static void unbreak_backslash_broken_lines(struct token_list *tl, tok_message_queue *mq) {
const char *s = tl->orig, *e = s+tl->orig_size;
darray_char *txt = talloc_darray(tl);
darray(const char*) *olines = talloc_darray(tl);
darray(const char*) *tlines = talloc_darray(tl);
do {
const char *line_start = s, *line_end;
const char *lnw; //last non-white
size_t start_offset = txt->size;
//scan to the next line and find the last non-white character in the line
while (s<e && !creturn(*s)) s++;
line_end = s;
lnw = s;
while (lnw>line_start && cspace(lnw[-1])) lnw--;
if (s<e && creturn(*s)) {
s++;
//check for non-standard newlines (i.e. "\r", "\r\n", or "\n\r")
if (s<e && *s=='\n'+'\r'-s[-1])
s++;
}
//add the backslash-break-free version of the text
if (lnw>line_start && lnw[-1]=='\\' && line_end<e) {
darray_append_items(*txt, line_start, lnw-1-line_start);
if (lnw<e && cspace(*lnw)) {
tok_msg_warn(spaces_after_backslash_break, lnw,
"Trailing spaces after backslash-broken line");
}
} else
darray_append_items(*txt, line_start, s-line_start);
//add the line starts for this line
darray_append(*olines, line_start);
darray_append(*tlines, (const char*)start_offset);
//Since the txt buffer moves when expanded, we're storing offsets
// for now. Once we're done building txt, we can add the base
// of it to all the offsets to make them pointers.
} while (s<e);
//stick a null terminator at the end of the text
darray_realloc(*txt, txt->size+1);
txt->item[txt->size] = 0;
//convert the line start offsets to pointers
{
const char **i;
darray_foreach(i, *tlines)
*i = txt->item + (size_t)(*i);
}
tl->olines = olines->item;
tl->olines_size = olines->size;
tl->txt = txt->item;
tl->txt_size = txt->size;
tl->tlines = tlines->item;
tl->tlines_size = tlines->size;
}
//Reads a C string starting at s until quoteChar is found or e is reached
// Returns the pointer to the terminating quote character or e if none was found
char *read_cstring(array_char *out, const char *s, const char *e, char quoteChar, tok_message_queue *mq) {
const char * const tokstart = s;
const char *p;
int has_endquote=0, has_newlines=0;
//tok_msg_debug(called, s, "Called read_cstring on `%s`", s);
#define append(startptr,endptr) array_append_items(*out, startptr, (endptr)-(startptr))
#define append_char(theChar) array_append(*out, theChar)
#define append_zero() do {array_append(*out, 0); out->size--;} while(0)
p = s;
while (p<e) {
char c = *p++;
if (c == '\\') {
append(s, p-1);
s = p;
if (p >= e) {
append_char('\\');
tok_msg_error(ended_in_backslash, p-1,
"read_cstring input ended in backslash");
break;
}
c = *p++;
if (c>='0' && c<='9') {
unsigned int octal = c-'0';
size_t digit_count = 0;
while (p<e && *p>='0' && *p<='9') {
octal <<= 3;
octal += (*p++) - '0';
if (++digit_count >= 2)
break;
}
if (p<e && *p>='0' && *p<='9') {
tok_msg_info(ambiguous_octal, s-2,
"Octal followed by digit");
}
if (octal > 0xFF) {
tok_msg_warn(octal_overflow, s-2,
"Octal out of range");
}
c = octal;
} else {
switch (c) {
case 'x': {
size_t digit_count = 0;
size_t zero_count = 0;
unsigned int hex = 0;
while (p<e && *p=='0') p++, zero_count++;
for (;p<e;digit_count++) {
c = *p++;
if (c>='0' && c<='9')
c -= '0';
else if (c>='A' && c<='F')
c -= 'A'-10;
else if (c>='a' && c<='f')
c -= 'a'-10;
else {
p--;
break;
}
hex <<= 4;
hex += c;
}
if (zero_count+digit_count > 2) {
char *hex_string = strdup_rng(s-2, p);
tok_msg_warn(ambiguous_hex, s-2,
"Hex escape '%s' is ambiguous", hex_string);
if (digit_count > 2)
tok_msg_warn(hex_overflow, s-2,
"Hex escape '%s' out of range", hex_string);
free(hex_string);
}
c = hex & 0xFF;
} break;
case 'a':
c=0x7;
break;
case 'b':
c=0x8;
break;
case 'e':
c=0x1B;
break;
case 'f':
c=0xC;
break;
case 'n':
c=0xA;
break;
case 'r':
c=0xD;
break;
case 't':
c=0x9;
break;
case 'v':
c=0xB;
break;
case '\\':
break;
default:
if (c == quoteChar)
break;
if (c=='\'' && quoteChar=='"') {
/* tok_msg_info(escaped_single_quote, s-2,
"Single quote characters need not be escaped within double quotes"); */
break;
}
if (c=='"' && quoteChar=='\'') {
/* tok_msg_info(escaped_double_quote, s-2,
"Double quote characters need not be escaped within single quotes"); */
break;
}
if (c=='?') // \? is needed in some situations to avoid building a trigraph
break;
tok_msg_warn(unknown_escape, s-2,
"Unknown escape sequence '\\%c'", c);
break;
}
}
s = p;
append_char(c);
} else if (c == quoteChar) {
p--;
has_endquote = 1;
break;
} else if (creturn(c)) {
has_newlines = 1;
}
}
append(s, p);
append_zero();
if (!has_endquote) {
tok_msg_error(missing_endquote, tokstart,
"Missing endquote on %s literal",
quoteChar=='\'' ? "character" : "string");
} else if (has_newlines) {
tok_msg_warn(quote_newlines, tokstart,
"%s literal contains newline character(s)",
quoteChar=='\'' ? "Character" : "String");
}
return (char*)p;
#undef append
#undef append_char
#undef append_zero
}
