octoleo/qpdf
2
1
Fork 0
mirror of https://github.com/qpdf/qpdf.git synced 2024-11-02 03:42:30 +00:00
Code Issues Releases Wiki Activity

Optimise QPDFParser::parse for #311 problem

Browse Source 
Avoid creating new null objects that later will be discarded and made
implicit.

Part of #729
This commit is contained in:
m-holger 2022-08-30 13:32:54 +01:00
parent 97a7ad1d80
commit 70d985f942
1 changed files with 20 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
libqpdf/QPDFParser.cc
View File

@ -51,9 +51,12 @@ QPDFParser::parse(bool& empty, bool content_stream)
int bad_count = 0; int bad_count = 0;
int good_count = 0; int good_count = 0;
bool b_contents = false; bool b_contents = false;
bool is_null = false;
auto null_oh = QPDFObjectHandle::newNull();
while (!done) { while (!done) {
bool bad = false; bool bad = false;
is_null = false;
auto& frame = stack.back(); auto& frame = stack.back();
auto& olist = frame.olist; auto& olist = frame.olist;
parser_state_e state = state_stack.back(); parser_state_e state = state_stack.back();
@ -83,7 +86,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
case QPDFTokenizer::tt_bad: case QPDFTokenizer::tt_bad:
QTC::TC("qpdf", "QPDFParser bad token in parse"); QTC::TC("qpdf", "QPDFParser bad token in parse");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
break; break;
case QPDFTokenizer::tt_brace_open: case QPDFTokenizer::tt_brace_open:
@ -91,7 +94,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
QTC::TC("qpdf", "QPDFParser bad brace"); QTC::TC("qpdf", "QPDFParser bad brace");
warn("treating unexpected brace token as null"); warn("treating unexpected brace token as null");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
break; break;
case QPDFTokenizer::tt_array_close: case QPDFTokenizer::tt_array_close:
@ -101,7 +104,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
QTC::TC("qpdf", "QPDFParser bad array close"); QTC::TC("qpdf", "QPDFParser bad array close");
warn("treating unexpected array close token as null"); warn("treating unexpected array close token as null");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
} }
break; break;
@ -112,7 +115,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
QTC::TC("qpdf", "QPDFParser bad dictionary close"); QTC::TC("qpdf", "QPDFParser bad dictionary close");
warn("unexpected dictionary close token"); warn("unexpected dictionary close token");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
} }
break; break;
@ -122,7 +125,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
QTC::TC("qpdf", "QPDFParser too deep"); QTC::TC("qpdf", "QPDFParser too deep");
warn("ignoring excessively deeply nested data structure"); warn("ignoring excessively deeply nested data structure");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
state = st_top; state = st_top;
} else { } else {
state = st_start; state = st_start;
@ -140,7 +143,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
break; break;
case QPDFTokenizer::tt_null: case QPDFTokenizer::tt_null:
object = QPDFObjectHandle::newNull(); is_null = true;
break; break;
case QPDFTokenizer::tt_integer: case QPDFTokenizer::tt_integer:
@ -195,7 +198,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
// We just saw endobj without having read // We just saw endobj without having read
// anything. Treat this as a null and do not move // anything. Treat this as a null and do not move
// the input source's offset. // the input source's offset.
object = QPDFObjectHandle::newNull(); is_null = true;
input->seek(input->getLastOffset(), SEEK_SET); input->seek(input->getLastOffset(), SEEK_SET);
empty = true; empty = true;
} else { } else {
@ -228,16 +231,16 @@ QPDFParser::parse(bool& empty, bool content_stream)
warn("treating unknown token type as null while " warn("treating unknown token type as null while "
"reading object"); "reading object");
bad = true; bad = true;
object = QPDFObjectHandle::newNull(); is_null = true;
break; break;
} }
if ((!object.isInitialized()) && if (!object.isInitialized() && !is_null &&
(!((state == st_start) || (state == st_stop) || (!((state == st_start) || (state == st_stop) ||
(state == st_eof)))) { (state == st_eof)))) {
throw std::logic_error("QPDFObjectHandle::parseInternal: " throw std::logic_error("QPDFObjectHandle::parseInternal: "
"unexpected uninitialized object"); "unexpected uninitialized object");
object = QPDFObjectHandle::newNull(); is_null = true;
} }
if (bad) { if (bad) {
@ -254,7 +257,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
// intervening successful objects. Give up. // intervening successful objects. Give up.
warn("too many errors; giving up on reading object"); warn("too many errors; giving up on reading object");
state = st_top; state = st_top;
object = QPDFObjectHandle::newNull(); is_null = true;
} }
switch (state) { switch (state) {
@ -266,7 +269,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
// In content stream mode, leave object uninitialized to // In content stream mode, leave object uninitialized to
// indicate EOF // indicate EOF
if (!content_stream) { if (!content_stream) {
object = QPDFObjectHandle::newNull(); is_null = true;
} }
break; break;
@ -279,7 +282,7 @@ QPDFParser::parse(bool& empty, bool content_stream)
object.setParsedOffset(input->getLastOffset()); object.setParsedOffset(input->getLastOffset());
} }
set_offset = true; set_offset = true;
olist.push_back(object); olist.push_back(is_null ? null_oh : object);
break; break;
case st_top: case st_top:
@ -387,11 +390,14 @@ QPDFParser::parse(bool& empty, bool content_stream)
if (state_stack.back() == st_top) { if (state_stack.back() == st_top) {
done = true; done = true;
} else { } else {
stack.back().olist.push_back(object); stack.back().olist.push_back(is_null ? null_oh : object);
} }
} }
} }
if (is_null) {
object = QPDFObjectHandle::newNull();
}
if (!set_offset) { if (!set_offset) {
setDescriptionFromInput(object, offset); setDescriptionFromInput(object, offset);
object.setParsedOffset(offset); object.setParsedOffset(offset);