\\':
c=*(++p);
switch (c) {
case 'r':
c = '\r';
break;
case 'n':
c = '\n';
break;
case 'f':
c = '\f';
break;
case 't':
c = '\t';
break;
case 'x': {
int d1,d2;
c = '?';
++p;
d1 = tohex(*p++);
if(d1 < 0) {
daperror(state,"Illegal \\xDD in TOKEN_STRING");
} else {
d2 = tohex(*p++);
if(d2 < 0) {
daperror(state,"Illegal \\xDD in TOKEN_STRING");
} else {
c=(((unsigned int)d1)<<4) | (unsigned int)d2;
}
}
}
break;
default:
break;
}
break;
default:
break;
}
#endif /*!DAP2ENCODE*/
if(more) dapaddyytext(lexstate,c);
}
token=WORD_STRING;
} else if(strchr(lexstate->wordchars1,c) != NULL) {
int isdatamark = 0;
/* we have a WORD_WORD */
dapaddyytext(lexstate,c);
while((c=*(++p))) {
#ifdef URLCVT
if(c == '%' && p[1] != 0 && p[2] != 0
&& strchr(hexdigits,p[1]) != NULL
&& strchr(hexdigits,p[2]) != NULL) {
int d1,d2;
d1 = tohex(p[1]);
d2 = tohex(p[2]);
if(d1 >= 0 || d2 >= 0) {
c=(((unsigned int)d1)<<4) | (unsigned int)d2;
p+=2;
}
} else {
if(strchr(lexstate->wordcharsn,c) == NULL) {
p--;
break;
}
}
dapaddyytext(lexstate,c);
#else
if(strchr(lexstate->wordcharsn,c) == NULL) {
p--;
break;
}
dapaddyytext(lexstate,c);
#endif
}
/* Special check for Data: */
tmp = ocbytescontents(lexstate->yytext);
if(strcmp(tmp,"Data")==0 && *p == ':') {
dapaddyytext(lexstate,*p);
p++;
if(p[0] == '\n') {
token = SCAN_DATA;
isdatamark = 1;
p++;
} else if(p[0] == '\r' && p[1] == '\n') {
token = SCAN_DATA;
isdatamark = 1;
p+=2;
}
}
if(!isdatamark) {
/* check for keyword */
token=WORD_WORD; /* assume */
for(i=0;; i++) {
if(keywords[i] == NULL) break;
if(strcasecmp(keywords[i],tmp)==0) {
token=keytokens[i];
break;
}
}
}
} else { /* illegal */
}
}
lexstate->next = p;
strncpy(lexstate->lasttokentext,ocbytescontents(lexstate->yytext),MAX_TOKEN_LENGTH);
lexstate->lasttoken = token;
if(ocdebug >= 2)
dumptoken(lexstate);
/*Put return value onto Bison stack*/
if(ocbyteslength(lexstate->yytext) == 0)
lval = NULL;
else {
lval = ocbytesdup(lexstate->yytext);
oclistpush(lexstate->reclaim,(void*)lval);
}
if(lvalp) *lvalp = lval;
return token; /* Return the type of the token. */
}
int
daplex(YYSTYPE* lvalp, DAPparsestate* state)
{
DAPlexstate* lexstate = state->lexstate;
int token;
int c;
unsigned int i;
char* p=lexstate->next;
char* tmp;
token = 0;
ocbytesclear(lexstate->yytext);
/* invariant: p always points to current char */
for(p=lexstate->next;token==0&&(c=*p);p++) {
if(c == '\n') {
lexstate->lineno++;
} else if(c <= ' ' || c == '\177') {
/* whitespace: ignore */
} else if(c == '#') {
/* single line comment */
while((c=*(++p))) {if(c == '\n') break;}
} else if(strchr(lexstate->worddelims,c) != NULL) {
/* don't put in lexstate->yytext to avoid memory leak */
token = c;
} else if(c == '"') {
int more = 1;
/* We have a string token; will be reported as SCAN_WORD */
while(more && (c=*(++p))) {
#ifdef NONSTDCVT
switch (c) {
case '"': more=0; break;
case '\\':
c=*(++p);
switch (c) {
case 'r': c = '\r'; break;
case 'n': c = '\n'; break;
case 'f': c = '\f'; break;
case 't': c = '\t'; break;
case 'x': {
int d1,d2;
c = '?';
++p;
d1 = tohex(*p++);
if(d1 < 0) {
daperror(state,"Illegal \\xDD in TOKEN_STRING");
} else {
d2 = tohex(*p++);
if(d2 < 0) {
daperror(state,"Illegal \\xDD in TOKEN_STRING");
} else {
c=(((unsigned int)d1)<<4) | (unsigned int)d2;
}
}
} break;
default: break;
}
break;
default: break;
}
#else /*!NONSTDCVT*/
if(c == '"')
more = 0;
else if(c == '\\') {
c=*(++p);
if(c == '\0') more = false;
if(c != '"') {c = '\\'; --p;}
}
#endif /*!NONSTDCVT*/
if(more) dapaddyytext(lexstate,c);
}
token=SCAN_WORD;
} else if(strchr(lexstate->wordchars1,c) != NULL) {
/* we have a SCAN_WORD */
dapaddyytext(lexstate,c);
while((c=*(++p))) {
#ifdef URLCVT
if(c == '%' && p[1] != 0 && p[2] != 0
&& strchr(hexdigits,p[1]) != NULL
&& strchr(hexdigits,p[2]) != NULL) {
#ifdef WRONG /* Should not unescape %xx occurrences */
int d1,d2;
d1 = tohex(p[1]);
d2 = tohex(p[2]);
if(d1 >= 0 || d2 >= 0) {
c=(((unsigned int)d1)<<4) | (unsigned int)d2;
p+=2;
}
#endif
} else {
if(strchr(lexstate->wordcharsn,c) == NULL) {p--; break;}
}
dapaddyytext(lexstate,c);
#else
if(strchr(lexstate->wordcharsn,c) == NULL) {p--; break;}
dapaddyytext(lexstate,c);
#endif
}
/* Special check for Data: */
tmp = ocbytescontents(lexstate->yytext);
if(strcmp(tmp,"Data")==0 && *p == ':') {
dapaddyytext(lexstate,*p); p++;
token = SCAN_DATA;
} else {
/* check for keyword */
token=SCAN_WORD; /* assume */
for(i=0;;i++) {
if(keywords[i] == NULL) break;
if(strcasecmp(keywords[i],tmp)==0) {
token=keytokens[i];
break;
}
}
}
} else { /* illegal */
}
}
lexstate->next = p;
strncpy(lexstate->lasttokentext,ocbytescontents(lexstate->yytext),MAX_TOKEN_LENGTH);
lexstate->lasttoken = token;
if(ocdebug >= 2)
dumptoken(lexstate);
/*Put return value onto Bison stack*/
if(ocbyteslength(lexstate->yytext) == 0)
*lvalp = NULL;
else {
*lvalp = ocbytesdup(lexstate->yytext);
oclistpush(lexstate->reclaim,(ocelem)*lvalp);
}
return token; /* Return the type of the token. */
}
