mirror of
https://github.com/qpdf/qpdf.git
synced 2024-11-18 02:25:08 +00:00
1037 lines
25 KiB
C++
1037 lines
25 KiB
C++
#include <qpdf/QPDFTokenizer.hh>
|
|
|
|
// DO NOT USE ctype -- it is locale dependent for some things, and
|
|
// it's not worth the risk of including it in case it may accidentally
|
|
// be used.
|
|
|
|
#include <qpdf/QIntC.hh>
|
|
#include <qpdf/QPDFExc.hh>
|
|
#include <qpdf/QPDFObjectHandle.hh>
|
|
#include <qpdf/QTC.hh>
|
|
#include <qpdf/QUtil.hh>
|
|
|
|
#include <stdexcept>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
|
|
static bool
|
|
is_delimiter(char ch)
|
|
{
|
|
return (strchr(" \t\n\v\f\r()<>[]{}/%", ch) != nullptr);
|
|
}
|
|
|
|
namespace
|
|
{
|
|
class QPDFWordTokenFinder: public InputSource::Finder
|
|
{
|
|
public:
|
|
QPDFWordTokenFinder(
|
|
std::shared_ptr<InputSource> is, std::string const& str) :
|
|
is(is),
|
|
str(str)
|
|
{
|
|
}
|
|
virtual ~QPDFWordTokenFinder() = default;
|
|
virtual bool check();
|
|
|
|
private:
|
|
std::shared_ptr<InputSource> is;
|
|
std::string str;
|
|
};
|
|
} // namespace
|
|
|
|
bool
|
|
QPDFWordTokenFinder::check()
|
|
{
|
|
// Find a word token matching the given string, preceded by a
|
|
// delimiter, and followed by a delimiter or EOF.
|
|
QPDFTokenizer tokenizer;
|
|
QPDFTokenizer::Token t = tokenizer.readToken(is, "finder", true);
|
|
qpdf_offset_t pos = is->tell();
|
|
if (!(t == QPDFTokenizer::Token(QPDFTokenizer::tt_word, str))) {
|
|
QTC::TC("qpdf", "QPDFTokenizer finder found wrong word");
|
|
return false;
|
|
}
|
|
qpdf_offset_t token_start = is->getLastOffset();
|
|
char next;
|
|
bool next_okay = false;
|
|
if (is->read(&next, 1) == 0) {
|
|
QTC::TC("qpdf", "QPDFTokenizer inline image at EOF");
|
|
next_okay = true;
|
|
} else {
|
|
next_okay = is_delimiter(next);
|
|
}
|
|
is->seek(pos, SEEK_SET);
|
|
if (!next_okay) {
|
|
return false;
|
|
}
|
|
if (token_start == 0) {
|
|
// Can't actually happen...we never start the search at the
|
|
// beginning of the input.
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::reset()
|
|
{
|
|
state = st_before_token;
|
|
type = tt_bad;
|
|
val.clear();
|
|
raw_val.clear();
|
|
error_message = "";
|
|
before_token = true;
|
|
in_token = false;
|
|
char_to_unread = '\0';
|
|
inline_image_bytes = 0;
|
|
string_depth = 0;
|
|
bad = false;
|
|
}
|
|
|
|
QPDFTokenizer::Token::Token(token_type_e type, std::string const& value) :
|
|
type(type),
|
|
value(value),
|
|
raw_value(value)
|
|
{
|
|
if (type == tt_string) {
|
|
raw_value = QPDFObjectHandle::newString(value).unparse();
|
|
} else if (type == tt_name) {
|
|
raw_value = QPDFObjectHandle::newName(value).unparse();
|
|
}
|
|
}
|
|
|
|
QPDFTokenizer::QPDFTokenizer() :
|
|
allow_eof(false),
|
|
include_ignorable(false)
|
|
{
|
|
reset();
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::allowEOF()
|
|
{
|
|
this->allow_eof = true;
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::includeIgnorable()
|
|
{
|
|
this->include_ignorable = true;
|
|
}
|
|
|
|
bool
|
|
QPDFTokenizer::isSpace(char ch)
|
|
{
|
|
return ((ch == '\0') || QUtil::is_space(ch));
|
|
}
|
|
|
|
bool
|
|
QPDFTokenizer::isDelimiter(char ch)
|
|
{
|
|
return is_delimiter(ch);
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::presentCharacter(char ch)
|
|
{
|
|
handleCharacter(ch);
|
|
|
|
if (this->in_token) {
|
|
this->raw_val += ch;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::handleCharacter(char ch)
|
|
{
|
|
// State machine is implemented such that the final character may not be
|
|
// handled. This happens whenever you have to use a character from the
|
|
// next token to detect the end of the current token.
|
|
|
|
switch (this->state) {
|
|
case st_top:
|
|
inTop(ch);
|
|
return;
|
|
|
|
case st_in_space:
|
|
inSpace(ch);
|
|
return;
|
|
|
|
case st_in_comment:
|
|
inComment(ch);
|
|
return;
|
|
|
|
case st_lt:
|
|
inLt(ch);
|
|
return;
|
|
|
|
case st_gt:
|
|
inGt(ch);
|
|
return;
|
|
|
|
case st_in_string:
|
|
inString(ch);
|
|
return;
|
|
|
|
case st_name:
|
|
inName(ch);
|
|
return;
|
|
|
|
case st_number:
|
|
inNumber(ch);
|
|
return;
|
|
|
|
case st_real:
|
|
inReal(ch);
|
|
return;
|
|
|
|
case st_string_after_cr:
|
|
inStringAfterCR(ch);
|
|
return;
|
|
|
|
case st_string_escape:
|
|
inStringEscape(ch);
|
|
return;
|
|
|
|
case st_char_code:
|
|
inCharCode(ch);
|
|
return;
|
|
|
|
case st_literal:
|
|
inLiteral(ch);
|
|
return;
|
|
|
|
case st_inline_image:
|
|
inInlineImage(ch);
|
|
return;
|
|
|
|
case st_in_hexstring:
|
|
inHexstring(ch);
|
|
return;
|
|
|
|
case st_in_hexstring_2nd:
|
|
inHexstring2nd(ch);
|
|
return;
|
|
|
|
case st_name_hex1:
|
|
inNameHex1(ch);
|
|
return;
|
|
|
|
case st_name_hex2:
|
|
inNameHex2(ch);
|
|
return;
|
|
|
|
case st_sign:
|
|
inSign(ch);
|
|
return;
|
|
|
|
case st_decimal:
|
|
inDecimal(ch);
|
|
return;
|
|
|
|
case (st_before_token):
|
|
inBeforeToken(ch);
|
|
return;
|
|
|
|
case (st_token_ready):
|
|
inTokenReady(ch);
|
|
return;
|
|
|
|
default:
|
|
throw std::logic_error(
|
|
"INTERNAL ERROR: invalid state while reading token");
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inTokenReady(char ch)
|
|
{
|
|
throw std::logic_error("INTERNAL ERROR: QPDF tokenizer presented character "
|
|
"while token is waiting");
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inBeforeToken(char ch)
|
|
{
|
|
// Note: we specifically do not use ctype here. It is
|
|
// locale-dependent.
|
|
if (isSpace(ch)) {
|
|
this->before_token = !this->include_ignorable;
|
|
this->in_token = this->include_ignorable;
|
|
if (this->include_ignorable) {
|
|
this->state = st_in_space;
|
|
this->val += ch;
|
|
}
|
|
} else if (ch == '%') {
|
|
this->before_token = !this->include_ignorable;
|
|
this->in_token = this->include_ignorable;
|
|
this->state = st_in_comment;
|
|
if (this->include_ignorable) {
|
|
this->val += ch;
|
|
}
|
|
} else {
|
|
this->before_token = false;
|
|
this->in_token = true;
|
|
inTop(ch);
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inTop(char ch)
|
|
{
|
|
switch (ch) {
|
|
case '(':
|
|
this->string_depth = 1;
|
|
this->state = st_in_string;
|
|
return;
|
|
|
|
case '<':
|
|
this->state = st_lt;
|
|
return;
|
|
|
|
case '>':
|
|
this->state = st_gt;
|
|
return;
|
|
|
|
case (')'):
|
|
this->type = tt_bad;
|
|
QTC::TC("qpdf", "QPDFTokenizer bad )");
|
|
this->error_message = "unexpected )";
|
|
this->val += ch;
|
|
this->state = st_token_ready;
|
|
return;
|
|
|
|
case '[':
|
|
this->type = tt_array_open;
|
|
this->state = st_token_ready;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case ']':
|
|
this->type = tt_array_close;
|
|
this->val += ch;
|
|
this->state = st_token_ready;
|
|
return;
|
|
|
|
case '{':
|
|
this->type = tt_brace_open;
|
|
this->state = st_token_ready;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '}':
|
|
this->type = tt_brace_close;
|
|
this->state = st_token_ready;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '/':
|
|
this->state = st_name;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '0':
|
|
case '1':
|
|
case '2':
|
|
case '3':
|
|
case '4':
|
|
case '5':
|
|
case '6':
|
|
case '7':
|
|
case '8':
|
|
case '9':
|
|
this->state = st_number;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '+':
|
|
case '-':
|
|
this->state = st_sign;
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '.':
|
|
this->state = st_decimal;
|
|
this->val += ch;
|
|
return;
|
|
|
|
default:
|
|
this->state = st_literal;
|
|
this->val += ch;
|
|
return;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inSpace(char ch)
|
|
{
|
|
// We only enter this state if include_ignorable is true.
|
|
if (!isSpace(ch)) {
|
|
this->type = tt_space;
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
this->state = st_token_ready;
|
|
return;
|
|
} else {
|
|
this->val += ch;
|
|
return;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inComment(char ch)
|
|
{
|
|
if ((ch == '\r') || (ch == '\n')) {
|
|
if (this->include_ignorable) {
|
|
this->type = tt_comment;
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
this->state = st_token_ready;
|
|
} else {
|
|
this->state = st_before_token;
|
|
}
|
|
} else if (this->include_ignorable) {
|
|
this->val += ch;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inString(char ch)
|
|
{
|
|
switch (ch) {
|
|
case '\\':
|
|
this->state = st_string_escape;
|
|
return;
|
|
|
|
case '(':
|
|
this->val += ch;
|
|
++this->string_depth;
|
|
return;
|
|
|
|
case ')':
|
|
if (--this->string_depth == 0) {
|
|
this->type = tt_string;
|
|
this->state = st_token_ready;
|
|
return;
|
|
}
|
|
|
|
this->val += ch;
|
|
return;
|
|
|
|
case '\r':
|
|
// CR by itself is converted to LF
|
|
this->val += '\n';
|
|
this->state = st_string_after_cr;
|
|
return;
|
|
|
|
case '\n':
|
|
this->val += ch;
|
|
return;
|
|
|
|
default:
|
|
this->val += ch;
|
|
return;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inName(char ch)
|
|
{
|
|
if (isDelimiter(ch)) {
|
|
// A C-locale whitespace character or delimiter terminates
|
|
// token. It is important to unread the whitespace
|
|
// character even though it is ignored since it may be the
|
|
// newline after a stream keyword. Removing it here could
|
|
// make the stream-reading code break on some files,
|
|
// though not on any files in the test suite as of this
|
|
// writing.
|
|
|
|
this->type = this->bad ? tt_bad : tt_name;
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
this->state = st_token_ready;
|
|
} else if (ch == '#') {
|
|
this->char_code = 0;
|
|
this->state = st_name_hex1;
|
|
} else {
|
|
this->val += ch;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inNameHex1(char ch)
|
|
{
|
|
this->hex_char = ch;
|
|
|
|
if ('0' <= ch && ch <= '9') {
|
|
this->char_code = 16 * (int(ch) - int('0'));
|
|
this->state = st_name_hex2;
|
|
|
|
} else if ('A' <= ch && ch <= 'F') {
|
|
this->char_code = 16 * (10 + int(ch) - int('A'));
|
|
this->state = st_name_hex2;
|
|
|
|
} else if ('a' <= ch && ch <= 'f') {
|
|
this->char_code = 16 * (10 + int(ch) - int('a'));
|
|
this->state = st_name_hex2;
|
|
|
|
} else {
|
|
QTC::TC("qpdf", "QPDFTokenizer bad name 1");
|
|
this->error_message = "name with stray # will not work with PDF >= 1.2";
|
|
// Use null to encode a bad # -- this is reversed
|
|
// in QPDF_Name::normalizeName.
|
|
this->val += '\0';
|
|
this->state = st_name;
|
|
inName(ch);
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inNameHex2(char ch)
|
|
{
|
|
if ('0' <= ch && ch <= '9') {
|
|
this->char_code += int(ch) - int('0');
|
|
|
|
} else if ('A' <= ch && ch <= 'F') {
|
|
this->char_code += 10 + int(ch) - int('A');
|
|
|
|
} else if ('a' <= ch && ch <= 'f') {
|
|
this->char_code += 10 + int(ch) - int('a');
|
|
|
|
} else {
|
|
QTC::TC("qpdf", "QPDFTokenizer bad name 2");
|
|
this->error_message = "name with stray # will not work with PDF >= 1.2";
|
|
// Use null to encode a bad # -- this is reversed
|
|
// in QPDF_Name::normalizeName.
|
|
this->val += '\0';
|
|
this->val += this->hex_char;
|
|
this->state = st_name;
|
|
inName(ch);
|
|
return;
|
|
}
|
|
if (this->char_code == 0) {
|
|
QTC::TC("qpdf", "QPDFTokenizer null in name");
|
|
this->error_message = "null character not allowed in name token";
|
|
this->val += "#00";
|
|
this->state = st_name;
|
|
this->bad = true;
|
|
} else {
|
|
this->val += char(this->char_code);
|
|
this->state = st_name;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inSign(char ch)
|
|
{
|
|
if (QUtil::is_digit(ch)) {
|
|
this->state = st_number;
|
|
this->val += ch;
|
|
} else if (ch == '.') {
|
|
this->state = st_decimal;
|
|
this->val += ch;
|
|
} else {
|
|
this->state = st_literal;
|
|
inLiteral(ch);
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inDecimal(char ch)
|
|
{
|
|
if (QUtil::is_digit(ch)) {
|
|
this->state = st_real;
|
|
this->val += ch;
|
|
} else {
|
|
this->state = st_literal;
|
|
inLiteral(ch);
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inNumber(char ch)
|
|
{
|
|
if (QUtil::is_digit(ch)) {
|
|
this->val += ch;
|
|
} else if (ch == '.') {
|
|
this->state = st_real;
|
|
this->val += ch;
|
|
} else if (isDelimiter(ch)) {
|
|
this->type = tt_integer;
|
|
this->state = st_token_ready;
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
} else {
|
|
this->state = st_literal;
|
|
this->val += ch;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inReal(char ch)
|
|
{
|
|
if (QUtil::is_digit(ch)) {
|
|
this->val += ch;
|
|
} else if (isDelimiter(ch)) {
|
|
this->type = tt_real;
|
|
this->state = st_token_ready;
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
} else {
|
|
this->state = st_literal;
|
|
this->val += ch;
|
|
}
|
|
}
|
|
void
|
|
QPDFTokenizer::inStringEscape(char ch)
|
|
{
|
|
this->state = st_in_string;
|
|
switch (ch) {
|
|
case '0':
|
|
case '1':
|
|
case '2':
|
|
case '3':
|
|
case '4':
|
|
case '5':
|
|
case '6':
|
|
case '7':
|
|
this->state = st_char_code;
|
|
this->char_code = 0;
|
|
this->digit_count = 0;
|
|
inCharCode(ch);
|
|
return;
|
|
|
|
case 'n':
|
|
this->val += '\n';
|
|
return;
|
|
|
|
case 'r':
|
|
this->val += '\r';
|
|
return;
|
|
|
|
case 't':
|
|
this->val += '\t';
|
|
return;
|
|
|
|
case 'b':
|
|
this->val += '\b';
|
|
return;
|
|
|
|
case 'f':
|
|
this->val += '\f';
|
|
return;
|
|
|
|
case '\n':
|
|
return;
|
|
|
|
case '\r':
|
|
this->state = st_string_after_cr;
|
|
return;
|
|
|
|
default:
|
|
// PDF spec says backslash is ignored before anything else
|
|
this->val += ch;
|
|
return;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inStringAfterCR(char ch)
|
|
{
|
|
this->state = st_in_string;
|
|
if (ch != '\n') {
|
|
inString(ch);
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inLt(char ch)
|
|
{
|
|
if (ch == '<') {
|
|
this->val += "<<";
|
|
this->type = tt_dict_open;
|
|
this->state = st_token_ready;
|
|
return;
|
|
}
|
|
|
|
this->state = st_in_hexstring;
|
|
inHexstring(ch);
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inGt(char ch)
|
|
{
|
|
if (ch == '>') {
|
|
this->val += ">>";
|
|
this->type = tt_dict_close;
|
|
this->state = st_token_ready;
|
|
} else {
|
|
this->val += ">";
|
|
this->type = tt_bad;
|
|
QTC::TC("qpdf", "QPDFTokenizer bad >");
|
|
this->error_message = "unexpected >";
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
this->state = st_token_ready;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inLiteral(char ch)
|
|
{
|
|
if (isDelimiter(ch)) {
|
|
// A C-locale whitespace character or delimiter terminates
|
|
// token. It is important to unread the whitespace
|
|
// character even though it is ignored since it may be the
|
|
// newline after a stream keyword. Removing it here could
|
|
// make the stream-reading code break on some files,
|
|
// though not on any files in the test suite as of this
|
|
// writing.
|
|
|
|
this->in_token = false;
|
|
this->char_to_unread = ch;
|
|
this->state = st_token_ready;
|
|
this->type = (this->val == "true") || (this->val == "false")
|
|
? tt_bool
|
|
: (this->val == "null" ? tt_null : tt_word);
|
|
} else {
|
|
this->val += ch;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inHexstring(char ch)
|
|
{
|
|
if ('0' <= ch && ch <= '9') {
|
|
this->char_code = 16 * (int(ch) - int('0'));
|
|
this->state = st_in_hexstring_2nd;
|
|
|
|
} else if ('A' <= ch && ch <= 'F') {
|
|
this->char_code = 16 * (10 + int(ch) - int('A'));
|
|
this->state = st_in_hexstring_2nd;
|
|
|
|
} else if ('a' <= ch && ch <= 'f') {
|
|
this->char_code = 16 * (10 + int(ch) - int('a'));
|
|
this->state = st_in_hexstring_2nd;
|
|
|
|
} else if (ch == '>') {
|
|
this->type = tt_string;
|
|
this->state = st_token_ready;
|
|
|
|
} else if (isSpace(ch)) {
|
|
// ignore
|
|
|
|
} else {
|
|
this->type = tt_bad;
|
|
QTC::TC("qpdf", "QPDFTokenizer bad hexstring character");
|
|
this->error_message =
|
|
std::string("invalid character (") + ch + ") in hexstring";
|
|
this->state = st_token_ready;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inHexstring2nd(char ch)
|
|
{
|
|
if ('0' <= ch && ch <= '9') {
|
|
this->val += char(this->char_code + int(ch) - int('0'));
|
|
this->state = st_in_hexstring;
|
|
|
|
} else if ('A' <= ch && ch <= 'F') {
|
|
this->val += char(this->char_code + 10 + int(ch) - int('A'));
|
|
this->state = st_in_hexstring;
|
|
|
|
} else if ('a' <= ch && ch <= 'f') {
|
|
this->val += char(this->char_code + 10 + int(ch) - int('a'));
|
|
this->state = st_in_hexstring;
|
|
|
|
} else if (ch == '>') {
|
|
// PDF spec says odd hexstrings have implicit trailing 0.
|
|
this->val += char(this->char_code);
|
|
this->type = tt_string;
|
|
this->state = st_token_ready;
|
|
|
|
} else if (isSpace(ch)) {
|
|
// ignore
|
|
|
|
} else {
|
|
this->type = tt_bad;
|
|
QTC::TC("qpdf", "QPDFTokenizer bad hexstring 2nd character");
|
|
this->error_message =
|
|
std::string("invalid character (") + ch + ") in hexstring";
|
|
this->state = st_token_ready;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inCharCode(char ch)
|
|
{
|
|
if (('0' <= ch) && (ch <= '7')) {
|
|
this->char_code = 8 * this->char_code + (int(ch) - int('0'));
|
|
if (++(this->digit_count) < 3) {
|
|
return;
|
|
}
|
|
// We've accumulated \ddd. PDF Spec says to ignore
|
|
// high-order overflow.
|
|
}
|
|
this->val += char(this->char_code % 256);
|
|
this->state = st_in_string;
|
|
return;
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::inInlineImage(char ch)
|
|
{
|
|
this->val += ch;
|
|
if (this->val.length() == this->inline_image_bytes) {
|
|
QTC::TC("qpdf", "QPDFTokenizer found EI by byte count");
|
|
this->type = tt_inline_image;
|
|
this->inline_image_bytes = 0;
|
|
this->state = st_token_ready;
|
|
}
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::presentEOF()
|
|
{
|
|
switch (this->state) {
|
|
case st_name:
|
|
case st_name_hex1:
|
|
case st_name_hex2:
|
|
case st_number:
|
|
case st_real:
|
|
case st_sign:
|
|
case st_decimal:
|
|
case st_literal:
|
|
QTC::TC("qpdf", "QPDFTokenizer EOF reading appendable token");
|
|
// Push any delimiter to the state machine to finish off the final
|
|
// token.
|
|
presentCharacter('\f');
|
|
this->in_token = true;
|
|
break;
|
|
|
|
case st_top:
|
|
case st_before_token:
|
|
this->type = tt_eof;
|
|
break;
|
|
|
|
case st_in_space:
|
|
this->type = this->include_ignorable ? tt_space : tt_eof;
|
|
break;
|
|
|
|
case st_in_comment:
|
|
this->type = this->include_ignorable ? tt_comment : tt_bad;
|
|
break;
|
|
|
|
case st_token_ready:
|
|
break;
|
|
|
|
default:
|
|
QTC::TC("qpdf", "QPDFTokenizer EOF reading token");
|
|
this->type = tt_bad;
|
|
this->error_message = "EOF while reading token";
|
|
}
|
|
this->state = st_token_ready;
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::expectInlineImage(std::shared_ptr<InputSource> input)
|
|
{
|
|
if (this->state != st_before_token) {
|
|
throw std::logic_error("QPDFTokenizer::expectInlineImage called"
|
|
" when tokenizer is in improper state");
|
|
}
|
|
findEI(input);
|
|
this->before_token = false;
|
|
this->in_token = true;
|
|
this->state = st_inline_image;
|
|
}
|
|
|
|
void
|
|
QPDFTokenizer::findEI(std::shared_ptr<InputSource> input)
|
|
{
|
|
if (!input.get()) {
|
|
return;
|
|
}
|
|
|
|
qpdf_offset_t last_offset = input->getLastOffset();
|
|
qpdf_offset_t pos = input->tell();
|
|
|
|
// Use QPDFWordTokenFinder to find EI surrounded by delimiters.
|
|
// Then read the next several tokens or up to EOF. If we find any
|
|
// suspicious-looking or tokens, this is probably still part of
|
|
// the image data, so keep looking for EI. Stop at the first EI
|
|
// that passes. If we get to the end without finding one, return
|
|
// the last EI we found. Store the number of bytes expected in the
|
|
// inline image including the EI and use that to break out of
|
|
// inline image, falling back to the old method if needed.
|
|
|
|
bool okay = false;
|
|
bool first_try = true;
|
|
while (!okay) {
|
|
QPDFWordTokenFinder f(input, "EI");
|
|
if (!input->findFirst("EI", input->tell(), 0, f)) {
|
|
break;
|
|
}
|
|
this->inline_image_bytes = QIntC::to_size(input->tell() - pos - 2);
|
|
|
|
QPDFTokenizer check;
|
|
bool found_bad = false;
|
|
// Look at the next 10 tokens or up to EOF. The next inline
|
|
// image's image data would look like bad tokens, but there
|
|
// will always be at least 10 tokens between one inline
|
|
// image's EI and the next valid one's ID since width, height,
|
|
// bits per pixel, and color space are all required as well as
|
|
// a BI and ID. If we get 10 good tokens in a row or hit EOF,
|
|
// we can be pretty sure we've found the actual EI.
|
|
for (int i = 0; i < 10; ++i) {
|
|
QPDFTokenizer::Token t = check.readToken(input, "checker", true);
|
|
token_type_e type = t.getType();
|
|
if (type == tt_eof) {
|
|
okay = true;
|
|
} else if (type == tt_bad) {
|
|
found_bad = true;
|
|
} else if (t.isWord()) {
|
|
// The qpdf tokenizer lumps alphabetic and otherwise
|
|
// uncategorized characters into "words". We recognize
|
|
// strings of alphabetic characters as potential valid
|
|
// operators for purposes of telling whether we're in
|
|
// valid content or not. It's not perfect, but it
|
|
// should work more reliably than what we used to do,
|
|
// which was already good enough for the vast majority
|
|
// of files.
|
|
bool found_alpha = false;
|
|
bool found_non_printable = false;
|
|
bool found_other = false;
|
|
for (char ch: t.getValue()) {
|
|
if (((ch >= 'a') && (ch <= 'z')) ||
|
|
((ch >= 'A') && (ch <= 'Z')) || (ch == '*')) {
|
|
// Treat '*' as alpha since there are valid
|
|
// PDF operators that contain * along with
|
|
// alphabetic characters.
|
|
found_alpha = true;
|
|
} else if (
|
|
(static_cast<signed char>(ch) < 32) && (!isSpace(ch))) {
|
|
// Compare ch as a signed char so characters
|
|
// outside of 7-bit will be < 0.
|
|
found_non_printable = true;
|
|
break;
|
|
} else {
|
|
found_other = true;
|
|
}
|
|
}
|
|
if (found_non_printable || (found_alpha && found_other)) {
|
|
found_bad = true;
|
|
}
|
|
}
|
|
if (okay || found_bad) {
|
|
break;
|
|
}
|
|
}
|
|
if (!found_bad) {
|
|
okay = true;
|
|
}
|
|
if (!okay) {
|
|
first_try = false;
|
|
}
|
|
}
|
|
if (okay && (!first_try)) {
|
|
QTC::TC("qpdf", "QPDFTokenizer found EI after more than one try");
|
|
}
|
|
|
|
input->seek(pos, SEEK_SET);
|
|
input->setLastOffset(last_offset);
|
|
}
|
|
|
|
bool
|
|
QPDFTokenizer::getToken(Token& token, bool& unread_char, char& ch)
|
|
{
|
|
bool ready = (this->state == st_token_ready);
|
|
unread_char = !this->in_token && !this->before_token;
|
|
ch = this->char_to_unread;
|
|
if (ready) {
|
|
token = (this->type == tt_bad)
|
|
? Token(
|
|
this->type, this->raw_val, this->raw_val, this->error_message)
|
|
: Token(this->type, this->val, this->raw_val, this->error_message);
|
|
|
|
this->reset();
|
|
}
|
|
return ready;
|
|
}
|
|
|
|
bool
|
|
QPDFTokenizer::betweenTokens()
|
|
{
|
|
return this->before_token;
|
|
}
|
|
|
|
QPDFTokenizer::Token
|
|
QPDFTokenizer::readToken(
|
|
std::shared_ptr<InputSource> input,
|
|
std::string const& context,
|
|
bool allow_bad,
|
|
size_t max_len)
|
|
{
|
|
qpdf_offset_t offset = input->fastTell();
|
|
|
|
while (this->state != st_token_ready) {
|
|
char ch;
|
|
if (!input->fastRead(ch)) {
|
|
presentEOF();
|
|
|
|
if ((this->type == tt_eof) && (!this->allow_eof)) {
|
|
// Nothing in the qpdf library calls readToken
|
|
// without allowEOF anymore, so this case is not
|
|
// exercised.
|
|
this->type = tt_bad;
|
|
this->error_message = "unexpected EOF";
|
|
offset = input->getLastOffset();
|
|
}
|
|
} else {
|
|
handleCharacter(ch);
|
|
if (this->before_token) {
|
|
++offset;
|
|
}
|
|
if (this->in_token) {
|
|
this->raw_val += ch;
|
|
}
|
|
if (max_len && (this->raw_val.length() >= max_len) &&
|
|
(this->state != st_token_ready)) {
|
|
// terminate this token now
|
|
QTC::TC("qpdf", "QPDFTokenizer block long token");
|
|
this->type = tt_bad;
|
|
this->state = st_token_ready;
|
|
this->error_message =
|
|
"exceeded allowable length while reading token";
|
|
}
|
|
}
|
|
}
|
|
|
|
Token token;
|
|
bool unread_char;
|
|
char char_to_unread;
|
|
getToken(token, unread_char, char_to_unread);
|
|
input->fastUnread(unread_char);
|
|
|
|
if (token.getType() != tt_eof) {
|
|
input->setLastOffset(offset);
|
|
}
|
|
|
|
if (token.getType() == tt_bad) {
|
|
if (allow_bad) {
|
|
QTC::TC("qpdf", "QPDFTokenizer allowing bad token");
|
|
} else {
|
|
throw QPDFExc(
|
|
qpdf_e_damaged_pdf,
|
|
input->getName(),
|
|
context,
|
|
offset,
|
|
token.getErrorMessage());
|
|
}
|
|
}
|
|
|
|
return token;
|
|
}
|