3.8.0

Changes (Release 3.8.0)

Summary (pypdfium2)

• Major improvements to GitHub Workflows.
• Fixed setup on macOS with Python 3.11. (Wheel users were not impacted.)
• Fixed build script on macOS (GN configuration).
• Fixed build script on Windows (patches and RC).
• Fixed test suite on Windows (cleanup failure).
• Fixed test suite on Python < 3.7.

PDFium commit log

Commits between 5406 and 5418 (latest commit first):

• a98c08922 Ensure checkout_skia is defined before use
• 4948d0e50 Only GN check //skia if checkout_skia=True
• f8cf2e34d Add ARM64 expectations for pdfium embedder tests
• f0e239249 Change CBC_OneDimWriter::AppendPattern() to use spans.
• d7fe73817 Better span arithmetic in CPDF_Image::SetImage()
• 496ff1161 Prepare CBC_OneDimWriter::AppendPattern() for span conversion.
• 055a8a210 Clean up CBC_OnedCode39Writer.
• 65cbd7d93 Return vector from CPDF_Stream::ReadRawData()
• 06943a994 Use more spans in streams, part 5
• 1852aa019 Include more Skia files in GN checks
• 1d94eaddb [Skia] Remove bug_880920.pdf from the suppression list
• c2ab68313 Suppress GN check issues with Skia includes in fpdf_view*.cpp.
• 1ed2d0bd9 Change CBC_OneDimWriter::Encode() to return DataVector.
• dc2219869 Change CBC_OneDimWriter::RenderResult() to use a span.
• 898c5e841 Remove dead code in CBC_OneDimWriter::RenderResult().
• 4fbd4ab67 [Skia] Fix the issue that unwanted black borders are rendered.
• b5de5e89e Get rid of a multiplier in the 1D barcode implementation.
• 5c042e4aa Allow spans to flow further down into CFX_DIBBase.
• 8ebc9515f Use more spans in streams, part 4.
• 1f8d66db1 Convert CFX_BitStream to store a span<>
• ff0c8aa86 Remove needless multiplication by 1 in CPDF_Image::SetImage()
• 868d94c69 Pass spans further down into jpx/jbig2 decoders
• 52696f63d Return span<> from CFX_DIBBase::GetBuffer().
• 477cbda97 Improve SkPictureRecorder ownership
• 91e0c3f14 Use bounds-checked spancpy() in cpdf_rendershading.cpp
• e05af4345 Reduce layers of wrapper functions in CBC_OneDimWriter::Encode().
• f673e82fd Remove height out-parameter from 1D barcode encoding methods.
• 818c4cd3c Run GN check on third_party/icu.
• a42db2e05 Use spancpy() in a few more places.
• 6472a498b Pass spans to ProgressiveDecoder::GifReadScanline().
• ed2828ff7 Replace sprintf calls with snprintf in third_party
• 6a1250879 Use retained locals in CPDF_FormField
• 804fd9ee2 Fix typos in FPDFAnnotEmbedderTest.InkAnnotation
• 3d4231cfb Roll Instrumented Libraries from 673b76772c2e to a8992bf8abb3 (10 revisions)

3.7.0

Changes (Release 3.7.0)

Summary (pypdfium2)

pypdfium2 commit log

Commits between 3.6.0 and 3.7.0 (latest commit first):

• e220e13 [autorelease] update changelog and version file
• 039d3a3 Add additional testing workflows
• 632ff95 GH pages: nevermind, build on push to main
• caf99d3 GH Pages: Daily build at 11 o'clock UTC
• 734bec3 Add workflow for GH Pages
• 31b7ad2 Blacklist broken wheel release
• 95b4a04 workflow: address set-output deprecation

PDFium commit log

Commits between 5392 and 5406 (latest commit first):

• 7b8416e79 Add NOLINTNEXTLINE missed in prior CL.
• 490d68130 Add NOLINTNEXTLINE() to public methods with murky ownership.
• 8df158b30 Roll third_party/abseil-cpp/ 751b72a8c..384efd2e8 (11 commits)
• a85fe5432 Roll build/ 371b6c310..53351eb16 (37 commits)
• 15255e1a4 Move-assign object in cpdf_object_walker_unittest.cpp
• 010d211f3 Use retained locals in cpdf_pagecontentgenerator_unittest.cpp
• af616af06 Use retained local in CPDF_StructTree::LoadPageTree()
• 5b5987975 Remove RetainedTreeNode.
• 9f138e5d1 Restructure CPDF_Object methods to avoid analyzer warnings
• 4da347bd4 Remove RetainPtr::operator=(T*)
• d89d2e00f Use retained locals to avoid implicit .Get() calls
• 20390f831 Turn off PartitionAlloc when building with GCC as well.
• d3d2ad3d2 Take small suggestions from clang-analyzer run over fpdfapi/
• 3e08c2abb Avoid fatal OOM in CStretchEngine::StartStretchHorz().
• 5fdfb25eb Use more spans in CStretchEngine.
• d6f9fe053 Fix some nits in fx_memory code.
• 6c5cfd873 Use FixedTryAllocZeroedDataVector in CPDF_ContentParser.
• e5aa62574 Implement FixedTryAllocZeroedDataVector.
• e5bd69272 Actually fix clang-tidy report
• 257a3d763 Change CPDF_SyntaxParser::ReadBlock() to take a span.
• 204d96e1b Re-enable OOM test cases.
• e1c2bcf6c Use more spans in streams, part 3.
• b33a5a61d Build SkOpts_crc32 on ARM64
• 8636cff40 Fix copyright headers in GN files
• 5c00919b2 Speculative change for clang-analyzer.
• 1f8663684 Use more spans in streams, part 2.
• b81de425a Remove const-ref retained stream arguments, step 3.
• b9898333d Fuzz the new CJPX_Decoder::Create() parameter.
• 6a39df1f7 Use more spans in fxcrt streams.
• 9b46097ae [Skia] Remove some corpus tests from the suppression list
• cb1aee119 Add DCHECK() for non-zero sized spans with null data members.
• fb3d15edc Roll tools/memory/ 98140a694..8581381d9 (2 commits)
• b3cb989f7 Roll buildtools/third_party/libunwind/trunk/ 7ff728a97..aabcd8753 (8 commits)
• 7d89cb013 Roll buildtools/third_party/libc++abi/trunk/ 5c3e02e92..8dd405113 (10 commits)
• 75f0ff464 Remove const-ref retained stream arguments, step 2.
• 06d220f2b Add stronger checks in span_util.h
• 732e38a17 Remove const-ref retained stream arguments, step 1
• c323bf9ba Disable LargeImageDoesNotRenderBlank test case in debug builds.
• 027326013 Add a test case for rendering a large image.
• 832a694e5 Update old-style copyright headers
• f01a27d2b Fix non-compliant copyright headers
• 5e33cbd12 Enable pan-project checks
• 8d2f700af Improve integer overflow checking in CStretchEngine::StartStretchHorz().
• ab6c904a2 Update gn_version to git_revision:11dc0b1f438bd26380774e9d50fd4c63f346d41a
• dd7089fa8 Roll Catapult from 16eb7278843e to 3a19f7afaa58 (59 revisions)
• 59f26709f Roll MarkupSafe from 1b882ef6372b to 13f4e8c9e206 (1 revision)
• 9cab1a6e7 Roll Code Coverage from 690a8f131004 to 5afe9c957e79 (1 revision)
• be6fac607 Roll Jinja2 from ee69aa00ee85 to 4633bf431193 (1 revision)
• bdbf62394 Roll third_party/nasm/ 9215e8e1d..fc8e0bd89 (3 commits)

3.6.0

Changes (Release 3.6.0)

Summary (pypdfium2)

pypdfium2 commit log

Commits between 3.5.0 and 3.6.0 (latest commit first):

• 81638c9 [autorelease] update changelog and version file

PDFium commit log

Commits between 5378 and 5392 (latest commit first):

• 16812d6c1 Return object number from CPDF_IndirectObjectHolder::AddIndirectObject()
• 5b59f6d89 Add FPDFText_HasUnicodeMapError() API.
• b899fba4e Fix one unretained dictionary usage in CPDF_DIB
• 03571c10b Delete shebangs from some skia_gold scripts.
• 7faffd69b Remove "from future" Python imports.
• b915e047f Change safetynet scripts to run with python3.
• 231e7e159 Optimization of decoding large JPX images
• 82cad2c70 Update comments about ownership in cpdf_indirect_object_holder.h
• b76ba9742 [Skia] Fix the black borders of small images
• f84a8a581 Roll third_party/depot_tools/ e3ed6a8e0..6f2321d1d (55 commits; 23 trivial rolls)
• 8dda52ced Roll Fuchsia SDK from version:9.20220929.2.1 to version:10.20221027.2.1
• 6f144ed06 Remove remaining UnownedPtr<>::Get() calls from core/
• 8f819ac22 Roll base/allocator/partition_allocator/ bb627f4ab..cb1f2e8e0 (48 commits)
• bb9bf7d99 Remove explicit UnownedPtr<>::Get() calls from core/fpdf*/
• c4acde581 Remove explicit UnownedPtr<>::Get() calls from fpdfsdk/
• 654d3591d Remove explicit UnownedPtr<>::Get() calls from fxjs/
• b4d26b104 Remove explicit UnownedPtr<>::Get() calls from fxbarcode/
• 0c3768e5d Remove explicit UnownedPtr<>::Get() calls from xfa/
• 5cc1afc13 Pass retained argument to ReplaceAbbr()
• 5ade14bea Return retained objects from CPDF_FormField methods.
• 245050a61 Report more detail to ResultDB for pixel tests
• ddc99ced8 Add test for CFFL_ComboBox::SetActionData()
• dab891757 Roll third_party/zlib/ 6fe4ce825..cff720848 (3 commits)
• 4eda5e1f5 Roll third_party/icu/ 20f8ac695..35132cceb (2 commits)
• cc90b0cd9 Hide CPDF_FormField from CFFL_ComboBox
• 59445b0a8 Roll tools/clang/ a5e0d7234..38497dbb2 (21 commits)
• 2aa86baac [Skia] Fix issue that a dot annotation is not rendered
• 1f7ef7c1d Remove some CPDF_Dictionary* locals.
• 7c96f4765 Remove CPDF_Creator::m_pMetaData.
• 2bd16e8c9 Make CFFL_ComboBox embedder test use higher-level methods.
• 0df0082af Add embedder test for FFL_ComboBox::GetSelectExportText()
• ec350d928 Create CFFL layer embedder test.
• cb6e71594 Remove CPDF_FormField::GetForm()
• a5bd2fa0f Hide CPDF_FormControl from CFFL_Button
• 1f8e56793 Rename CPWLComboBoxEditEmbedderTest::GetCFFLFormFiller()
• c44513e09 Make CPDF_PageView::GetInteractiveForm() a private method.
• 9fe8223f0 Forward declare CPDF_ImageLoder in cpdf_imagerenderer.h
• 8482a83de Replace CPDF_Page::SetImageCache() with AddImageCache()
• 98204b688 Move pixel test-only code to TestPixel()
• 61a662172 Make CFX_RenderDevice::GetDeviceDriver() a protected method
• b6c3a4de2 Revert "[Skia] Fix issue that a dot annotation is not rendered"
• 22396a1ac [Skia] Fix issue that a dot annotation is not rendered
• 7ae9c064b Roll third_party/skia/ 3050ef5ea..648827391 (287 commits)
• 371f0b063 Remove const form of CPDF_RenderContext::Layer::GetObjectHolder()
• 31fe0722b Add Skia expectation for an AGG-only line drawing issue.
• 927f85782 Roll third_party/skia/ ebfa2b71b..3050ef5ea (18 commits)
• 3bd5f342d Move CPDF_ImageLoader to page/ layer.
• 2c9b918c2 Roll third_party/skia/ 03cb8e7d8..ebfa2b71b (4 commits)
• 68fdc4913 Remove render layer concepts from CPDF_ImageLoader.

3.5.0

Changes (Release 3.5.0)

Summary (pypdfium2)

• Fixed an issue with the __new__ method of PdfPageObject that caused __init__ to run twice.
• Fixed faulty examples in the Readme.

pypdfium2 commit log

Commits between 3.4.0 and 3.5.0 (latest commit first):

• 0599f11 [autorelease] update changelog and version file
• b5560d9 Update Readme and Changelog
• 76b656a pageobj: do not run init twice
• eb9e7c8 Readme: Fix image insertion example
• 0e77758 docs: improve changelog integration for latest builds again

PDFium commit log

Commits between 5364 and 5378 (latest commit first):

• 16a4de2bb Report test_runner.py test case output to ResultDB
• 8841d123a Shorten CPDF_PageImageCache::ImageCacheEntry to Entry
• 57a6f3d3d Remove CPDF_Page::RenderCacheIface
• 167a560c7 Move CPDF_PageImageCache to page/ layer.
• b96600ee9 Rename CPDF_PageRenderCache to CPDF_PageImageCache.
• c988eee72 Remove render-layer concepts from CPDF_PageRenderCache.
• bce5eb53f Processing the NoRotate flag when displaying annotations
• 152084631 De-duplicate document pointers in CPDF_PageRenderCache::ImageCacheEntry
• 74faf2c3b Add experimental FPDFText_IsGenerated() API.
• 607b84ebd Roll v8/ f666aea4d..7ff976805 (323 commits)
• 4e52af07a Mark CPDF_Page::m_pPDFDocument as const
• deba3a79b Update PartitionAlloc comment
• f0f61c87d Remove CPDF_StreamAcc::GetData().
• a46ea9b04 Remove CPDF_Stream's span ctor.
• 6dc8a147d Remove CPDF_Stream ctor that takes a std::unique_ptr.
• 5e5c2bd24 Limited build support for iOS.
• b9a3dd089 Roll third_party/freetype/src/ 905270925..0b62c1e43 (18 commits)
• 5ea4ecf11 Add a pixel test for annotations with the NoRotate flag.
• dab79f117 Remove data parameter from CPDF_Stream::InitStream() and rename.
• a3c920fb3 Replace some CPDF_Stream std::unique_ptr ctor usage.
• 235e56abd Change CPDF_Stream to store memory-based data in a DataVector.
• d61dac0d9 Use Google-style naming for CPDF_Stream member variables.
• 9ab44983c Change CPDF_Stream to use absl::variant.
• 11d077ffc Run Skia Gold as part of each pixel test
• 784cc119a Avoid another use of CPDF_Document::GetPageData().
• 91d13f4aa Reduce use of CPDF_Document::GetPageData()
• 07d4ab7ec Change CPDF_Stream::GetInMemoryRawData() to return a span.
• 9b0ee1fb2 Disentangle CPDF_ContentParser from CPDF_Form.
• baa3f372f Make CPDF_Pattern::document() a protected method.
• 7e9f3de25 Use indirect object holder superclass in CPDF_PageContentManager.
• fb6dbcd0b Realize objects created with FPDF_NewFormObjectFromXObject().
• d07ce743f Returned retained result from CXFA_TextParser::GetParentStyle().
• 47abf3795 Roll v8/ 261f00697..f666aea4d (42 commits)
• 60a072f74 Report test duration to ResultDB
• 56c4682ae Don't pickle TestRunner
• 2a09fff07 Separate test_runner.py's per-process state
• d08ac75a8 Also check FPDFPageObj_GetBounds() in ImportPageToXObject test.
• ee31732f6 Avoid exposing CFFL_InteractiveFormFiller::CallbackIface to callers.
• 821be7591 Remove CPDF_RenderOptions::GetOCContext().
• b50fad7d0 Make CPWL_Wnd::GetParent() a protected method.
• cb75a6ce4 Remove CXFA_FFNotify::GetWidgetHandler().
• 61a96c1db Revert "Disable ResultDB for test_runner.py on Windows"
• f0a2d55c9 Roll build/ 01569374d..371b6c310 (103 commits)
• 44528e4e1 Remove stray cppgc::Member<>::Get() calls
• 0351b27f9 Convert CFXJSE_FormCalcContext::m_pDocument to WeakPersistent<>.
• 756333545 Remove CFXJSE_ResolveProcessor::GetNodeHelper().
• aad1a8481 Remove CPDF_Annot::GetPopupAnnot()
• 4adc95560 Enable 2 CFWLEditEmbedderTests

3.4.0

Changes (Release 3.4.0)

Summary (pypdfium2)

• Added a new support model PdfImageObject (which inherits from PdfPageObject). This can be used to insert a JPEG image into a page, get metadata, etc.
• Docs: The changelog page now selectively includes an entry for the next release that may be shown on latest builds.

pypdfium2 commit log

Commits between 3.3.0 and 3.4.0 (latest commit first):

• 75a3854 [autorelease] update changelog and version file
• 1c6bde4 Improve insert_jpeg() API
• 9998d5a Revert "Remove PdfDocument._rendering_input and related features"
• d92ba57 misc: replace private method
• e7fad89 Move colour validation
• b175e3c readme: add jpeg insertion example
• e10f083 jpegtopdf: enhance comment
• 7c34f76 pageobjects: add initial image support model (#151)
• a631a1a docs/matrix: remove trivial information
• 18017dc docs: group C wrappers and pure-python helpers
• ab934e4 Use a smarter way to keep data in memory
• c1b0261 Inline utility methods
• 613f10f document: make assignment order nicer
• 90a6c38 Smarter inclusion of changelog_staging
• a829800 docs: include changelog_staging depending on branch
• 7aaa2e8 setup_base: nvm, use previous ext_modules approach
• 549e02f Inline opener into document
• bfbaf0f Clean up unused imports
• a2d0762 Make file access utilities public
• cc0c2b6 opener: make get_fileaccess() a standalone function
• 950db18 Slightly improve docs for static attributes
• 5991e8b setup: don't pin ctypesgen anymore
• bfd4e53 setup: this if should have been an elif
• f1f510c Remove PdfDocument._rendering_input and related features

PDFium commit log

Commits between 5351 and 5364 (latest commit first):

• 336fae19c Remove unused method CPDFSDK_Widget::GetInteractiveForm()
• fdbd4bb76 Use fxcrt::MakeFakeUniquePtr() in a few more places.
• 3deb8857c Disable ResultDB for test_runner.py on Windows
• bc05517b5 Remove CXFA_FFDoc::GetPDFFontMgr()
• 52370c146 Make CPDF_PageObjectHolder::RemovePageObject() clearer
• d1aaef4aa Replace tuples with dataclasses in test_runner.py
• 89acf61db Use pdfium::Contains() to check CPDF_Form membership.
• 72312d25b Add CFXFontMapperSystemFontInfoTest.
• 85c3cc8c7 [Skia] Remove path_9.pdf from the suppression list
• df597389f Remove .vpython.
• 0e1f9b468 Replace CFX_Font::SetSubData() with AllocSubData().
• b214cdcb3 Wrap CFXJSE_Engine::GetJseContext() with GetJseContextForTest().
• 146766408 Avoid returning unowned reference from CPDF_PSOP::GetProc()
• 20ab33b4a Simplify test case dispatch in test_runner.py
• 3c5b5abbb Remove CPDF_Parser::GetSyntax()
• 8bad4d5fd Remove 2 corpus tests from the suppression list
• 9d2c662f5 Fix CHECK failures with GetFontData() calls.
• 318ad8501 Rename CPDF_DataAvail::GetHintTables() to GetHintTablesForTest()
• 30814aedc De-virtualize CPDFXFA_Context::GetPDFDoc().
• 9ab5ed692 Make CPDF_ColorState::SetPattern() less confusing.
• 2f7efe5d8 Pass retained args and arrays as rvalues to color.
• a997e9b48 Add ResultDB support to test_runner.py
• 029e389ef Mark GetDocument() protected in CPDF_Document::PageDataIface.
• 820c7b418 Mark pdfium::span<> as TRIVIAL_ABI and GSL_POINTER
• 755476171 Encapsulate CFX_SubstFont::m_bFlagMM.
• 6a29d9e55 Add GSL_POINTER and TRIVIAL_ABI to compiler_specific.h
• ef47a11d4 Remove CPDF_Annot::GetDocument()
• 7516d8827 Remove CPDF_InteractiveForm::GetDocument().
• f369dce2e Consolidate Python code to locate PDFium root
• 529184818 Return retained reference from CPDF_Font::GetFontDict()
• 9e4521d1f Fix overlapping annot rendering with focused annot.
• 95b363233 Fix CPDF_DocPageData::MaybePurgeFontFileStreamAcc()
• 67a546fe6 Introduce CPDF_StreamAcc::KeyForCache()
• 560d03e25 Roll Instrumented Libraries from e09c4b66b6e8 to 673b76772c2e (1 revision)

3.3.0

Changes (Release 3.3.0)

Summary (pypdfium2)

• Disruption: PdfPage.insert_text() does not generate page content automatically anymore. The new PdfPage.generate_content() method now needs to be called to apply changes, to avoid generating content repeatedly.
• Disruption: In PdfTextPage.get_text(), the boundary values are now only set to defaults if they're None, not 0. This change was necessary because the underlying PDFium function requires PDF box values rather than normalised values.
• pypdfium2 finally implements automatic object finalisation. Calling the close() methods is not mandatory anymore. The context manager API of PdfDocument is retained for backwards compatibility, but exiting the context manager does not close the document anymore, since this would increase the risk of closing objects in wrong order.
• A new text extraction method get_text_range() was added. get_text() was renamed to get_text_bounded() but is still available as deprecated alias.
• Text pages now have an n_chars attribute. count_chars() is still available as deprecated alias.
• Added a helper class for transform matrices.
• Added support models to capture pages as XObjects, to get page objects for XObjects, to transform them with matrices, and to insert page objects into a page. This may be used to implement a custom N-up compositor, for instance.
• The document level renderer now uses a shortcut if processing just a single page.
• When rendering, pypdfium2 now checks if the document has forms before initialising/exiting a form environment.
• The Sphinx documentation now includes the Readme file.

pypdfium2 commit log

Commits between 3.2.0 and 3.3.0 (latest commit first):

• ba76f9b [autorelease] update changelog and version file
• 16b5bb8 page: move comment to right place
• 213ca63 workflow: re-activate schedule
• 593f1b6 Remove commented-out close calls (#150)
• 44f4aaa Remove excessive comments
• a0f79ef Introduce get_text_bounded() right now
• 7cf488f Mainly documentation improvements
• 464f6a7 textpage: make error handling caller-configurable
• 0f9ca15 textpage.get_text(): check coords are within mediabox
• dbc5419 Comment out excessive debug messages
• 379b820 Merge development branch (#148)
• 02b2e77 doc/render: add shortcut for single-page pdfs
• be2770d Improve docs for PdfDocument.get_page_size() and PdfPage.get_size()
• 7ab2314 page: add notes regarding matrix rendering
• d4b6025 Improve matrix-related docs
• 640aad0 Add matrix helper methods (#147)
• 4e5a486 page.insert_object(): always set pageobj.pdf
• e2dc4f1 Update readme
• 78c1102 Add helpers for XObjects (#145)

PDFium commit log

Commits between 5337 and 5351 (latest commit first):

• fe212103e Return retained object from CPDF_FormControl::GetWidgetDict()
• 14b3b8cee Introduce CPDF_Dest::GetScrollPositionsArray().
• 4ede0790c Return retained references from CPDF_PageObjectHolder methods.
• 57595d4af Remove CPDF_StreamAcc::GetDict().
• cfa4d05c9 Add death test for bad pdfium::span<> lifetimes
• 756e1fe08 Use std::as_const() in a few places.
• 411532e63 [Skia] Fix coons patch mesh rendering
• a88c2be5a Encapsulate the CPDF_Page annots array.
• c42ec58fe Encapsulate CPDF_StructElement::m_pDict
• d691ae619 Remove unused CPDF_InteractiveForm methods
• 5680de5ef Better encapsulate CPDF_InteractiveForm::m_pFormNotify.
• 643278dc3 Add tools/resultdb to DEPS
• 61f2c1da3 Make CPDF_FormField::GetFieldAttr() a non-static method
• bf0f077f4 Avoid vending dictionary pointers from CPDF_StructElement.
• ad9ec040e Delete unused CPDFSDK_AnnotReverseIteration.
• 8176850eb Remove duplicate CPDF_FormField::GetDict() method.
• b0d3d7ab9 Stop vending dictionary pointers from CPDF_FlateEncoder
• 05332a80d Minimize vending pointers from CPDF_Action (single A)
• 6938b0c12 Fix PDF annotation drawing order.
• a1428d380 Avoid vending dictionary pointers from CPDF_AAction
• bbf75918a Add iterator stability comment to CPDFSDK_FormFIllEnvironment.
• 191afc503 Convert compile-disabled Skia tests to runtime-disabled
• 5c18b65d8 [xfa] Ignore invalid BMP pixel data offsets
• c1b28b90c [xfa] Fix image decoding across block boundaries
• 80cfc5e86 Roll Code Coverage from 06c29fb95a45 to 690a8f131004 (4 revisions)
• 48215d6b5 Roll Catapult from 7ee071132a53 to 16eb7278843e (57 revisions)

3.2.0

Changes (Release 3.2.0)

Summary (pypdfium2)

pypdfium2 commit log

Commits between 3.1.0 and 3.2.0 (latest commit first):

• b3ba56c [autorelease] update changelog and version file
• 16c0169 autorelease: improve versioning logic
• f37a0cf packaging_base: add note
• 73b4d87 packaging_base: add FIXME
• a045696 setup.py: improve error messages for invalid states
• fbeb84d setup_base: add further exists() checks
• 4cfa51c craft_wheels: address FIXME
• 4e8464f craft_wheels: add a FIXME related to the artefact stash
• 6830de0 build: replace dynamic values in resources.rc
• bcea29c readme update
• d1ae5cb Update tasks
• 15796bf render_base(): remove allocator note
• 67c8647 workflow: re-enable schedule
• b3d6702 docs/python_api: fix missing word
• 58a9d2f code comment: fix missing words

PDFium commit log

Commits between 5323 and 5337 (latest commit first):

• cfc766988 Remove unused method CPDF_Link::GetDict()
• 986e12acf Avoid vending pointers to CPDF_ColorSpace::m_pArray
• 4df3e6b4f Support runtime renderer selection in FPDFProgressiveRenderEmbedderTest
• 61eb8063f Roll third_party/skia/ 73fb58ce0..03cb8e7d8 (492 commits; 1 trivial rolls)
• 79a701c67 Roll buildtools/ d66231001..cccaf48c8 (17 commits)
• 3061eb8e9 Roll Fuchsia SDK from 9.20220830.0.1 to 9.20220929.2.1
• 807603890 Roll third_party/icu/ bbdc7d893..20f8ac695 (1 commit)
• a37c9f0c9 Roll third_party/depot_tools/ 3528d4d3c..e3ed6a8e0 (45 commits; 13 trivial rolls)
• d442cee2b Roll tools/clang/ 0a2285903..a5e0d7234 (23 commits)
• 2fd6393c6 Roll third_party/zlib/ 926ac230d..6fe4ce825 (10 commits)
• 60b3a0116 Roll tools/memory/ e5f1a8a76..98140a694 (1 commit)
• 5945cf536 Update gn_version to cc28efe62ef0c2fb32455f414a29c4a55bb7fbc4
• ae04146f0 Roll v8/ 340416187..261f00697 (180 commits)
• 73c5b8d68 Roll third_party/libjpeg_turbo/ 22f1a22c9..ed683925e (1 commit)
• 1c35dd500 Roll base/trace_event/common/ 640fc6dc8..521ac34eb (3 commits)
• 1a0312c74 Roll buildtools/third_party/libc++abi/trunk/ 48afced8a..5c3e02e92 (2 commits)
• 8cb439419 Roll buildtools/third_party/libunwind/trunk/ 42aa6de55..7ff728a97 (3 commits)
• 310c450c2 Only validate indexes for paletted BMPs
• 144d4729b Add basic unit tests for BMP decoding
• b4a309d9d Fix initial BMP pixel offset
• 2e5da1ba9 Roll third_party/freetype/src/ 8e68439a6..905270925 (10 commits)
• 0969211eb Enable FPDFEditEmbedderTest.SaveAndRender for Skia
• 9fe159009 Return retained result from GetShadingObject()
• e282b0fab Pass retained arguments to still more ctors/setters.
• 14e8a0128 Fix roll-freetype.sh to work with sed on macOS
• 5850f38fb Roll build/ 06b221966..01569374d (154 commits)
• da5d81449 Roll base/allocator/partition_allocator/ d47a38e4a..bb627f4ab (24 commits)
• 5e8077158 Roll third_party/lss/ 0d6435b73..ce877209e (8 commits)
• b01b2cdd1 Pass retained arguments to fpdfdoc/ class constructors.
• f504a7655 Fix retained argument TODOs, part 8.
• e1d691d54 Fix retained argument TODOs, part 7.
• 7870d8d06 Fix retained argument TODOs, part 6.
• c90fa00b4 Enable FPDFEditEmbedderTest.GetRenderedBitmapHandlesSetMatrix for Skia
• 6d8dfa196 Remove image_8bit_devicergb_dctdecode.pdf from the suppression list
• 6723234d3 Return retained object from CPDF_Form and CPDF_GeneralState methods.
• 4ab13cbdb Document test runner positional arguments
• 525a4270b Fix retained argument TODOs, part 5.
• 0fbd8fc96 Use retained references in fpdfsdk/fpdf_*.cpp files
• c62c31ef1 Fix retained argument TODOs, part 4.
• cda6ca689 Fix retained argument TODOs, part 3.
• cd3d3798c Rework CPDFSDK_FormFillEnvironment::JS_appResponse() callers.
• bb040742c Return retained objects frm CPDF_Image methods
• 5a3f1d31e Fix retained argument TODOs, part 2.
• 06749a7a5 Pass/return retained args from CPDF_NumberTree.
• 3df809cbc Fix retained argument TODOs, part 1.
• e08aa1ef0 Pass retained arguments to CPDF_PageObjectHolder.
• f9491f39e Fix unowned/retain ptr conversion and assignment.
• af0fcf1ed Pass retained arguments to CPDF_Action constructor
• [82b558c70](https://pdfium.googl...

3.1.0

Changes (Release 3.1.0)

Summary (pypdfium2)

• render_to() was fixed to actually allow callables. The code passage in question was unreachable due to an unguarded issubclass() check before.

pypdfium2 commit log

Commits between 3.0.0 and 3.1.0 (latest commit first):

• b6e8e6d [autorelease] update changelog and version file
• 9117f42 Noneify .raw attributes on close
• 23b2b77 Update changelog_staging
• f59dc65 Revert automatic closing via del
• b37b326 page.render_to: fix compatibility with callables
• 7d3e7ee converter: correct a docstring
• 662f12b setupsrc: slightly enhance set_version()
• 7ec0744 setup.cfg: reset sphinx to normal
• 2f3680b workflow: comment out schedule again
• fc5bcd2 document: remove note on closing
• 47d5c78 closing: check self before parents
• fb62c5a setup.cfg: temporarily block sphinx 5.2.0.post0
• b068b9e Automatically close objects via __del__ (#142)
• 69b3e9b autorelease: log beta releases, without manual summary
• f22d6f7 changelog: fix up autorelease note on pdfium version

PDFium commit log

Commits between 5316 and 5323 (latest commit first):

• 014c24029 Revert "Remove ALLOW_UNUSED macros."
• 2785e0210 Pass retained array to CPDF_Dest constructor.
• 953ff0e12 Merge FixedUninitDataVector and FixedZeroedDataVector code.
• 38c2dcc58 Make UnownedPtr and RetainPtr more similar to each other.
• 99762aa82 Re-order RetainPtr/UnownedPtr operator=() methods
• 42fc37599 Replace CFX_MemoryStream::GetBuffer() with GetSpan().
• d1f85fdbb Avoid ambiguity in CPDF_{Dictionary,Array}Locker constructors
• 11a2745cd Return retained result from CPDF_Dictionary::GetPageDictionary()
• 5b0176124 Switch from DataVector to FixedDataVector CPDF_CMap.
• 0820b6b7a Use FixedDataVector in fxbarcode/ when applicable.
• 7043b271f Add a FixedZeroedDataVector container.
• 0dc2e884d Do more input checks inside BC_DefaultPlacement.cpp.
• 065698acf Fix interaction between RetainPtr and transparent comparisons.
• 22a8e1526 Clean up CBC_DefaultPlacement.
• f10010959 Switch CPDF_TextPage::m_CharIndices back to DataVector.
• f24061bf2 Create helper function and move a constant to a header file
• 07d7f6b34 Remove ALLOW_UNUSED macros.
• ba4ecd0a5 Use UnownedPtr<> keys in CPDF_FontGlobals.
• 275ee19d0 Add transparent Contains() test to unowned_ptr_unittest.cpp
• 081fa0b6f Prevent pdfium::Contains() from falling back to linear search.
• 101e5c6be Prove whether unowned/retain pointer collections compare transparently
• 9524684f4 [Skia] Remove DCHECKs on RGB channels during upsampling
• 9b9170e56 Add unit test for transparent comparison of UnownedPtr.
• 8aad22ae5 Make CPDF_{Array,Dictionary} methods returning pointers internal-only.
• 4424a54c6 Refactor Skia renderer runtime checks.
• 7784e6a6b Runtime renderer selection for embedder tests
• 786e8f65d Add FPDFAnnot_GetFormFieldAlternateName() API

3.0.0

Changes (Release 3.0.0)

Summary (pypdfium2)

• Support model

  API-breaking changes

  • If the target page of a bookmark cannot be identified, OutlineItem.page_index is now None rather than -1, to avoid accidental reverse list indexing and to enforce that callers properly handle this case. Moreover, OutlineItem.is_closed is now None rather than False if the bookmark has no kids.
  • PdfPageObject.get_type() was replaced with a type attribute.
  • PdfPage.count_objects() was removed. Use PdfPage.get_objects() or the raw PDFium API instead.
  • If a negative index is passed to PdfDocument.new_page(), it is now interpreted in reversed direction, rather than inserting at the beginning.
  • PDFium is now provided with an external, python-allocated buffer for rendering. This has numerous advantages, most notably that callers don't need to free resources anymore. PdfPage.render_base() now directly returns a ctypes ubyte array; BitmapDataHolder has been removed.
  • Changed rendering parameters
    • annotations was renamed to draw_annots
    • colour was renamed to fill_colour and now only takes a list of 4 values for simplicity - it may not be 3 values or None anymore
    • no_antialias has been replaced with separate boolean options no_smoothtext, no_smoothimage, and no_smoothpath

  Other changes

  • OutlineItem now contains information on the number of sub-items (n_kids attribute).
  • All document-level methods that take a page index now accept negative values for reverse indexing (except the rendering methods).
  • New method PdfDocument.get_page_size() to retrieve page size by index without needing to load a PdfPage (uses FPDF_GetPageSizeByIndexF() under the hood).
  • New rendering parameters added: colour_scheme, fill_to_stroke, force_halftone, draw_forms, rev_byteorder, prefer_bgrx, force_bitmap_format, extra_flags, allocator, and memory_limit.
  • Added new render_to() functions to PdfPage and PdfDocument that take a custom bitmap converter, to transform the ctypes array to a different object. A set of built-in converters is provided with the BitmapConv class. Confer the updated API documentation for details. The previous rendering functions (render_topil() render_tobytes(), ...) are still around as aliases but might be deprecated eventually.
  • New rendering target numpy_ndarray added.
  • The pil_image rendering target now accepts a prefer_la parameter to request automatic conversion of BGRA/RGBA to LA if rendering in greyscale mode with alpha channel.
  • PdfPage.get_objects() can now recursively descend into Form XObjects.
  • Form environments are now initialised/exited on document level rather than on page rendering. In the course of this work, a segmentation fault source was eliminated, related to a formerly undocumented requirement of PDFium regarding object lifetime. Whether the segmentation fault would actually take place was dependent on Python garbage collection behaviour. This did not appear to happen under normal circumstances, so the issue remained unnoticed for a long time.

• Setup code

  • $PYP_TARGET_PLATFORM was renamed to $PDFIUM_BINARY, the value sdist was renamed to none.
  • When doing an automatic release, repository changes are now only pushed after successful wheel building, to avoid leaving the repository in an invalid state in case some earlier step fails.
  • pypdfium2 now declares a no-op setuptools extension to prevent wheel content from landing in a purelib folder. Some systems use this information to separate platform-dependent packages from pure-python packages (i. e. /usr/lib64 instead of /usr/lib). Confer PEP 427.
  • Autorelease now properly takes existing beta tags into account for its version changes.
  • PDFium's commit log is now shown with GitHub releases.
  • The wheel packaging script now restores in-tree artefacts from a possible editable install.
  • Platform files are now detected in a more robust way. If missing, a proper exception will be raised.
  • Platform data directories are now annotated with a text file storing the pdfium version, to prevent a possible mismatch between the state of version.py and the actual version of the used binary. The update and build scripts do not directly change the main version file anymore, but defer the changes to setup.py.
  • Missing platform files are now always procured implicitly on installation. If platform files exist already but are outdated, they will be updated by default. You may opt out by creating an empty file called .lock_autoupdate.txt in data/.
  • A MANIFEST.in file was added to avoid being dependent on setuptools_scm.
  • On setup, dependency checks are now only done for system commands. The wonky implicit installation of missing Python packages was removed. If opting out of build isolation, callers should properly install setup dependencies beforehand.

• Documentation

  • Rewrote the project's README.md. Added more support model examples and an extensive guide regarding the raw PDFium/ctypes API.
  • Improved docstrings and included type hints.

pypdfium2 commit log

Commits between 3.0.0b1 and 3.0.0 (latest commit first):

• d2ddcee [autorelease] update changelog and version file
• bf24758 Fix incompatibility with Python 3.6

3.0.0b1

Changes (Release 3.0.0b1)

Summary (pypdfium2)

• Support model

  API-breaking changes

  • If the target page of a bookmark cannot be identified, OutlineItem.page_index is now None rather than -1, to avoid accidental reverse list indexing and to enforce that callers properly handle this case. Moreover, OutlineItem.is_closed is now None rather than False if the bookmark has no kids.
  • PdfPageObject.get_type() was replaced with a type attribute.
  • PdfPage.count_objects() was removed. Use PdfPage.get_objects() or the raw PDFium API instead.
  • If a negative index is passed to PdfDocument.new_page(), it is now interpreted in reversed direction, rather than inserting at the beginning.
  • PDFium is now provided with an external, python-allocated buffer for rendering. This has numerous advantages, most notably that callers don't need to free resources anymore. PdfPage.render_base() now directly returns a ctypes ubyte array; BitmapDataHolder has been removed.
  • Changed rendering parameters
    • annotations was renamed to draw_annots
    • colour was renamed to fill_colour and now only takes a list of 4 values for simplicity - it may not be 3 values or None anymore
    • no_antialias has been replaced with separate boolean options no_smoothtext, no_smoothimage, and no_smoothpath

  Other changes

  • OutlineItem now contains information on the number of sub-items (n_kids attribute).
  • All document-level methods that take a page index now accept negative values for reverse indexing (except the rendering methods).
  • New method PdfDocument.get_page_size() to retrieve page size by index without needing to load a PdfPage (uses FPDF_GetPageSizeByIndexF() under the hood).
  • New rendering parameters added: colour_scheme, fill_to_stroke, force_halftone, draw_forms, rev_byteorder, prefer_bgrx, force_bitmap_format, extra_flags, allocator, and memory_limit.
  • Added new render_to() functions to PdfPage and PdfDocument that take a custom bitmap converter, to transform the ctypes array to a different object. A set of built-in converters is provided with the BitmapConv class. Confer the updated API documentation for details. The previous rendering functions (render_topil() render_tobytes(), ...) are still around as aliases but might be deprecated eventually.
  • New rendering target numpy_ndarray added.
  • The pil_image rendering target now accepts a prefer_la parameter to request automatic conversion of BGRA/RGBA to LA if rendering in greyscale mode with alpha channel.
  • PdfPage.get_objects() can now recursively descend into Form XObjects.
  • Form environments are now initialised/exited on document level rather than on page rendering. In the course of this work, a segmentation fault source was eliminated, related to a formerly undocumented requirement of PDFium regarding object lifetime. Whether the segmentation fault would actually take place was dependent on Python garbage collection behaviour. This did not appear to happen under normal circumstances, so the issue remained unnoticed for a long time.

• Setup code

  • $PYP_TARGET_PLATFORM was renamed to $PDFIUM_BINARY, the value sdist was renamed to none.
  • When doing an automatic release, repository changes are now only pushed after successful wheel building, to avoid leaving the repository in an invalid state in case some earlier step fails.
  • pypdfium2 now declares a no-op setuptools extension to prevent wheel content from landing in a purelib folder. Some systems use this information to separate platform-dependent packages from pure-python packages (i. e. /usr/lib64 instead of /usr/lib). Confer PEP 427.
  • Autorelease now properly takes existing beta tags into account for its version changes.
  • PDFium's commit log is now shown with GitHub releases.
  • The wheel packaging script now restores in-tree artefacts from a possible editable install.
  • Platform files are now detected in a more robust way. If missing, a proper exception will be raised.
  • Platform data directories are now annotated with a text file storing the pdfium version, to prevent a possible mismatch between the state of version.py and the actual version of the used binary. The update and build scripts do not directly change the main version file anymore, but defer the changes to setup.py.
  • Missing platform files are now always procured implicitly on installation. If platform files exist already but are outdated, they will be updated by default. You may opt out by creating an empty file called .lock_autoupdate.txt in data/.
  • A MANIFEST.in file was added to avoid being dependent on setuptools_scm.
  • On setup, dependency checks are now only done for system commands. The wonky implicit installation of missing Python packages was removed. If opting out of build isolation, callers should properly install setup dependencies beforehand.

• Documentation

  • Rewrote the project's README.md. Added more support model examples and an extensive guide regarding the raw PDFium/ctypes API.
  • Improved docstrings and included type hints.

pypdfium2 commit log

Commits between 2.11.0 and 3.0.0b1 (latest commit first):

• 03b379f [autorelease] update changelog and version file
• 75b972c update_pdfium: make max_workers configurable
• 69a0b77 autorelease: nevermind, always include the log
• 76fc08e Finalise changelog_staging
• 21dc9b0 Misc improvements
• b02b67b workflow: re-activate schedule
• 0378257 Slightly improve converter docs
• 6a24cc4 Revert "textpage: replace count_chars() with char_count attribute"
• 7f2e9ff Update some docs
• 18221b4 readme: suggest using get_functype()
• 1b8d7f1 tocprinter: avoid some variable assignments
• ff52551 Update tasks
• 6d00654 Make most of _utils.py public
• 2c82267 readme: fix use of non-public member
• cc70047 utils: rename bg_colour to fill_colour
• d960830 Rename back from color to colour
• 69c780f rendering: rework colour integration
• 26bf175 textpage: replace count_chars() with char_count attribute
• 038feb2 Improve string decoding
• 6b14960 docs: decide not to deprecate render_to* methods just yet
• 97d2243 Update rendering CLI
• fc736ec auto_bitmap_format(): flatten if/else tree
• c44d81a multipage renderer: add missing close() calls
• f8608e8 Set OutlineItem.is_closed to None if it has no kids
• 1f2b75b toc/OutlineItem: Add n_kids attribute
• 1093579 Provide means to plug in custom converters (#140)
• e513026 render_topil: use frombuffer() rather than frombytes()
• d17cffd autorelease: make commit log collapsible
• 9d4a3fc autorelease: allow skipping commit log
• 3df1533 autorelease: improve changelog handling f...