octoleo/qpdf
2
1
Fork 0
mirror of https://github.com/qpdf/qpdf.git synced 2024-11-14 08:44:04 +00:00
Code Issues Releases Wiki Activity
1,767 Commits 16 Branches 95 Tags 40 MiB
df067c9ab6
Commit Graph

797 Commits

Author SHA1 Message Date
Jay Berkenbilt
09027344b9 Refactor: separate code that determines whether to filter a stream 2020-12-28 12:58:19 -05:00
Jay Berkenbilt
39bfa01307 Implement user-provided stream filters 
Refactor QPDF_Stream to use stream filter classes to handle supported
stream filters as well.
 2020-12-28 12:58:19 -05:00
Jay Berkenbilt
cc8895078a Add QPDFObjectHandle::makeDirect(bool allow_streams) 2020-12-26 08:48:18 -05:00
Jay Berkenbilt
573b6eb8b1 Provide qpdf write progress reporting from C API (fixes #487) 2020-12-20 14:43:24 -05:00
Jay Berkenbilt
2050977099 Add QPDFObjectHandle manipulation to C API 2020-11-28 19:48:07 -05:00
Jay Berkenbilt
78b9d6bfd4 Prepare 10.0.4 release 2020-11-21 13:50:02 -05:00
Jay Berkenbilt
bd79138c84 Treat direct page as runtime rather than logic error (fuzz issue 27393) 2020-11-11 09:50:43 -05:00
Jay Berkenbilt
47f4ebcdac Ignore unused field in xref entry, avoiding range error (fixes #482) 2020-11-04 07:46:46 -05:00
Jay Berkenbilt
fbe40b800d Prepare 10.0.3 release 2020-10-31 13:47:03 -04:00
Jay Berkenbilt
6971f78ff6 Fix stack overflow on direct root (fuzz issue 26761) 2020-10-31 13:10:39 -04:00
Jay Berkenbilt
ffe6af6f77 Add comments explaining the foreign object copying code 
These are the comments I would have liked to have been able to read
while fixing #449 and #478.
 2020-10-31 12:14:26 -04:00
Jay Berkenbilt
96767fb104 Fix foreign stream copying bug (fixes #478) 
This reverts an incorrect fix to #449 and codes it properly. The real
problem was that we were looking at the local dictionaries rather than
the foreign dictionaries when saving the foreign stream data. In the
case of direct objects, these happened to be the same, but in the case
of indirect objects, the object references could be pointing anywhere
since object numbers don't match up between the old and new files.
 2020-10-31 12:14:26 -04:00
Jay Berkenbilt
da7540794a Prepare 10.0.2 release 2020-10-27 11:57:48 -04:00
Jay Berkenbilt
09bd1fafb1 Improve efficiency of number to string conversion 2020-10-27 11:57:48 -04:00
Jay Berkenbilt
bcea54fcaa Revert removal of unreadCh change for performance 
Turns out unreadCh is much more efficient than seek(-1, SEEK_CUR).
Update comments and code to reflect this.
 2020-10-27 11:57:48 -04:00
Jay Berkenbilt
b30deaeeab Avoid merging adjacent tokens when concatenating contents (fixes #444) 2020-10-23 08:00:04 -04:00
Jay Berkenbilt
8a11feacc3 Avoid leak by resolving object streams more than once (fuzz issue 23642) 2020-10-22 15:39:36 -04:00
Jay Berkenbilt
30bb4c64ee Minor code cleanup 
* Return rather than exiting from realmain in qpdf.cc
* Remove extraneous blank line
* Don't assign temporary to const reference
 2020-10-22 15:39:36 -04:00
Jay Berkenbilt
232f5fc9f3 Handle jpeg library fuzz false positives 
The jpeg library has some assembly code that is missed by the compiler
instrumentation used by memory sanitization. There is a runtime
environment variable that is used to work around this issue.
 2020-10-22 06:31:52 -04:00
Jay Berkenbilt
c1684eae91 Check for overflow in page labels (fuzz issue 23599) 2020-10-22 05:49:24 -04:00
Jay Berkenbilt
7f4a4df919 Add range_check method to QIntC 2020-10-22 05:48:40 -04:00
Jay Berkenbilt
24196c08cb Fix loop detection error (fuzz issue 23172) 2020-10-22 05:48:35 -04:00
Jay Berkenbilt
956c8f6432 Obscure bug fix copying foreign streams in special cases (fixes #449) 
Specifically, if a stream had its stream data replaced and had
indirect /Filter or /DecodeParms, it would result in non-silent loss
of data and/or internal error.
 2020-10-21 19:23:23 -04:00
Jay Berkenbilt
98f6c00dad Protect numeric conversion against user's locale (fixes #459) 2020-10-21 16:42:51 -04:00
Jay Berkenbilt
bed165c9fc Stop using InputSource::unreadCh 2020-10-18 07:43:05 -04:00
Dean Scarff
153060a0c5 Check integer overflow in resolveObjectsInStream 
Fixes a crash found by fuzzing.
 2020-10-16 20:09:24 -04:00
Dean Scarff
9a3791c53b Properly detect OPENSSL_IS_BORINGSSL 
OPENSSL_IS_BORINGSSL is not actually set by configure, so it will be
undefined until a BoringSSL header is included.  Hence the #ifdef logic
in QPDFCrypto_openssl.h would usually never apply.

This still worked because evp.h transitively included BoringSSL's
cipher.h and digest.h, but the latter are the correct (documented)
headers.

By re-ordering the includes, we can ensure the macro is defined when we
use it.

Also: fix case in the header guards.
 2020-10-16 20:04:36 -04:00
Dean Scarff
2ff84aa2c9 Include detailed OpenSSL error messages 
Fixes qpdf/qpdf#450
 2020-10-16 19:58:11 -04:00
James R. Barlow
3fc7c99d02 Replace memchr with manual memory search 
On large files with predominantly \n line endings, memchr(..'\r'..)
seems to waste a considerable amount of time searching for a line
ending candidate that we don't need.

On the Adobe PDF Reference Manual 1.7, this commit is 8x faster at
QPDF::processMemoryFile().
 2020-10-16 19:57:29 -04:00
oltolm
3221022fc9 fix WindowsCryptProvider fixes #432 2020-10-16 19:56:33 -04:00
Jay Berkenbilt
ff65e272a8 Fix printf formatting for newer msvc 
Use autoconf rather than ifdefs to determine what format string to use
for long long.
 2020-10-16 07:02:23 -04:00
Jay Berkenbilt
88b8f8ec86 Remove redundant check found by lgtm.com 2020-10-15 14:47:43 -04:00
Jay Berkenbilt
26514ab731 Write linearization errors to stderr (fixes #438) 2020-04-29 17:33:34 -04:00
Jay Berkenbilt
92d3cbecd4 Fix warnings reported by -Wshadow=local (fixes #431) 2020-04-16 12:41:43 -04:00
Jay Berkenbilt
578c5ac66c Use more references when iterating 
When possible, use `for (auto&` or `for (auto const&` when iterating
using C++-11 style iterators.
 2020-04-10 13:30:33 -04:00
Jay Berkenbilt
821a701851 Prepare 10.0.1 release 2020-04-09 11:48:26 -04:00
Jay Berkenbilt
1a7d3700a6 Fix unnecessary copies in auto iter (fixes #426) 
Also switch to colon-style iteration in some cases. Thanks to Dean
Scarff for drawing this to my attention after detecting some
unnecessary copies with
https://clang.llvm.org/extra/clang-tidy/checks/performance-for-range-copy.html
 2020-04-08 20:45:26 -04:00
Jay Berkenbilt
4977a7efa5 Bug fix: getStreamData should on unfilterable stream (fixes #425) 2020-04-08 18:52:04 -04:00
Jay Berkenbilt
1e629c278a Prepare 10.0.0 release 2020-04-06 11:30:15 -04:00
Jay Berkenbilt
c996f4ac33 Don't include <cwchar> if not building with wchar 2020-04-06 11:23:02 -04:00
Jay Berkenbilt
77198d5310 Delegate random number generation to crypto provider (fixes #418) 2020-04-06 11:23:02 -04:00
Jay Berkenbilt
52749b85df Make random data provider code thread-safe 
This uses C++-11 thread-safe static initializers now.
 2020-04-06 10:00:43 -04:00
Jay Berkenbilt
619d294e9d Remove QUtil::srandom 2020-04-06 09:49:02 -04:00
Dean Scarff
0f2507234f Add OpenSSL/BoringSSL crypto provider 
Fixes qpdf/qpdf#417
 2020-04-06 09:01:55 -04:00
Jay Berkenbilt
893d38b87e Allow propagation of errors and retry through StreamDataProvider 
StreamDataProvider::provideStreamData now has a rich enough API for it
to effectively proxy to pipeStreamData.
 2020-04-05 20:07:13 -04:00
Jay Berkenbilt
7246404177 JSON: implement pattern keys in schema 2020-04-04 18:06:32 -04:00
Dean Scarff
c5c1a028cd Use deterministic assignments for unique_id 
Fixes qpdf/qpdf#419
 2020-04-04 08:29:28 -04:00
Jay Berkenbilt
2100b4ce15 Allow qpdf to be built on systems without wchar_t (fixes #406) 2020-04-03 21:39:44 -04:00
Jay Berkenbilt
6a4117add9 Avoid potential segfault in warning methods 2020-04-03 21:39:20 -04:00
Jay Berkenbilt
4f3b89991b placeFormXObject: allow control of shrink/expand (fixes #409) 2020-04-03 21:39:17 -04:00
Jay Berkenbilt
b76b73b229 C API: accept any non-zero value as TRUE 2020-04-03 17:33:44 -04:00
Jay Berkenbilt
54726930df Remove redundant methods in QUtil 
This was being saved until we had to break ABI.
 2020-04-03 12:17:57 -04:00
Jay Berkenbilt
5806e5c60c QPDFPageObjectHelper::placeFormXObject: use std::string const& (fixes #374) 2020-04-03 12:17:57 -04:00
Jay Berkenbilt
97de12343b Performance: remove Members indirection for Pipeline 2020-04-03 12:17:57 -04:00
Jay Berkenbilt
bfda941519 Use an unordered map for SparseOHArray for efficiency 
This was added in C++11.
 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
ee271fd2f2 Use auto for iterating over sparse array 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
70665cb381 Internally use unsafeShallowCopy where we can 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
38afdcea7b Add QPDFObjectHandle::unsafeShallowCopy 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
07afb668b1 Performance: remove indirection through Members for QPDFObject 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
89f19b7099 Performance: remove Members indirection for QPDFObjectHandle 2020-04-03 12:16:24 -04:00
Jay Berkenbilt
dac65a21fb Look in form XObjects when removing unreferenced resources (fixes #373) 
If a page contains a form XObject, also filter the form XObject and
remove its unreferenced resources.
 2020-03-31 17:39:20 -04:00
Jay Berkenbilt
278710fbe8 Refactor QPDFPageObjectHelper::removeUnreferencedResources() 
Refactor removeUnreferencedResources to prepare for filtering form
XObjects.
 2020-03-31 17:39:20 -04:00
Jay Berkenbilt
bb6768b8f0 Include header for wcslen (fixes #405) 2020-02-29 08:43:33 -05:00
Jay Berkenbilt
bb3137296d Handle root /Pages pointing to other than page tree root (fixes #398) 2020-02-22 11:10:31 -05:00
Jay Berkenbilt
52a2e95dd5 Prepare 9.1.1 release 2020-01-26 18:49:04 -05:00
Jay Berkenbilt
57c01ef81f In qdf mode, don't write extra XRef streams (fixes #386) 
fix-qdf assumes there is exactly one XRef stream and that it is at the
end of the file.
 2020-01-26 16:50:57 -05:00
Jay Berkenbilt
bbc2f8ffae Bug fix: handle ColorSpace lookup for inline images (fixes #392) 
If the value of /CS in the inline image dictionary was is key in the
page's /Resource -> /ColorSpace dictionary, properly resolve it by
referencing the proper colorspace, and not just the name, in the
external image dictionary.
 2020-01-26 15:29:10 -05:00
Cloudmersive
a8b6ff5763 Fix for Windows unable to acquire crypt context with new keyset (fixes #387) 
Fix is based on guidance
https://support.microsoft.com/en-us/help/238187/cryptacquirecontext-use-and-troubleshooting
and is the proper fix for #285/#286
 2020-01-14 18:45:54 -05:00
Jay Berkenbilt
a44b5a34a0 Pull wmain -> main code from qpdf.cc into QUtil.cc 2020-01-14 11:40:51 -05:00
Jay Berkenbilt
ab4061f1ee Add error detection for read_lines_from_file(FILE*) 2020-01-14 11:07:09 -05:00
Jay Berkenbilt
211a7f57be QUtil::read_lines_from_file: optional EOL preservation 2020-01-13 11:26:18 -05:00
Jay Berkenbilt
9a398504ca Refactor QUtil::read_lines_from_file 
This commit adds the preserve_eol flags but doesn't implement EOL
preservation yet.
 2020-01-13 09:19:53 -05:00
Jay Berkenbilt
9b0c6022d7 Prepare 9.1.0 release 2019-11-16 22:29:54 -05:00
Jay Berkenbilt
5e6dfc938e Prepare 9.1.rc1 release 2019-11-09 22:00:53 -05:00
Jay Berkenbilt
c4478e5249 Allow odd/even modifiers in numeric range (fixes #364) 2019-11-09 13:23:12 -05:00
Jay Berkenbilt
5508f74603 Allow /P in encryption dictionary to be positive (fixes #382) 
Even though this is disallowed by the spec, files like this have been
encountered in the wild.
 2019-11-09 12:33:15 -05:00
Jay Berkenbilt
127a957aee Allow runtime inspection/override of crypto provider 2019-11-09 09:53:42 -05:00
Jay Berkenbilt
88bedb41fe Implement gnutls crypto provider (fixes #218) 
Thanks to Zdenek Dohnal <zdohnal@redhat.com> for contributing the code
used for the gnutls crypto provider.
 2019-11-09 09:53:38 -05:00
Jay Berkenbilt
cc14523440 Update autoconf to support crypto selection 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
d0a53cd3ea Fix typos in configure.ac 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
c03ced09c0 Isolate source files used for native crypto 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
d1ffe46c04 AES_PDF: move CBC logic from pipeline to AES_PDF implementation 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
c8cda4f965 AES_PDF: switch to pluggable crypto 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
bb427bd117 SHA2: switch to pluggable crypto 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
eadc222ff9 Rename SHA2 implementation (non-bisectable) 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
4287fcc002 RC4: switch to pluggable crypto 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
0cdcd10228 Rename RC4 implementation (non-bisectable) 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
ce8f9b6608 MD5: switch to pluggable crypto 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
5c3e856e9f Rename MD5 implementation (non-bisectable) 
Just rename MD5 -> MD5_native in place so that git annotate will show
the lines as having originated there.
 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
2de41856a0 QPDFCryptoProvider: initial implementation 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
700f5b961e Remove int type checks -- subsumed by C++-11 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
653ce3550d Require C++-11 
Includes updates to m4/ax_cxx_compile_stdcxx.m4 to make it work with
msvc, which supports C++-11 with no flags but doesn't set __cplusplus
to a recent value.
 2019-11-09 08:18:02 -05:00
Jay Berkenbilt
9094fb1f8e Fix two additional fuzz test cases 2019-11-03 18:59:12 -05:00
Masamichi Hosoda
5a842792b6 Parse Contents in signature dictionary without encryption 
Various PDF digital signing tools do not encrypt /Contents value in
signature dictionary. Adobe Acrobat Reader DC can handle a PDF with
the /Contents value not encrypted.

Write Contents in signature dictionary without encryption

Tests ensure that string /Contents are not handled specially when not
found in sig dicts.
 2019-10-22 16:20:21 -04:00
Masamichi Hosoda
cdc46d78f4 Add QPDFObject::getParsedOffset() 2019-10-22 16:19:06 -04:00
Masamichi Hosoda
50b329ee9f Add QPDFWriter::getWrittenXRefTable() 2019-10-22 16:16:16 -04:00
Masamichi Hosoda
5cf4090aee Add QPDFWriter::getRenumberedObjGen() 2019-10-22 16:16:16 -04:00
Masamichi Hosoda
46ac3e21b3 Add QPDF::getXRefTable() 2019-10-22 16:16:16 -04:00
Masamichi Hosoda
06b818dcd3 Exclude signature dictionary from compressible objects 
It seems better not to compress signature dictionaries. Various PDF
digital signing tools, including Adobe Acrobat Reader DC, do not
compress signature dictionaries.

Table 8.93 "Entries in a signature dictionary" in PDF 1.5 reference
describes that /ByteRange in the signature dictionary shall be used to
describe a digest that does not include the signature value
(/Contents) itself.

The byte ranges cannot be determined if the dictionary is compressed.
 2019-10-22 16:16:16 -04:00
Masamichi Hosoda
5e0ba12687 Fix /Contents value representation in a signature dictionary 
Table 8.93 "Entries in a signature dictionary" in PDF 1.5 reference
describes that the value of Contents entry is a hexadecimal string
representation when ByteRange is specified.

This commit makes QPDF always uses hexadecimal strings representation
instead of literal strings for it.
 2019-10-22 16:16:16 -04:00