Skip to content

CHANGELOG.md

Latest commit

 

History

History
1615 lines (889 loc) · 35.2 KB

CHANGELOG.md

File metadata and controls

1615 lines (889 loc) · 35.2 KB

[Unreleased]

Add

  • Add LoadOptions struct with password, filter, and strict fields for extensible loading configuration
  • Add load_with_options, load_from_with_options, and load_mem_with_options methods (sync + async)

Deprecate

  • Deprecate load_filtered, load_from_with_password, load_mem_with_password in favor of _with_options variants

v0.38.0 (2025-08-26)

Add

  • Add enhanced PDF decryption support for encrypted documents with empty passwords
  • Add automatic decryption during document loading for better pdftk compatibility
  • Add raw object extraction before parsing to handle encrypted content
  • Add support for decrypting PDFs with compressed object streams
  • Add comprehensive test suite for PDF decryption functionality
  • Add assets/encrypted.pdf test file for decryption testing
  • Add examples demonstrating decryption capabilities (test_decryption.rs, verify_decryption.rs)

Enhance

  • Enhance Reader::read() to detect and handle encrypted PDFs automatically
  • Enhance document loading to attempt empty password authentication by default
  • Enhance object processing to decrypt objects after parsing
  • Enhance support for encrypted PDFs containing object streams

Fix

  • Fix encrypted object parsing by extracting raw bytes before decryption
  • Fix object stream handling in encrypted documents
  • Fix decryption workflow to match pdftk's approach

Implementation Details

  • Modified src/reader.rs to add load_encrypted_document() method
  • Added extract_raw_object() method for raw byte extraction
  • Added parse_raw_object() method for parsing extracted bytes
  • Store raw object bytes in Reader::raw_objects field for deferred decryption
  • Process compressed objects from object streams after decryption

v0.37.0 (2025-08-08)

Add

  • Add complete PDF object streams write support enabling 11-61% file size reduction (#XXX)
  • Add save_modern() method for easy object streams and cross-reference streams usage
  • Add SaveOptions struct with builder pattern for configuring compression settings
  • Add ObjectStreamBuilder for creating object streams programmatically
  • Add cross-reference stream support for PDF 1.5+ compliance
  • Add replace_partial_text() function for partial text replacement in PDFs
  • Add comprehensive test suite with 50+ tests for object streams functionality
  • Add object streams write capability (previously read-only)
  • Add implementation documentation in OBJECT_STREAMS_IMPLEMENTATION.md

Fix

  • Fix pdfutil build error - missing derive feature for clap dependency
  • Fix async feature compilation - 25 examples/tests failing with --all-features
  • Fix 31 clippy linting errors blocking CI with #![deny(clippy::all)]
  • Fix object compression eligibility - structural objects (Catalog, Pages, Page) now properly compressed
  • Fix trailer-referenced objects compression - only encryption dictionary excluded from compression
  • Fix linearization detection for proper Catalog handling per PDF specification
  • Fix compilation warnings

Update

  • Update to Rust 2024 edition with minimum Rust 1.85 requirement

Maintain

  • Maintain full backward compatibility - all existing APIs unchanged

v0.36.0 (2025-03-15)

Add

  • Add support for Revision 5 (#401)
  • Add more checks to the encryption/decryption logic (#399)
  • Add sanity checks for PDF encryption, add examples for decrypting/encrypting PDF files and various bug fixes (#397)
  • Add encrypt function to crypt filters
  • Add support for jiff and make both chrono and time optional features

Avoid

  • Avoid parsing encrypted object streams early and correctly parse object streams upon decryption (#385)
  • Avoid decrypting cross-reference streams (#381)

Check

  • Check if the security handler is the standard one

Clarify

  • Clarify datetime parsing logic using the PDF specification

Compute

  • Compute the file encryption key (revision 6)

Declare

  • Declare and implement crypt filters

Ensure

  • Ensure the document is actually encrypted

Fix

  • Fix unused imports
  • Fix warning for rotate example
  • Fix warnings about nom_parser
  • Fix clippy warning about operator precedence
  • Fix typo in comment

Gracefully

  • Gracefully handle the is_aes check without throwing errors (#376)

Handle

  • Handle cases where the stream objects override the crypt filter

Implement

  • Implement Document::encrypt() (#396)
  • Implement password authentication (revision 6)
  • Implement decrypt with and without password sanitization
  • Implement authentication functions
  • Implement password sanitization from string
  • Implement password algorithms 2-7
  • Implement encrypt_object function
  • Implement and use PKCS#5 padding instead
  • Implement function to parse the available crypt filters
  • Implement 256-bit AES-CBC crypt filter
  • Implement TryFrom rather than TryInto

Improve

  • Improve the AES decryption with some sanity checks (#383)

Merge

  • Merge remaining algorithms functions into PasswordAlgorithm implementation

Missing

  • Missing import to test

Only

  • Only encode EncryptMetadata when V >= 4 (#400)

Provide

  • Provide revision-agnostic functions for the password algorithms

Randomly

  • Randomly generate file encryption key for V5 in encrypt example (#403)

Recurse

  • Recurse into arrays and dictionaries to fully decrypt all strings/streams (#378)

Release

  • Release 0.36

Remove

  • Remove the old implementation
  • Remove workflow that used to enable the nom_parser feature
  • Remove nom_parser feature

Reorganize

  • Reorganize encryption code

Sanitize

  • Sanitize passwords (revision 6)

Try

  • Try decrypting with an empty password

Unpack

  • Unpack objects after decrypting object streams (#382)

Update

  • Update to nom 8.0 and nom_locate 5.0 (#402)

Update

  • update changelog

Use

  • Use a hasher instead of allocating a Vec
  • Use the new implementation to compute the file encryption key
  • Use the default stream and string crypt filter if present
  • Use Unix epoch if time feature is not enabled
  • Use get_deref for the Kids array to handle indirect references (#379)

Validate

  • Validate encryption dictionary for revision 5 (#405)

Validate

  • validate binary comment during parsing (#392)

0.35.0 (2025-01-19)

Add

  • Add test for supported color types in PDF image embedding
  • Add function for text chunks extraction. (#342)

Added

  • added binary comment as attribute and for load and write. Binary Comment is gonna be important for pdf in A/2, A/3 format. (#370)

Allow

  • Allow parsing off-spec PDF files with prefixes before the header (#362)

Also

  • Also accept ASCII85 streams without EOD marker (#354)

Fix

  • Fix clippy warning
  • Fix incorrect image data handling in PDF content stream
  • Fix BitsPerComponent calculation and improper ColorSpace mapping
  • Fix incorrect color type detection for JPEG images
  • Fix mulitplication overflow in ascii85 decode (#348)
  • Fix out of memory bug (#347)
  • Fix addition overflow (#346)
  • Fix lowercase s of Procset and no space target string(J-F-Liu#323) (#324)

Ignore

  • Ignore space after byte index of startxref (#371)

Implement

  • Implement ToUnicode for variadic len encodings (#328)

Improve

  • Improve JPEG processing efficiency by avoiding unnecessary decode (#345)
  • Improve cmap parsing and internal error handling (#335)

Inline

  • Inline images (#356)

Keep

  • keep existing values when extending dictionary (#322)

Properly

  • Properly support document prefixes (#365)

Refactor

  • Refactor and optimize image processing logic in xobject.rs

Release

  • Release 0.35

Remove

  • remove misleading Object::as_string (#350)

Remove

  • Remove superfluous ref keyword (#361)
  • Remove pom parser (#355)
  • Remove /Prev from trailer (#333)

Replace

  • Replace debug assert with Result (#349)
  • Replace unwrap with error handling (#351)

Rework

  • Rework errors (#358)

Specify

  • Specify minimum Rust version in Cargo.toml (#320)

Support

  • Support UTF-16 encoding for bookmark titles with non-ASCII characters (#364)
  • Support AES encryption and revision 4 (#343)

Throw

  • Throw error if xref stream cannot be uncompressed (#339)

Update

  • update changelog

v0.34.0 (2024-08-31)

Add

  • Add ASCII85 decoding (#317)
  • Add text extraction based on ToUnicode cmap (#314)
  • Add error handling to object stream (#299)
  • Add PDFDocEncoding (#296)

Cleanup

  • Cleanup comments and cargo fmt (#290)

Detect

  • Detect reference cycles when going through trailers (#308)
  • Detect reference cycles when parsing streams (with nom_parser) (#300)
  • Detect reference cycles when collecting page resources (#298)

Fix

  • Fix unicode fonts extraction in extract text example. (#315)
  • Fix clippy warings

Implement

  • Implement encoding and decoding of text strings (PDF1.7 section 7.9.2.2) (#297)

Improve

  • Improve error handling (#307)

Refactor

  • Refactor get_or_create_resources() (#291)

Release

  • Release 0.34

Replace

  • Replace unwrap with returning error (#310)
  • Replace LinkedHashMap with IndexMap (#293)

Update

  • Update dependencies (#309)
  • Update readme of pdfutil (#295)

v0.33.0 (2024-08-31)

Accept

  • Accept comments in content parsing (#261)

Added

  • Added a new feature to get images info from the pdf page. (#275)

Async

  • Async Examples (#266)

AsyncReader

  • AsyncReader (#265)

Fix

  • Fix parse outline failed, the key ’D‘ might be an object id (#274)
  • Fix parse outline failed(#270) (#271)

Indexmap

  • indexmap use in TOC for sorted TOC (#267)

Release

  • Release 0.33

Replace

  • Replace md5 with md-5 (#272)

v0.32.0 (2024-08-31)

Add

  • Add debug format for hexadecimal (#240)

Added

  • Added big generation value parsing (#257)

Added

  • added object parse to get_page_fonts (#249)
  • added meta info decryption (#237)

Fix

  • Fix clippy warning and format code
  • Fix clippy warnings
  • Fix typo in README.md (#251)

Fixed

  • Fixed parsing of the PDFs with incorrect xrefs to indirect objects (#254)

Fixed

  • fixed clippy issues (#238)

Handle

  • Handle references to arrays in get_page_contents() (#245)

Object

  • Object and related types implement PartialEq (#236)

Release

  • Release 0.32

v0.31.0 (2023-05-10)

Add

  • Add example of page rotation (#230)
  • Add decryption of documents using RC4 encryption. (#228)

Annotate

  • Annotate feature usage (#229)

Fix

  • Fix typo in README.md (#233)

PDF

  • PDF 2.0 is now a free specification

Release

  • Release 0.31

Remove

  • Remove extraneous Q operation from insert_image (#227)

v0.30.0 (2023-04-09)

Add

  • Add support for extracting TOC, Outlines and NamedDestinations (#211)
  • Add example extract_text (#212)
  • Add get_encrypted and is_encrypted (#210)
  • Add load_filtered method (#198)
  • Add as_string method to Object (#196)

Adding

  • Adding Comments to examples (#220)

Fix

  • Fix clippy warning
  • Fix cliippy warnings
  • Fix datetime using time crate
  • Fix Cargo.toml (#213)
  • Fix ci build issue (#209)
  • Fix extract_text to split text at word boundaries.
  • Fix embed_image feature

Make

  • Make some more objects public. (#199)

Readd

  • Readd accidently deleted pdf files in assets (#204)

Release

  • Release 0.30

Remove

  • Remove obsolete lifetime

Replace

  • Replace unmaitained encoding crate with encoding_rs (#222)

Set

  • Set default to nom_parser and rayon (#208)

Update

  • Update time dependency (#206)
  • Update nom dependency
  • Update time dependency
  • Update edition and some dependencies.

v0.29.0 (2023-04-09)

Add

  • Add function get_page_annotations and include an example (#184)

Added

  • Added documentation and improved tests (#178)

Allow

  • Allow mutable access to the document catalog (#189)

Extend

  • Extend match layout change and Full bookmark example in merge. (#179)

Fix

  • Fix nom parser
  • Fix clippy warnings
  • Fix add_barcode example
  • Fix Incremental.pdf
  • Fix documentation issues and make README testable (#171)
  • Fix pdfutil build error
  • Fix extend definition confusion bug (#161)

Fixed

  • Fixed #175 and some clippy issues. (#182)

Guard

  • Guard example based on if the "parser" feature is enabled (#173)

Made

  • made XREF parser accept an optional space character after 'xref' (#167)

Make

  • Make add_xobject follow references (#187)
  • Make xref public ,fix line endings and Fix Xref output so Adobe will open them again. (#181)

Merge

Release

  • Release 0.29
  • Release 0.28

Remove

  • Remove --no-default-features test

Remove

  • remove unneccessary time 0.1 dependency (#163)

Reorder

  • Reorder Pages before Renumbering Objects. (#193)

Support

  • Support Incremental Updates (#176)

Switch

  • switch to single-precision floating point (#190)

Update

  • Update itoa dependency to 1.0 (#162)

v0.27.0 (2021-12-16)

Add

  • Add GitHub Actions build matrix (#127)
  • Add Change Log

Added

  • Added Object::as_float() to convert numerical values to float. (#124)
  • Added Object::as_bool (#123)

Avoid

  • Avoid panic when encounters negative stream length

Bookmarks

Change

  • Change indent_style to space

Check

  • Check stream length

Do

  • Do not limit Real precision to two digits (#155)

Fix

  • Fix document save race in parser_aux::load_and_save and creator::create_document (#151)
  • Fix clippy warnings & add clippy build job (#128)

Preserve

  • Preserve the eol characters in literal strings (#131)

Reduce

  • Reduce allocation by reusing the iterator (#129)

Release

  • Release 0.27
  • Release pdfutil 0.4

Replace

  • Replace lzw with weezl (#140)

Return

  • Return early on error in Stream::filters (#130)

Unwrap

  • Unwrap the text (#119)

Update

  • Update nom to 6.0 (#126)

v0.26.0 (2020-09-29)

Add

  • Add as_str, as_str_mut methods to Object (#107)

Dtoa

  • dtoa may write real number in exponential format which is not allowed in PDF

Genericize

  • Genericize Content to allow AsRef<[Operation]> (#111)

Make

  • Make pom dependency optional (but default) (#112)
  • Make rayon dependency optional (#108)

Merge

  • Merge document PDF logic with some fixes (#117)

Various

  • Various improvements, updated libraries and image features (#118)

v0.25.0 (2020-06-25)

Add

  • add indexing checks (#98)

Add

Bugfix

  • Bugfix for xref_start. (#105)

Check

  • check that the buffer is big enough for startxref (#93)

Create

  • Create rust.yml (#104)

Extend

  • extend recursion limit to non-local references (#100)

Fix

  • Fix compilation error&test error (#102)

Keep

  • keep looking for the last pattern (#94)

Limit

  • Limit allowed bracket depth. (#97)

Limit

  • limit recursion to the number of objects (#92)

Move

  • Move bracket depth checking into parsers. (#101)

Release

  • Release 0.25

Return

  • Return Result from as_array_mut() (#106)

Update

  • Update itoa and linked-hash-map (#91)

v0.24.0 (2020-02-17)

Compute

  • Compute an accurate iterator size when the page tree is sane.

Fix

  • Fix datetime parser (#89)

More

  • More permissive datetime parsing (#90)

Release

  • Release 0.24

Validate

  • Validate expected id in pom parser.
  • Validate the expected id when reading indirect objects.

v0.23.0 (2019-07-14)

Adapt

  • Adapt pom parser.

Add

  • Add error descriptions.
  • Add a proper error type and remove some more panics.

Allow

  • Allow loading a document from a memory slice.

Avoid

  • Avoid allocating an intermediate collection for iteration.
  • Avoid unwraps when already returning an Option for failure.

Error

  • Error signaling around compression and image handling.

Escape

  • Escape fix (#68)

Export

  • Export dereference function as it is useful for PDF consumers.
  • Export filters module.

Get_font_encoding

  • get_font_encoding seems more at home with Dictionary.

Handle

  • Handle stream filter chains (#66)

Hex

  • Hex fix (#67)

Implement

  • Implement LZW decompression.

Improve

  • Improve hex parsing performance.

Make

  • Make a page iterator.
  • Make Reader::read consume the Reader.
  • Make content operations faillible.

Protect

  • Protect against reference loops.
  • Protect against a corrupted page tree.

Refactor

  • Refactor a bit to allow a utility function.

Release

  • Release 0.23.0

Remove

  • Remove intermediate assignation.
  • Remove unsafe code around FilterType.
  • Remove unsafe code on get_object_mut.
  • Remove some 'if let' for readability.
  • Remove more panic paths in xref parsing.

Replace

  • Replace unwraps in processor.rs.

Return

  • Return results when appropriate.

Separate

  • Separate decompression into two functions.

Take

  • Take care of panic that I actually hit on the pom side.
  • Take care of creator.rs.

Unify

  • Unify buffer creation.

Use

  • Use lifetime ellision.
  • Use TryInto.
  • Use writeln where appropriate.
  • Use error enum in reader.
  • Use stable cloned.

v0.22.0 (2019-05-13)

Add

  • Add parsing benchmark.
  • Add nom dependency.

Also

  • Also test with nom parsing feature enabled.

Array

  • Array and dictionary parsing.

Avoid

  • Avoid using format! when writing.

Be

  • Be explicit about trait objects.

Boolean

  • Boolean and null parsing.

Content

  • Content parsing.

Duplicate

  • Duplicate pom parser for incremental replacement with nom 5.

Ease

  • Ease off on rayon a bit.

Escape

  • Escape sequence parsing.

Extern

  • extern crate is not required anymore with 2018 edition.

Fix

  • Fix last ugly parser.
  • Fix octal parser.
  • Fix pdfutil build

Float

  • Float parsing.

Header

  • Header parsing.

Hex

  • Hex string parsing.

Indirect

  • Indirect object and stream parsing.

Literal

  • Literal string syntax.

Make

  • Make sure Stream.start_position is relative to the whole file.

Merge

  • Merge remote-tracking branch 'upstream/master' into nom5
  • Merge remote-tracking branch 'upstream/master' into nom5

More

  • More 2018 edition lints.
  • More cleanup.
  • More cleanup.
  • More simplifications.

Object

  • Object id and reference parsing.

Octal

  • Octal and hexadecimal parsing.

Parallel

  • Parallel object stream parsing.

Rayon

  • Rayon usage proof of concept.

Release

  • Release 0.22.0

Remove

  • Remove pom dependency in tests.

Replace

  • Replace name parser.

Resolve

  • Resolve name collisions.

Simplify

  • Simplify lifetime annotations.

Slowly

  • Slowly replace the cute pom parser with nom.

Trailer

  • Trailer and xref start.

Turns

  • Turns out "contained" already exists in nom.

Unify

  • Unify both variants of the parsing functions.

Use

  • Use a BufWriter when saving to path.
  • Use parse_at(&self.buffer, offset) to read indirect_object
  • Use nom digit testing functions.
  • Use lifetime ellision.
  • Use nom sequence operators.

Useless

  • Useless move.

Xref

  • Xref stream and trailer parsing.
  • Xref parsing.

v0.21.0 (2019-04-26)

Avoid

  • Avoid allocating a String.

Check

  • Check offsets read from file to avoid panics
  • Check and correct Size entry of trailer dictionary

Clean

  • Clean up bytes_to_string, string_to_bytes iterators

Fix

  • Fix clippy warnings
  • Fix .editorconfig

Fixed

  • fixed finally

Redundant

  • Redundant imports with 2018 edition.

Release

  • Release 0.21.0

Update

  • Update example
  • Update Cargo.toml

Use

  • Use env_logger in pdfutil

v0.20.0 (2019-03-07)

Release

  • Release 0.20.0

Replace

  • Replace println with log macros

Use

  • Use Rust 2018
  • Use pom 3.0

v0.19.0 (2018-10-24)

Allow

  • Allow xref section has zero entries

Dictionary

  • Dictionary key type changed to Vec

Format

  • Format code with rustfmt

Improve

  • Improve codestyle (simplify loops, remove closures, use is_empty() etc.)

Move

  • Move image dependency to embed_image feature

Release

  • Release 0.19.0

Skip

  • Skip corrupt deflate stream

v0.18.0 (2018-10-05)

Able

  • Able to read stream when it's length is in object stream

Adress

  • Adress timezone formatting problem from #34

Insert

  • insert image on page

v0.17.0 (2018-09-19)

Make

  • Make chrono crate optional

Release

  • Release 0.17.0

Update

  • Update add_barcode example

v0.16.0 (2018-09-18)

Add

  • Add form xobject to page
  • Add extract_stream subcommand

Compress

  • Compress created Form xobject

Compress

  • compress page content after change

Fix

  • Fix collect_fonts_from_resources for referenced resources
  • Fix add xobject to page resources as direct object

v0.15.3 (2018-09-14)

Decompress

  • Decompress Form XObject

Disable

  • Disable auto format markdown

Fix

  • Fix bug in reading incremental updated document
  • Fix build warning
  • Fix string_to_bytes method

Hexadecimal

  • Hexadecimal strings can contain white space.

Remove

  • Remove println in extract_text

Update

  • Update example code
  • Update example

v0.15.0 (2018-02-04)

Add

  • add get_object_mut
  • add method as_array_mut

Extract

  • Extract text from specified pages

Replace

  • Replace text of specified page

v0.14.1 (2017-11-03)

Add

  • Add impl From<_> for Object for more numeric types
  • Add an Object::string_literal constructor
  • Add a dictionary! macro that creates a Dictionary
  • Add impl From<ObjectId> for Object creating Object::Reference

Derive

  • Derive Clone for lopdf::Document

Release

  • Release 0.14.0

Remove

  • Remove the Seek bound on Document::save_to

v0.13.0 (2017-10-02)

Avoid

  • Avoid decompress flate stream which has Subtype

Debug

  • Debug with lldb

Fix

  • Fix get_object for created document

Ignore

  • Ignore invalid objects when reading all object in xref table

Impl

  • impl fmt::Debug for Object

Pdfutil

  • pdfutil add extract_pages command

Read

  • Read optional space at the end of xref subsection header line

Release

  • Release 0.13.0
  • Release 0.12.0

Store

  • Store compressed stream objects and normal objects together

v0.11.0 (2017-08-21)

Release

  • Release 0.11.0

Use

  • Use itoa and dtoa to improve writing performance

v0.10.0 (2017-07-20)

Added

  • Added optional allows_compression for Stream object

Release

  • Release 0.10.0

v0.9.0 (2017-05-24)

Add

  • Add pdfutil readme

Added

  • Added unit test for load_from() and save_to()
  • Added Document::with_version + refactored save() and load()
  • Added Debug trait for lopdf::Document

Apply

  • Apply multiple operations in one command

Build

  • Build with Rust stable
  • Build with Rust beta

Fix

  • Fix delete_zero_length_streams

Fixed

  • Fixed unit tests
  • Fixed breaking API changes

Release

  • Release 0.9.0

v0.8.0 (2017-03-16)

Change

  • Change Name(String) to Name(Vec)

Delete_object

  • delete_object and delete_unused_objects

Get_pages

  • get_pages and delete_pages

Handle

  • Handle zero length stream

Release

  • Release 0.8.0

Traverse

  • Traverse objects from trailer recursively

v0.7.0 (2017-03-07)

Add

  • Add Content::decode() function

Build

  • Build on Rust 1.17

Create

  • Create String object for DateTime

Parse

  • Parse PDF datetime value

Read

  • Read xref stream in hybrid-reference file

Update

  • Update create_document example
  • Update README

v0.6.0 (2017-02-16)

Add

  • Add Stream::decompressed_content() method

Read

  • Read previous Xrefs of linearized or incremental updated document

v0.5.0 (2017-02-10)

Add

  • Add size field to Xref
  • Add Xref struct

Decode

  • Decode PNG frame after FlateDecode

Read

  • Read compressed objects from object stream
  • Read xref stream

Update

  • Update README

Use

  • Use pom 0.9.0

XrefEntry

  • XrefEntry as enum type

v0.4.0 (2017-01-29)

Add

  • Add Operation constructor
  • Add modify_text test
  • Add travis-ci build status
  • Add FAQ in Readme
  • Add print_xref_size() for debuging

Decode

  • Decode content stream

Encode

  • Encode content operations

Fix

Optimize

  • Optimize parser code

Solve

  • Solve mutual reference problem between Pages and Page objects

Trigger

  • Trigger new release to pass build on docs.rs

Update

  • Update create PDF example

v0.3.0 (2017-01-18)

Add

  • Add compress/decompress subcommands to pdfutil

Create

  • create PDF parser using pom instead of nom

Dictionary

  • Dictionary preserve key insert order

Update

  • Update README
  • Update parser to use pom 0.6.0

Use

  • Use reader to get stream length if it is a reference object

v0.2.0 (2017-01-05)

Add

  • Add pdfutil program

Fix

  • Fix parsing PDF array error

Improve

  • Improve documentation

v0.1.0 (2016-12-23)

Editor

  • Editor config

Impl

  • impl Document add_object method

Improve

  • Improve Document::save functional type

Initial

  • Initial commit

PDF

  • PDF objects and document definition

Parse

  • Parse and load PDF document

Read

  • Read objects from xref table instead of sequentially from file stream

Save

  • Save PDF document to file

Store

  • Store max_id as a field of document