2.8.1 (April 2025)

New features:

* -remove-struct-tree removes whole structure tree
* -mark-as-artifact marks a whole file as an artifact

Extended features:

o -obj JSON output
* -obj/-obj-json can follow a chain from an object number
* -obj/-obj-json can explore through arrays and name/number trees
* -print-dict-entry, -remove-dict-entry and -replace-dict-entry
  can follow a chain from each found dictionary entry
* Much more of the PDF/UA Matterhorn verification suite implemented
* PDF/UA verifier now returns extra information with most failures
* Font lister now returns the font itself
o Allow %objnum in output name for -extract-images
* -show-boxes, -trim-marks, -table-of-contents, -add-text
  -add-rectangle, -impose, -twoup, -twoup-stack now preserve
  PDF/UA and PDF/UA-2 compatibility
* Preserve PDF/UA and PDF/UA-2 when drawing on existing PDF
* Preserve PDF/UA and PDF/UA-2 with -stamp-on and -combine-pages
o Allow -scale-stamp-to-fit for -combine-pages
* Allow -underneath for -combine-pages
* Preserve roundtripping of names destinations / actions
* -preserve-actions allows PDF/UA-2 style roundtripping of
  outlines

Fixes:

o Harden auto-compression against malformed streams
o Add backup JPEG dimensions method in Cpdfimage
o Correct object stream behaviour when splitting
o Fix margin calculation for -table-of-contents

* = Supported by a grant from NLnet

2.8 (December 2024)

New features:

o New -center-to-fit centres pages on a given paper size
o New -jpeg-to-jpeg-scale and -jpeg-to-jpeg-dpi
o Rasterize PDFs by calling out to GhostScript
o Extract pages as PNG/JPEG by calling out to GhostScript
o Replace stream content with -replace-stream

Extended features:

o Expand page characteristics to cover Art, Trim, Bleed
o Add Piece Info to -composition[-json]
o Add @b<n>@ for trimming bookmark text to given length
o Allow bold, italic, colours for JSON bookmarks
o Show OpenAction in -info
o Show more form information in -info
o Allow JSON / PDF syntax in dict processing and object exploration
o Show %Bookmark text when stamping text
o Change units for -info[-json] and -page-info[-json]
o Optionally add dot leaders to tables of content
o Add -collate-n to extend -collate to multiple pages at once

Fixes:

o Clean up @B implementation for -split-on-bookmarks
o -merge-add-bookmarks now has proper titles for images
o Font operations now include fonts within xobjects
o Image extraction now includes images within xobjects within xobjects
* HTML manual now ranks equally with PDF manual

* = Supported by a grant from NLnet

2.7.2 (October 2024)

New features:

* New -args-json for better control files
* New -replace-obj to edit dictionary entries via chain
* Create PDF/UA files from scratch with -create-pdf-ua-[1|2]
* Create structure information for files with -draw
* Draw can now make paragraphs with -para, -paras
* Add structure information to -typeset
* -typeset can make PDF/UA documents
* -jpeg, -png and friends can make PDF/UA documents
* Merge PDF/UA-1 and PDF/UA-2 files, preserving standards compliance
o -stretch scales without preserving aspect ratio
o -redact removes whole pages

Extended features:

o Rectangles may be specified as x y x' y' rather than x y w h

Fixes:

o Removed setting of Producer field in AGPL version
* Removed long-deprecated -control
* More compact -print-struct-tree
o -image-resolution-json was listing all images

* = Supported by a grant from NLnet


2.7.1 (July 2024)

New features:

o Build PDF files from JPEG2000 (.jp2, .jpf/.jpx) files
* Mark a file as PDF/UA compliant, or remove such marking
* Partial verification to PDF/UA via the Matterhorn protocol
* Extract, edit and reapply document structure tree
* Print structure tree summary
* Split structure tree when splitting PDF to save size
* Combine structure trees when merging or stamping PDFs
* Set the natural language of a file

Extended features:

o Allow -obj to look up nested PDF information
* Merge structure trees better when merging files
* Report top-level natural language on -info
* Report mark information dictionary contents on -info

Backward-incompatible change:

* -process-struct-tree replaces -no-process-struct-tree
newly introduced in previous version

* = Supported by a grant from NLnet


2.7 (February 2024)

New features:

o Split files to max size with -split-max
o Spray splits a file to multiple outputs by alternating pages
o List document and page info in JSON format
o List page labels in JSON format
o List fonts in JSON format
o Identify PDF/A, PDF/X, PDF/E, PDF/VT, PDF/UA in -info
o Identify AcroForm in -info
o Extract font files from a document
o List images on a page with -list-images[-json]
o Chop pages up into sections with -chop
o Build PDF files from JBIG2 streams, including globals
o Reprocess images within PDFs to further compress them
o Extract streams to disk
o Explore PDFs by printing objects
o Shift page boxes with -shift-boxes

Extended features:

o -list-images-used[-json] extends -image-resolution
o Use -raw with -extract-images to get PNMs
o -extract-images can extract JBIG2 images and their globals
o More PNGs - greyscale 1, 2, 4, 8, 16bpp and RGB 16bpp
o Report number of annotations in -page-info
o Specify image based only on file extension
o -squeeze updates old compression methods
o Show page size summary in -info
o Add -no-process-struct-trees to prevent merging of structure trees

Fixes:

o Added opam file in-source
o Fixed -set-annotations with page links
o Allow Exif JPEGs as well as JFIF ones in -jpeg and -draw-jpeg
o Only compress a stream if it actually makes it smaller


2.6.1 (September 2023)

o Fixed regression in UTF8 text with -add-text


2.6 (July 2023)

New features:

o Create new PDF documents or draw on existing ones with -draw
o Embed TrueType fonts with -load-ttf
o Embed the 14 standard fonts if requested
o Add links to parts of text with -add-text as %URL[|]
o Convert JPEGs and PNGs to PDFs with -jpeg and -png
o Export, import, and thereby round-trip annotations via JSON
o Show composition of PDF with -composition[-json]
o Use page labels like <iii> and <A-2> in page specifications

Extended features:

o Allow -utf8 with -split-bookmarks -o @B.pdf to produce UTF8 filenames
o -merge-add-bookmarks now works with unicode filenames
o Better transformation of some annotation types
o -list-annotations[-json] now respects page range
o Merge now merges structure trees (tagged PDF)
o Merge now rewrites clashing name tree entries
o Preserve /GoTo actions in bookmarks when merging
o UTF8 option for JSON output
o -info now shows object stream, /ID data, page mode and layout
o More options for viewer preference control
o More default colours, by using the CSS colour list

Fixes:

o Updated Yojson to remove dependency on Stream, ready for OCaml 5
o -typeset was opening files in text mode, instead of binary
o Fixed behaviour of -squeeze-no-pagedata / -squeeze-no-recompress 
o Significant improvements to malformed file reading
o Allow DUP page specifications to use larger numbers
o Reworked functions transforming pages to better preserve patterns


2.5.1 (January 2022)

o Fix a regression where standard fonts could not be chosen


2.5 (January 2022)

New features:

o Can read as well as write PDFs in JSON format with -j
o New operation -typeset typesets a text file as a PDF
o New operation -table-of-contents makes table of contents from bookmarks
o New operations -impose and -impose-xy for document imposition
o New operation -print-font-table gives (charcode, unicode, glyph name) triples
o New -print-dict-entry operation prints values for a given key
o New -replace-dict-entry function to search & replace e.g URLs
o Prepend e.g 2DUP to a page range to make 1,2,3 --> 1,1,2,2,3,3 etc.
o Prepend NOT to a page range to invert it
o Output annotations in JSON form with -list-annotations-json
o Output bookmarks in JSON format with -list-bookmarks-json
o Load bookmarks in JSON format with -add-bookmarks-json
o New option -collate to collate pages when merging
o Text added in existing fonts is now encoding-aware (plus new raw mode)

Extended features:

o Extend -remove-dict-entry to allow search
o Annotation /QuadPoints processed in addition to /Rect when transforming pages
o When adding text or graphics, may choose CYMK or Grey instead of RGB
o The -list-fonts operation now obeys the range
o Can now search for a font by real name with -font
o Basic merging of AcroForms when merging documents
o Add -idir-only-pdfs to restrict -idir to just files ending in .pdf
o Option -debug-force now properly documented as -decrypt-force

Internal changes:

o Switch to Yojson for faster/better JSON input/output
o Environment variable CPDF_REPRODUCIBLE_DATES for testing
o Environment variable CPDF_DEBUG for -debug
o Effectively make stderr unbuffered
o Split cpdf.ml into separate modules


2.4 (June 2021)

o New operation -extract-images
o New operation -output-json et al. to export PDF files in JSON format
o New operations to manipulate Optional Content Groups
o New operation -stamp-as-xobject to add one PDF as an xobject in another
o Optional Content Groups now preserved when merging / stamping pages
o Listing, coalescing and modfiying Optional Content Groups.
o New -labels-progress option to improve page labels interface
o Appearance streams transformed when transforming annotations
o Bookmark destination positions transformed when transforming pages
o No longer depends on Bigarray or Unix modules


2.3 (patchlevel 1, December 2019)

o Fixed bug which prevented -info working on encrypted files
o Allow -shift with -add-text for additional adjustment
o Prepend and postpend directly to page content streams


2.3 (October 2019)

o Directly set and remove Trim, Art, and Bleed boxes
o Dump attachments to file
o Extended bookmark format, preserving all bookmark information
o New -pad-with, -pad-mutiple-before options
o Set or create XMP metadata
o Remove graphics clipping
o Extended support for reading malformed files
o Embed missing fonts by calling out to gs
o Set bookmarks open to a given level
o Create PDF files from scatch
o Remove single images by name
o Add trim marks


2.2 (patchlevel 1)

o Fix for inability to read null objects in streams
o Workaround for Adobe "Error 21" on re-saving encrypted files
o More efficient bookmark operations on files with many pages
o New operation -hard-box to clip contents to a given box


2.2 (March 2017)

o Perform arithmetic on dimensions when specifying size or position
o Add simple rectangles to PDF to blank things out
o Stamping operations now preserve annotations
o Decryption fully on-demand for speed.
o -keep-l keeps existing linearization status
o -remove-dict-entry to remove the contents of a dictionary entry
o -topline in addition to -midline
o -producer and -creator may be used to alter producer and creator
in conjunction with any other operation
o -topline and -midline now apply to stamps
o -list-spot-colours
o -bates-at-range and -bates-pad-to
o -print-page-labels
o -squeeze replaces smpdf
o Preserve more sharing of data when doing merges and page alterations


2.1 (December 2014)

o Encryption now much, much faster
o Faster parsing of delayed object streams on large files
o -decompress now leaves object streams decompressed
o Select pages by landscape-ness or portrait-ness in a page range
o New -open-at-page and -open-at-page-fit option to set the open action
o New -recrypt option to re-encrypt output
o Reads XMP metadata and outputs it on -info
o New -center position for text
o -stamp can now use positions, just like stamping text
o Better handling of permissions for files with user passwords
o Linearization excised
o Can encrypt or recrypt output of -split and -split-bookmarks now
o -args replaces -control with better behaviour
o Can scale a stamp to fit before stamping with -scale-stamp-to-fit