mirror of
https://github.com/qpdf/qpdf.git
synced 2024-06-02 02:10:52 +00:00
Add private method QPDF::validateStreamLineEnd
This commit is contained in:
parent
ef866d68cd
commit
aabd3f6f9b
|
@ -1009,6 +1009,7 @@ class QPDF
|
||||||
QPDFObjectHandle readTrailer();
|
QPDFObjectHandle readTrailer();
|
||||||
QPDFObjectHandle readObject(std::string const& description, QPDFObjGen og);
|
QPDFObjectHandle readObject(std::string const& description, QPDFObjGen og);
|
||||||
void readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset);
|
void readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset);
|
||||||
|
void validateStreamLineEnd(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset);
|
||||||
QPDFObjectHandle readObjectInStream(std::shared_ptr<InputSource>, QPDFObjGen og);
|
QPDFObjectHandle readObjectInStream(std::shared_ptr<InputSource>, QPDFObjGen og);
|
||||||
size_t recoverStreamLength(
|
size_t recoverStreamLength(
|
||||||
std::shared_ptr<InputSource> input, QPDFObjGen const& og, qpdf_offset_t stream_offset);
|
std::shared_ptr<InputSource> input, QPDFObjGen const& og, qpdf_offset_t stream_offset);
|
||||||
|
|
|
@ -1324,51 +1324,7 @@ QPDF::readObject(std::string const& description, QPDFObjGen og)
|
||||||
void
|
void
|
||||||
QPDF::readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
|
QPDF::readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
|
||||||
{
|
{
|
||||||
// The PDF specification states that the word "stream" should be followed by either a carriage
|
validateStreamLineEnd(object, og, offset);
|
||||||
// return and a newline or by a newline alone. It specifically disallowed following it by a
|
|
||||||
// carriage return alone since, in that case, there would be no way to tell whether the NL in a
|
|
||||||
// CR NL sequence was part of the stream data. However, some readers, including Adobe reader,
|
|
||||||
// accept a carriage return by itself when followed by a non-newline character, so that's what
|
|
||||||
// we do here. We have also seen files that have extraneous whitespace between the stream
|
|
||||||
// keyword and the newline.
|
|
||||||
while (true) {
|
|
||||||
char ch;
|
|
||||||
if (m->file->read(&ch, 1) == 0) {
|
|
||||||
// A premature EOF here will result in some other problem that will get reported at
|
|
||||||
// another time.
|
|
||||||
break;
|
|
||||||
}
|
|
||||||
if (ch == '\n') {
|
|
||||||
// ready to read stream data
|
|
||||||
QTC::TC("qpdf", "QPDF stream with NL only");
|
|
||||||
break;
|
|
||||||
}
|
|
||||||
if (ch == '\r') {
|
|
||||||
// Read another character
|
|
||||||
if (m->file->read(&ch, 1) != 0) {
|
|
||||||
if (ch == '\n') {
|
|
||||||
// Ready to read stream data
|
|
||||||
QTC::TC("qpdf", "QPDF stream with CRNL");
|
|
||||||
} else {
|
|
||||||
// Treat the \r by itself as the whitespace after endstream and start reading
|
|
||||||
// stream data in spite of not having seen a newline.
|
|
||||||
QTC::TC("qpdf", "QPDF stream with CR only");
|
|
||||||
m->file->unreadCh(ch);
|
|
||||||
warn(damagedPDF(
|
|
||||||
m->file->tell(), "stream keyword followed by carriage return only"));
|
|
||||||
}
|
|
||||||
}
|
|
||||||
break;
|
|
||||||
}
|
|
||||||
if (!QUtil::is_space(ch)) {
|
|
||||||
QTC::TC("qpdf", "QPDF stream without newline");
|
|
||||||
m->file->unreadCh(ch);
|
|
||||||
warn(damagedPDF(
|
|
||||||
m->file->tell(), "stream keyword not followed by proper line terminator"));
|
|
||||||
break;
|
|
||||||
}
|
|
||||||
warn(damagedPDF(m->file->tell(), "stream keyword followed by extraneous whitespace"));
|
|
||||||
}
|
|
||||||
|
|
||||||
// Must get offset before accessing any additional objects since resolving a previously
|
// Must get offset before accessing any additional objects since resolving a previously
|
||||||
// unresolved indirect object will change file position.
|
// unresolved indirect object will change file position.
|
||||||
|
@ -1406,6 +1362,56 @@ QPDF::readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
|
||||||
object = newIndirect(og, QPDF_Stream::create(this, og, object, stream_offset, length));
|
object = newIndirect(og, QPDF_Stream::create(this, og, object, stream_offset, length));
|
||||||
}
|
}
|
||||||
|
|
||||||
|
void
|
||||||
|
QPDF::validateStreamLineEnd(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
|
||||||
|
{
|
||||||
|
// The PDF specification states that the word "stream" should be followed by either a carriage
|
||||||
|
// return and a newline or by a newline alone. It specifically disallowed following it by a
|
||||||
|
// carriage return alone since, in that case, there would be no way to tell whether the NL in a
|
||||||
|
// CR NL sequence was part of the stream data. However, some readers, including Adobe reader,
|
||||||
|
// accept a carriage return by itself when followed by a non-newline character, so that's what
|
||||||
|
// we do here. We have also seen files that have extraneous whitespace between the stream
|
||||||
|
// keyword and the newline.
|
||||||
|
while (true) {
|
||||||
|
char ch;
|
||||||
|
if (m->file->read(&ch, 1) == 0) {
|
||||||
|
// A premature EOF here will result in some other problem that will get reported at
|
||||||
|
// another time.
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
if (ch == '\n') {
|
||||||
|
// ready to read stream data
|
||||||
|
QTC::TC("qpdf", "QPDF stream with NL only");
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
if (ch == '\r') {
|
||||||
|
// Read another character
|
||||||
|
if (m->file->read(&ch, 1) != 0) {
|
||||||
|
if (ch == '\n') {
|
||||||
|
// Ready to read stream data
|
||||||
|
QTC::TC("qpdf", "QPDF stream with CRNL");
|
||||||
|
} else {
|
||||||
|
// Treat the \r by itself as the whitespace after endstream and start reading
|
||||||
|
// stream data in spite of not having seen a newline.
|
||||||
|
QTC::TC("qpdf", "QPDF stream with CR only");
|
||||||
|
m->file->unreadCh(ch);
|
||||||
|
warn(damagedPDF(
|
||||||
|
m->file->tell(), "stream keyword followed by carriage return only"));
|
||||||
|
}
|
||||||
|
}
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
if (!QUtil::is_space(ch)) {
|
||||||
|
QTC::TC("qpdf", "QPDF stream without newline");
|
||||||
|
m->file->unreadCh(ch);
|
||||||
|
warn(damagedPDF(
|
||||||
|
m->file->tell(), "stream keyword not followed by proper line terminator"));
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
warn(damagedPDF(m->file->tell(), "stream keyword followed by extraneous whitespace"));
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
QPDFObjectHandle
|
QPDFObjectHandle
|
||||||
QPDF::readObjectInStream(std::shared_ptr<InputSource> input, QPDFObjGen og)
|
QPDF::readObjectInStream(std::shared_ptr<InputSource> input, QPDFObjGen og)
|
||||||
{
|
{
|
||||||
|
|
Loading…
Reference in New Issue
Block a user