summaryrefslogtreecommitdiff
path: root/core/fxcodec
AgeCommit message (Collapse)Author
2016-07-12Replace void* to actual type for jbig2 context to avoid castsweili
Also clear up a few variable names and unnecessary brackets. Review-Url: https://codereview.chromium.org/2143083002
2016-06-27Double AdobeCMYK_to_sRGB speed with faster roundingbrucedawson
FXSYS_round is painfully slow on Windows. It does range checking and then calls an extremely expensive function. It ends up consuming half the CPU time when decoding the images in PDFs such as this one: https://www.ets.org/Media/Tests/GRE/pdf/gre_research_validity_data.pdf SSE can be used to optimize this: __m128 cmyk = {c * 255, m * 255, y * 255, k * 255}; uint32_t output[4]; _mm_storeu_si128((__m128i*)output, _mm_cvtps_epi32(cmyk)); but is cryptic, only works for x86/x64, and gives basically identical performance to this solution - int(c * 255 + 0.5f); The rounding behavior is not identical but in practice this rarely matters, and in this specific case it does not matter because the edge cases that vary are not hit. The three divisions at the end were changed to multiplies because profiling showed they were a significant cost. This change reduces the image-decode stalls in the PDF listed above by about 40%, making for a noticeably better experience. Further optimizations are possible but would require significantly more time and testing. BUG=617365 Review-Url: https://codereview.chromium.org/2096723003
2016-06-20Fixup LoadImageInfo type checking.dsinclair
The ::DetectImageType method does more then just detecting the image type, it also sets up various needed structures to handle the decoding. Instead of skipping the ::DetectImageType call this CL changes the code to return early if the image check fails. This should allow us to stop working on images which do not match the required data format. BUG=chromium:621094 Review-Url: https://codereview.chromium.org/2085493002
2016-06-16Remove unused code.dsinclair
This cl removes code that exists but is never called. This includes: * cfwl_formtp * cfwl_widgetdelegate * cfwl_scrollbar * cfwl_theme * cfwl_tooltip * PWL_Label * PWL_ListCtrl * PWL_Signature * PWL_IconList * BC_ResultPoint * BC_CommonPerspectiveTransform * BC_CommonBitSource * BC_PDF417Codeword * fx_codec_flate.h (the .cpp file still exists) Review-Url: https://codereview.chromium.org/2071953002
2016-06-15Make code compile with clang_use_chrome_plugin (part IV)weili
This change mainly contains files in fpdfsdk/ directory. This is part of the efforts to make PDFium code compilable by Clang chromium style plugins. The changes are mainly the following: -- move inline constructor/destructor of complex class/struct out-of-line; -- add constructor/destructor of complex class/struct if not explicitly defined; -- add explicit out-of-line copy constructor when needed; -- move inline virtual functions out-of-line; -- Properly mark virtual functions with 'override'; -- some minor cleanups plus removing an unused file and splitting cxfa_eventparam out from fxfa.h BUG=pdfium:469 Review-Url: https://codereview.chromium.org/2062313002
2016-06-14Make code compile with clang_use_chrome_plugin (part II)weili
This change contains files in core directory which were not covered in part I. This is part of the efforts to make PDFium code compilable by Clang chromium style plugins. The changes are mainly the following: -- move inline constructor/destructor of complex class/struct out-of-line; -- add constructor/destructor of complex class/struct if not explicitly defined; -- add explicit out-of-line copy constructor when needed; -- move inline virtual functions out-of-line; -- Properly mark virtual functions with 'override'; -- some minor cleanups; BUG=pdfium:469 Review-Url: https://codereview.chromium.org/2060913003
2016-06-13Optionally skip image type detection in progressive decoder.dsinclair
The progressive decoder will attempt to verify that the provided image type matches the actual image content. We need to disable this check when running the fuzzer in order to target the fuzzing to specific decoders otherwise each fuzzer will end up fuzzing all of the decoders. BUG=chromium:587126 Review-Url: https://codereview.chromium.org/2061733002
2016-06-09Clean up fx_codec_tiff.cpp.thestig
Fix regressions from commit 4997b22. BUG=618164 Review-Url: https://codereview.chromium.org/2053573003
2016-06-07Get rid of NULLs in fpdfsdk/thestig
Review-Url: https://codereview.chromium.org/2031653003
2016-06-07Fix more code which has shadow variablesweili
The code has local variables that shadow struct or class member variables. Also, when this happens, different variable names should be used instead of namespaces. These were discovered by /Wshadow warning flag in Clang. Review-Url: https://codereview.chromium.org/2034253003
2016-06-07Get rid of NULLs in core/thestig
Review-Url: https://codereview.chromium.org/2032613003
2016-06-07Verify we have a CJBig2_Image before attempting use.dsinclair
In CJBig2_SDDProc::decode_Arith we will set a SDNEWSYMS value to nullptr if the height or width are 0. With the PDF from the bug, all of the decoders are set to nullptr. Then, we call into CJBig2_TRDProc::decode_Arith and pull out one of the nullptr decoders and attempt to use it, crashing. This CL adds a check that we have a non-null decoder before attempting to use the decoder. BUG=pdfium:511 Review-Url: https://codereview.chromium.org/2048683002
2016-06-06Remove unused PS generation code.thestig
CFX_PSRenderer, IFX_PSOutput, CPSOutput, CPSPrinterDriver. Also reland commit 2d63eaa. Review-Url: https://codereview.chromium.org/2019603002
2016-06-03Fix some code which causes warnings when compiled by /analyze toolweili
The code may not cause error conditions, but can be improved. These warnings include uninitialized variables, signed/unsigned mismatch, redundant condition, and using bool in arithmetic operations. Also remove a chunk of unused code. BUG=chromium:613623, chromium:427616 Review-Url: https://codereview.chromium.org/2036203004
2016-06-02Fix all the code which has duplicate variable declarationsweili
When there are duplicate variable declarations, the inner names shadow the outter ones. This is error prone and harder to read. Remove all the instances found by /analyze. BUG=chromium:613623, chromium:427616 Review-Url: https://codereview.chromium.org/2027273002
2016-05-26Fix a bug which may cause infinite loopweili
When the condition is "status == FXCODEC_STATUS_DECODE_TOBECONTINUE" while |status| never gets updated in the loop, it would enter infinite loop. Also, since Start_decode_MMR() never returns FXCODEC_STATUS_DECODE_TOBECONTINUE, there is no point to check on the return value for that. Review-Url: https://codereview.chromium.org/2013263002
2016-05-25Banish CFX_PrivateData to the XFA sidetsepez
Remove an unused inheritance. The remaining usages are a little more intricate. Review-Url: https://codereview.chromium.org/2015743002
2016-05-25Remove CFX_PrivateData from CPDF_Documenttsepez
Replace it with two generic slots for Links and Codec usage. Since the codec is at a lower layer than the document, we don't provide separate get/set methods, since having a document upon which to call these would be a layering violation. Do the same for the Links for simplicity. Review-Url: https://codereview.chromium.org/2005193003
2016-05-24Remove CFX_PrivateData from CPDF_ModuleMgrtsepez
Its only used to store one object, so replace it with a unique_ptr to a class with a virtual dtor. Rename the prototypical class with virtual dtor from CFX_DestructObject to CFX_Deletable. Rename the fx_basic_module.cpp to cfx_modulemgr.cpp to match the one class in it. Review-Url: https://codereview.chromium.org/2013483003
2016-05-16Convert border style defines to an enum class.dsinclair
There were two defines, BBS_ and PBS_ for the various border styles in the system. They were the same, except PBS_ had an extra SHADOW define which was never used. This CL combines both of those into a single BorderStyle enum class and updates the code as needed. Also, removes ADDBIT, GETBIT unused defines. Updates barcode code to use the util.h defines instead of redefinition. fsdk_baseannot names starting with _ were cleaned up and some #defines moved to constants. Review-Url: https://codereview.chromium.org/1980973002
2016-05-13Make CFX_ByteString(const CFX_ByteStringC&) explicit.tsepez
Add missing helper function to CFX_ByteTextBuf to avoid the anti-pattern CFX_ByteString(sBuf.AsStringC()), using the name "Make" to indicate there's an allocation going on in this case. Change some method arguments to take pre-existing ByteStrings where possible. Review-Url: https://codereview.chromium.org/1977093002
2016-05-11Replace some calls to Release() with direct delete, part 1.tsepez
Searching for the anti-pattern: void Release() { delete this; } We must be explicit on the ownership model. Add unique_ptrs as a result. Review-Url: https://codereview.chromium.org/1960673003
2016-04-29Avoid nullptr dereferences in sycc444_to_rgb().thestig
BUG=607739 Review-Url: https://codereview.chromium.org/1934483002
2016-04-27Standardize on ASSERT.dsinclair
There are currently three ways to assert in the code (ASSERT, FXSYS_assert and assert). This CL standardizes on ASSERT. The benefit of ASSERT is that it can be overridden if the platform requies and we can pickup the Chromium version if it has already been defined in the build. This does change behaviour. Currently FXSYS_assert is always defined but ASSERT is only defined in debug builds. So, the FXSYS_assert's would fire in Release builds. That will no longer happen. BUG=pdfium:219 Review-Url: https://codereview.chromium.org/1914303003
2016-04-14Replace calls to deprecated CFX_{Wide,Byte}String::Empty()tsepez
Use the more standard name "clear()" instead. Review URL: https://codereview.chromium.org/1888103002
2016-04-13Re-enable warning 4701 for GN build and some cleanupweili
Clean up three 4701 warnings (use potientially uninitialized variable) which are the only ones left; Re-enable warning 4701 for GN build; Remove an unused data structure; BUG=pdfium:29 Review URL: https://codereview.chromium.org/1885093002
2016-04-12Remove ICodec_* Interfaces.dsinclair
All of the ICodec_* interfaces had a single implementation. This CL removes the interfaces and uses the concrete classes in their place. BUG=pdfium:468 Review URL: https://codereview.chromium.org/1876023003
2016-04-08Remove CFX_ByteString::Load() and operator= for CFX_ByteTextBuftsepez
The few places these are called are better served by explicit calls to CFX_ByteString() itself. This helps make Byte and Wide strings more similar. Also prevents fx_string.h from having any knowledge of fx_basic.h's ByteTextBuf class. Review URL: https://codereview.chromium.org/1870043003
2016-04-07Remove unused ICodec_Jbig2Encoderdsinclair
Never implemented, never used. BUG=pdfium:468 Review URL: https://codereview.chromium.org/1867573004
2016-04-06Move include/fxcodec to fxcodec/includedsinclair
Review URL: https://codereview.chromium.org/1864153002
2016-04-06Rename FX_SAFE_DWORD to FX_SAFE_UINT32dsinclair
We removed the FX_DWORD typedef in favour of uint32_t. This CL cleans up the FX_SAFE_DWORD naming to match. BUG=pdfium:81, pdfium:470 Review URL: https://codereview.chromium.org/1861403002
2016-04-06Move core/include/fxge to core/fxge/include.dsinclair
This CL is a straight move of the fxge includes into core/fxge/include. Review URL: https://codereview.chromium.org/1868533002
2016-04-05Remove dead image decoding/caching code.thestig
Review URL: https://codereview.chromium.org/1860223002
2016-03-30Fix C4434 warning about 32-bit shift assigned to 64-bitsBruce Dawson
VS 2015 has a new or louder warning about 32-bit shifts that are then assigned to a 64-bit target. This type of code triggers it: int64_t size = 1 << shift_amount; Because the '1' being shifted is a 32-bit int the result of the shift will be a 32-bit result, so assigning it to a 64-bit variable is just misleading. In this case the code that triggers it is this: m_IAID.resize(1 << SBSYMCODELEN); The destination is a size_t so the warning only shows up on 64-bit builds and doesn't indicate a real bug. But, casting the '1' constant to size_t makes the behavior/intent more obvious and consistent and allows enabling C4334 in Chromium. BUG=593448 R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1843253002 .
2016-03-29Code change to avoid signed/unsigned mismatch warnings.Wei Li
This makes pdfium code on Linux and Mac sign-compare warning free. The warning flag will be re-enabled after checking on windows clang build. BUG=pdfium:29 R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1841643002 .
2016-03-28Reduce signed/unsigned comparison warningsWei Li
The warnings generated by Clang. This is part 1 for some simple cases. BUG=pdfium:29 R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1840483003 .
2016-03-25Remove FX_DWORD from core/ and delete definitionchromium/2695chromium/2694chromium/2693chromium/2692tsepez
Review URL: https://codereview.chromium.org/1832173003
2016-03-23Split core/include/fpdfapi/fpdf_resource.hdan sinclair
This CL splits apart the core/include/fpdfapi/fpdf_resource.h file and places the classes into individual files. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1824033002 .
2016-03-23Move core/include/fxcrt to core/fxcrt/include.Dan Sinclair
This CL moves the fxcrt code into the core/fxcrt directory. The only exception was fx_bidi.h which was moved into core/fxcrt as it is not used outside of core/. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1825953002 .
2016-03-22Re-enable MSVC warning 4702Wei Li
BUG=pdfium:29 R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1821423002 .
2016-03-21Remove FX_WORD in favor of uint16_t.Tom Sepez
It isn't buying us anthing, and it looks strange in a struct when other uint types are already present. R=dsinclair@chromium.org Review URL: https://codereview.chromium.org/1821043003 .
2016-03-21Make a few more const tables smaller.chromium/2689chromium/2688Tom Sepez
Remove some tables from .h file (risk of duplication). R=ochang@chromium.org Review URL: https://codereview.chromium.org/1814233005 .
2016-03-21Re-enable several MSVC warningsWei Li
Re-enable the following warnings: 4245: signed/unsigned conversion mismatch; 4310: cast may truncate data; 4389: operator on signed/unsigned mismatch; 4701: use potentially uninitialized local variable; 4706: assignment within conditional expression Clean up the code to avoid those warnings. BUG=pdfium:29 R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1801383002 .
2016-03-15core/include/fdpfapi cleanup Part I.Dan Sinclair
This Cl moves a bunch of the files from core/include/fpdfapi to their correct location outside the core/include tree. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1805603002 .
2016-03-15Reland "Fix sycc{420,422}_to_rgb issues."chromium/2680Oliver Chang
Fix an incorrect unit test result. This reverts commit 2df269c52741c12d63300cee806e96ae5d0b038f. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1803043002 .
2016-03-15Split CPDF_Stream/CPDF_StreamAcc into separate filesTom Sepez
R=dsinclair@chromium.org Review URL: https://codereview.chromium.org/1783933003 .
2016-03-15Revert "Fix sycc{420,422}_to_rgb issues."Oliver Chang
This reverts commit 48b934135171b55b1ffc510226317fdddef5f5b8. Broke unit tests. TBR=tsepez@chromium.org Review URL: https://codereview.chromium.org/1800153002 .
2016-03-15Fix sycc{420,422}_to_rgb issues.Oliver Chang
sycc422_to_rgb was incorrect. This patch adds a similar check to the one in sycc420, and fixes an issue with the previous sycc_420_size_is_valid function where truncation from division wasn't accounted for. BUG=591785 R=jun_fang@foxitsoftware.com, tsepez@chromium.org Review URL: https://codereview.chromium.org/1785323003 .
2016-03-14Move core/src/ up to core/.Dan Sinclair
This CL moves the core/src/ files up to core/ and fixes up the include guards, includes and build files. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1800523005 .