3.0.3 - #21
Merged
Merged
Conversation
* fix: restore clipping path support in PageDrawer The Note box text in certain PDFs (e.g. WoW manual) was invisible because the decorative frame image was painted over the text without the intended clipping restriction. Root cause 1 — W/W* operators were silently ignored clip() only stored clipWindingRule but never called intersectClippingPath(). The clipping path in the graphics state was therefore never updated, so images were drawn over the full page instead of just the frame/border area. Fixed by immediately calling intersectClippingPath(linePath) inside clip(), matching the behaviour of the original Apache PDFBox. Root cause 2 — drawImage() never applied the current clipping path Even after fixing clip(), images were drawn without respecting the active clip because drawImage() never called setClip() before painting. Fixed by adding setClip() at the start of drawImage(). Regression fix — image flipping on pages with tiling patterns After the above fixes, images on pages that use a Pattern fill (/P1 scn) were rendered upside-down. The cause was drawTilingPattern(): it switches to a temporary Canvas for the pattern, calls setClip() which updates clipSaveCount to a save-level valid only for that temporary Canvas, then returns without restoring clipSaveCount. The next setClip() call on the main Canvas called canvas.restoreToCount() with a stale value, rolling the main Canvas back past the page y-flip transform, which flipped every subsequent image. Fixed by saving and restoring clipSaveCount around the tiling pattern rendering, the same way lastClip is already handled.
…r disappear. Small images with an `/SMask` (soft mask) — commonly used to render checkboxes, form widgets, and icons in PDF forms — were not being drawn on the page starting from version 3.0.1. Root cause: the new `drawBufferedImageV2()` code path (introduced in 3.0.1) included a downscaling optimization that reset the Canvas transformation matrix via `canvas.setMatrix(new Matrix())` and attempted to reconstruct positioning from the CTM only. Because the DPI/flip transform applied in `PageDrawer.drawPage()` was not taken into account, the images ended up being drawn in the wrong coordinate space (typically outside the visible canvas), so nothing appeared on the rendered page. Fix: non-stencil images (regular XObject images, including those with soft masks such as checkboxes) are now drawn via `drawBitmap()` again — the same simple, correct path used in 3.0.0. Stencil images with patterns still use the newer code path where it is needed for correct pattern masking.
# Conflicts: # README.md # gradle.properties
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Image rendering: checkboxes and small images with soft masks no longer disappear.
Small images with an
/SMask(soft mask) — commonly used to render checkboxes, form widgets, and icons in PDF forms — were not being drawn on the page starting from version 3.0.1.Root cause: the new
drawBufferedImageV2()code path (introduced in 3.0.1) included a downscaling optimization that reset the Canvas transformation matrix viacanvas.setMatrix(new Matrix())and attempted to reconstruct positioning from the CTM only. Because the DPI/flip transform applied inPageDrawer.drawPage()was not taken into account, the images ended up being drawn in the wrong coordinate space (typically outside the visible canvas), so nothing appeared on the rendered page.Fix: non-stencil images (regular XObject images, including those with soft masks such as checkboxes) are now drawn via
drawBitmap()again — the same simple, correct path used in 3.0.0. Stencil images with patterns still use the newer code path where it is needed for correct pattern masking.Additional Changes
[PDFBOX-4076] - PDFBox cannot properly handle PDF Name objects containing bytes with values outside the US_ASCII range
[PDFBOX-6150] - PDComboBox constructAppearances wrong text
[PDFBOX-6173] - NullPointerException while merging PDFs with output intents
[PDFBOX-6176] - reported number of objects (7412) is not one plus the highest object number (7410)
[PDFBOX-6178] - PdfBox renames RadioButton with Umlaut
[PDFBOX-6179] - Removing fields not possible
[PDFBOX-6186] - IllegalArgumentException when calling PDDocumentCatalog.setThreads(null)
[PDFBOX-6187] - NullPointerException in AppearanceGeneratorHelper when PDFont.getName() returns null
[PDFBOX-6189] - Usage of PDF Controlled dimension may cause OOM
[PDFBOX-6191] - Print border doesn't work if using rastering
[PDFBOX-6193] - TrueTypeEmbedder.getTag() throws StringIndexOutOfBoundsException for negative Map.hashCode()
[PDFBOX-6194] - COSStream becomes COSDictionary after save — shared XObject reference replaced by Font
[PDFBOX-6202] - Possible regression when combining multiple PDFs
[PDFBOX-6203] - Splitter.split() corrupts result PDFs
[PDFBOX-6204] - PDType1FontEmbedder sets /Length to uncompressed PFB size
[PDFBOX-6206] - Rendering fails when bad annotations array
[PDFBOX-6209] - Regression in v3.0.7 causes Splitter to extract pages with text converted to symbols
[PDFBOX-6210] - Incorrect CJK Character Extraction for Shared Glyphs
[PDFBOX-6214] - PDDocument static AWT warm-up fails hard when AWT native libs are unavailable (GraalVM native-image)
[PDFBOX-6216] - Reject OpenType fonts using CFF2 outlines