0.5.15 2023-01-25T06:41:44+13:00
- Handle and throw on IO.open failures
0.5.14 2023-01-19T14:59:45+13:00
- Suppress warning from PDF::COS::Coercer. See
https://github.com/pdf-raku/PDF-Class-raku/issues/21
0.5.13 2022-12-11T09:21:41+13:00
- make pdf-rewriter.raku --render option only uncompress page content
- refactor of PDF::COS::Tie multi-methods. Added !tie-container() private
method for Hash/Dict coercement.
0.5.12 2022-07-27T05:13:06+12:00
- Ensure cross reference entries are sorted by object number #35
0.5.11 2022-07-04T15:46:11+12:00
- Fix locking for consistency in PDF::COS::Util ast_coerce()
0.5.10 2022-07-04T12:46:14+12:00
- Fix pdf-checker.raku for PDF 0.5.8+
- Add PDF::COS::Dict required-entries() and aliases method
- Deprecate Hash slow-path in PDF::IO::Writer
- Fix regression in Pod::To::PDF t/render-pdf-test-sheets.t
0.5.9 2022-06-09T08:20:15+12:00
- Fix writing of Any as 'null'
0.5.8 2022-06-07T16:19:57+12:00
- Use PDF::Grammar v0.2.9+ :lite mode for lighter AST's.
- Numeric objects are no longer cast to roles PDF::COS::Int,
PDF::COS::Bool, PDF::COS::Real unless they are indirect
objects.
- Utilize PDF::Native v0.0.2+ cross reference stream
unpack-xref-stream() and pack-xref-stream() subroutines.
- Add --compat option to pdf-rewriter.raku script.
0.5.7 2022-06-02T06:26:44+12:00
- Upgrade from 32 to 64 bit indexing in cross reference streams
to support opening very large PDF 1.5+ (> 3.2Gb) files.
- Added compatibility and automatic loading of PDF::Native. Soon
to be released optional performance enhancement module.
0.5.6 2022-05-21T09:43:39+12:00
- More thread protection for array/hash construction
0.5.5 2022-05-19T12:58:40+12:00
- Fix possible memory leaks in PDF object construction
0.5.4 2022-05-12T13:37:30+12:00
- Fix a thread-safety issue in PDF::COS::Util ast-coerce() sub.
0.5.3 2022-05-12T07:59:14+12:00
- Fix a thread-safety issue in PDF::COS::Tie tie() method.
0.5.2 2022-05-11T12:58:16+12:00
- Partially back-out PDF::COS required() method to fix PDF::Class
regressions from 0.5.0 release.
0.5.1 2022-05-07T09:32:57+12:00
- Further lock-protection of dynamically 'require'd classes
for thread safety
0.5.0 2022-05-06T13:19:02+12:00
- Improved thread safety of reading and fetching objects.
-- Protect PDF::COS loader() method
-- Protect Reader ind-obj method to allow concurrent fetching
-- Better scoping of %seen + general refactor in PDF::COS::{Dict|Array}
-- Some locking and deprecations in PDF::COS::Loader
-- Added test t/threads.t
This round of changes targets concurrent PDF reading only!
0.4.25 2022-03-31T16:42:09+13:00
- Replace role() method with induce(). Only mixin punned roles.
0.4.24 2022-03-26T07:41:03+13:00
- Further fixes to latin-1 in ByteString coercements
0.4.23 2022-03-13T17:25:53+13:00
- Ensure strings are latin-1 in ByteString coercements
- Simplify date string coercements
- Add X::PDF::NoTrailer error
0.4.22 2022-03-13T17:07:05+13:00
- Initial CPAN -> zef release
0.4.22 2022-02-24T13:21:35+13:00
- Replace role-model() method with role()
0.4.21 2022-02-23T14:06:44+13:00
- Attempted fix of flapping error in PDF::COS::Tie role-model() method
0.4.20 2022-02-06T11:53:46+13:00
- Fix PDF::IO::Filter::RunLength for pending Rakudo 2022.02; niner++
0.4.19 2022-01-12T06:47:11+13:00
- Refactor Array coercements for consistency with Dict and Stream
- Fixed PDF::IO::Writer name serialization to encode '#' as #23 (not ##)
- Fixed some test fails from git checkout on Windows #31
0.4.18 2022-01-01T07:42:10+13:00
- PDF::COS::Bool now does the PDF::COS role
- Fixed PDFDoc encoding by regenerating from PDF::ISO_32000 resource:
Table_D2-PDFDocEncoding_Character_Set.json
- Fixed PDF::COS::TextString to handle non-byte strings (e.g. /ActualText
in content streams).
- Improvements to PDF::COS::TextString utf16-be encoding
0.4.17 2021-09-29T14:46:25+13:00
- Refix PDF::COS::DateTime.new
0.4.16 2021-09-28T09:21:51+13:00
- Fix to PDF::COS::DateTime.new
0.4.15 2021-07-29T12:43:43+12:00
- Add error check for zero cross reference offset
- Change Bufs to Blobs in PDF::IO::Util :&pack and
PDF::IO::Filter::Predictors
- Fix overflow error in LZW decode filter #30
0.4.14 2021-07-02T13:14:48+12:00
- Refix PDF without file associations
0.4.13 2021-07-02T12:07:07+12:00
- Fix PDF::IO coercement of Blobs
- Ensure a PDF works without file associations. E.g. when rebuilt from:
$pdf .= open: $pdf.Blob;
0.4.12 2021-06-24T11:43:44+12:00
- Add COERCE method to PDF::IO::Blob, PDF::IO::Handle, PDF::IO:Str
0.4.11 2021-04-28T12:41:23+12:00
- Fix writing of enumerated names #29
0.4.10 2021-04-20T06:27:30+12:00
- Add --stream option to pdf-rewriter.raku and PDF::IO::Reader
- Default to :!stream in save-as() method.
0.4.9 2021-04-19T06:37:06+12:00
- Add PDF::IO::Writer stream-cos() and stream-body() methods for chunked
saving to a file-handle. Apply to PDF save-as() and update() methods.
0.4.8 2021-04-15T07:09:03+12:00
- Enum agnostic array access for latest Rakudo. See
https://github.com/rakudo/rakudo/issues/4310
0.4.7 2021-04-14T06:19:49+12:00
- Tweak AST construction for latest Rakudo
0.4.6 2021-04-12T12:39:56+12:00
- Use Hash::int for PDF::IO::Reader %!ind-obj-idx (indirect
object index).
- Add $pdf.id rw accessor. In particular, this allows assignment of
consistent IDs for the benefit of PDF tool-chain testing.
0.4.5 2021-01-12T06:00:31+13:00
- Rename classes PDF::Writer -> PDF::IO::Writer and PDF::Reader
-> PDF::IO::Reader. Free up PDF::Reader and PDF::Writer for
Raku ecosystem usage.
Release 0.4.4
- Add COERCE methods to COS roles and classes. For example:
my PDF::COS::TextString $s .= COERCE: 'Hi';
or using Rakudo 2020.11+ coercion semantics:
my PDF::COS::TextString() $s = 'Hi';
Release 0.4.3
- [PDF::Reader] add abstract base exception class X::PDF apply to existing
X::PDF::* exception classes
- [pdf-rewriter.raku]
-- added --render option to render and reformat pages. This option also requires
an upstream module such as PDF::Lite or PDF::Class.
-- Hide back trace for handled errors (type X::PDF)
- [PDF::COS::TextString] Decode as UTF16-BE / PDFDoc when input is a byte-string #26
- [PDF::COS::Util] rename pack-pp() -> pack-be() [big endian]
Release 0.4.2
- Streamlined cross-reference loading; adjust to PDF::Grammar 0.2.1
- Fix 'missing required field(s)' message on a tied COS attribute.
- Fix accidental coercement of byte to name strings; also date coercements
- Fix syntax level PDF/A non-conformance issues, as reported by verapdf:
-- follow 'obj' by \n (not space) when serialising indirect objects
-- have 4 binary characters (not 3) in header comment
Release 0.4.1
- [PDF::Writer] Compacted writing of smaller dictionaries
- [PDF] .save-as() method: don't copy when updating and source and destination are the same
Release 0.4.0
- Rename modules *.pm -> *.rakumod and scripts *.p6 -> *.raku
Release 0.3.8
- [PDF::Writer] - Fixed .write-content() method to sanitise any non-latin
unicode chars in comment strings, which otherwise breaks serialization.
Choosing to map them to #Xxx where Xxx is a hexadecimal number
- [PDF] Add $.open(:type) option. Asserts that the PDF file is of a given type 'PDF', 'FDF' etc.
- [pdf-rewriter.p6] --class option changed from a Boolean to a string. Any class
can now be loaded that is based on PDF (PDF::Lite, PDF::Class, PDF::API6).
- [PDF::COS::Tie][PDF::COS::Tie::Array][PDF::COS::Tie::Hash]
Some internal renaming of attribute roles and methods. TiedHash -> COSDictAttrHOW,
TiedArray -> COSArrayAttrHOW, $attr.tied -> $attr.cos
Release 0.3.7
- [PDF::COS::Type::XRef]
-- Fixed a bug in decode-index() method.
- [PDF::Reader]
-- Reworked !load-index method
- [PDF::COS]
-- Improved required() method.
Release 0.3.6
- [PDF::Writer]
-- Fix for issue#22 Incremental updates breaking Adobe
Reader. Incremental updates to a 1.5+ PDF that has
cross reference streams are now written with cross
reference streams.
Release 0.3.5
- [PDF]
-- Add top-level .cb-finish() hook. Call just prior to
serialization
Release 0.3.4
- [PDF::Reader][PDF::Writer]
-- Allow copying of raw object ASTs from input to output
PDFs. To reduce memory usage in full-save mode. Speeds up
saving of encrypted PDFs.
-- Fixed /Prev trailer entry in incremental save
-- Fixed pdf-rewriter.p6 --decrypt option
- [PDF::Reader]
-- Respect EncryptMetadata flag in existing PDF files.
Disallow for newly created PDF files.
-- Fixed handling of encryption dictionaries when they're
not an indirect object
-- Improved loading of cross reference index
- [PDF]
-- Allow reencryption of an already encrypted PDF.
-- Disallow update of a newly encrypted or re-encrypted PDF
- [pdf-rewriter.p6]
-- added owner-check and -drm option.
Release 0.3.3
- [PDF::COS::Tie]
-- added :default trait
- [PDF::COS::Coercer]
-- fixed coercement to a PDF::COS::TextString subset
- [PDF::Reader]
-- Add revision-xrefs() method [to distinguish revision
sections from hybrid cross referefence sections]
-- Tweaked filtering of object and cross reference streams
during serialization
Release 0.3.2
- Handle null entries in stream DecodeParms entry (as per spec)
- Rename PDF::IO.substr() method to byte-str()
- `save-as()` method now copies an input PDF, then updates it incrementally.
- Add constant PDF::COS::DateString::DateRegex
- Add PDF::COS::Tie entry :key trait option
- Fix AES-128 encryption for xpdf
Release 0.3.1
- Add Tie :array-or-item trait. For the common case, where a
single item or array of items is permitted.`
- add tmp/ directory for test JSON files, etc
- Adjust for 'pdf' -> 'ast' in PDF::Grammar 0.1.6.
- Changed the top level node from 'pdf' to 'cos' in JSON
serializations
- Handle utf-16 encoding in PDF::COS::DateString
Release 0.3.0
- Renamed PDF::DAO classes to PDF::COS, where COS stands for
Carousel Object System (see
https://en.wikipedia.org/wiki/Portable_Document_Format#File_structure)
- Improvements to PDF::COS.coerce($obj, $type) multi methods.
- Fixed class-level PDF::COS::Stream.gist
- Support Hybrid PDFs that contain both cross reference tables and streams
Release 0.2.8
- [PDF::DAO]
-- better support for lists and list to array coercion
- [PDF::DAO::Loader]
-- replaced :$fallback option with :$base-class
- [PDF::DAO::Tie]
-- add mixin method
Release 0.2.7
- [PDF::DAO::Tie]
-- reintroduce method :entry(:alias) sub-trait.
-- remove type-check method
Release 0.2.6
- [PDF::DAO::Util] setup libpdf at BEGIN time
- Rename or hide overused load() method in various classes/roles
-- [PDF::DAO] [PDF::DAO::Loader] deprecate load() -> load-delegate()
-- [PDF::Reader] rename load() -> load-pdf()
-- [PDF::IO::Crypt] made load() method private
- Fix PNG predictors for BitsPerComponent < 8 (issue #18)
Release 0.2.5
- [PDF::Reader][PDF::Writer]
-- adjust to PDF::Grammar v0.1.1 shaped cross reference streams.
-- upgrade index offsets from uint32 to uint64 (for PDF files > 4.2Gb)
-- don't escape binary characters when writing literal strings
- [PDF::DAO::Util]
-- add coercements for typed arrays
Release 0.2.4
- [PDF]
renamed .save-as :update option to :preserve
- Created classes PDF::DAO::Loader and PDF::DAO::Coercer. These
replace the PDF::DAO::Delegator class.
- [PDF::DAO::Type::Xref]
streamlined index decoding and encoding
- [PDF::Writer]
tweak write-real method to match Lib::PDF::Writer.write-real
Release 0.2.3
- [PDF::DAO::Tie]
-- Simplify method accessors. direct use of AT-KEY, ASSIGN-KEY,
etc. 20% speed up of test suite run.
- [PDF::DAO::Util]
-- convert enums to ints in to-ast()
-- rename sub to-ast-native() -> ast-coerce()
- [PDF::Writer]
-- use native functions from Lib::PDF::Writer, when available.
Release 0.2.2
- [PDF::Writer] handle ' (MoveShow) and " (MoveSetShow) operators
- [PDF::IO::Handle] Open I/O as :bin
- Experimental use of Lib::PDF
-- Speed up the slowest test xt/filter-predictors_bulk.t
-- Fast packing alternative for PDF::IO::Util
Optional, and for regression purposes, at this stage. Probably
wont persue seriously until Rakudo has better JIT.
Release 0.2.1
- Renamed classes PDF::IO to PDF::IO::Util and PDF::IO::Input::*
to PDF::IO::*.
- Don't cache misses in PDF::DAO::Delegator
- PDF::IO::Filter::Predictors now handles /BitsPerComponent < 8
- Replace PDF::IO::Util resample() with pack(), unpack()
Release 0.2.0
- Renamed class PDF::DAO::Type::PDF to PDF. Restored module name
to PDF
Release 0.1.1
- [PDF::Reader] now handles very large xref tables or streams (> 65K)
- [PDF::DAO] use 'but' to apply roles to simple scalars, not 'does'.
Release 0.1.0
Renamed class namespace PDF::Storage -> PDF::IO. Used as the module
name
Release 0.0.9
- class PDF::DAO::Stream is now based on PDF::DAO::Dict
- withdrew non-standard PDF::DAO::Dict.new( $dict) positionals,
in favour of PDF::DAO::Dict.new( :$dict ) named arguments.
Same also for stream and array object constructors.
- Use of sigil-less variables, where appropriate.
- Revamped handing of PDF::DAO::Tie :entry and :index traits.
- Fixes for PDF::Storage::Input::Str (byte-string reader) and
PDF::Storage::Filter::Predictors
Release 0.0.8
- [PDF::DAO::Doc] .update( :diffs(...) ) - now supports JSON
- Renamed classes
-- PDF::DAO::Doc -> PDF::DAO::Type::PDF
-- PDF::DAO::Storage::Crypt::Doc -> PDF::DAO::Storage::Crypt::PDF
- [PDF::DAO::Type::XRef] renamed methods
-- decode-to-stage2 -> decode-index
-- encode-from-stage2 -> encode-index
- [PDF::Writer]
-- refactored 'write' multi method into node-specific methods
(write-array, write-bool, ... etc)
-- handle comments in write-op method
- [PDF::Writer] [PDF::Storage::Filter::RunLength] fixed handling
of DOS newline graphemes "\r\n"
- [bin/pdf-rewriter.pl] fixed --rebuild option
Release 0.0.7
- [PDF::Storage::Crypt] Encryption refactored to use OpenSSL. Added support for
PDF 1.6+ AES-128 encryption.
- [PDF::DAO::Doc] add saves-as( :update ) option.
The input PDF is copied to the destination PDF, which is then incrementally
updated. This should generally be faster than a regular .save-as.
- [PDF::Reader] Bug fixes to JSON serialization
- [PDF::Storage::Filter] simplified filter dispatch
Release 0.0.6
- [bin/pdf-rewriter.pl] unbust rewriting/recompression of encrypted PDFs
- [PDF::Storage::Serializer][PDF::DAO::{Array|Dict|Stream}] use Perl 6 object hashes
-- Refactored `my %h; ...; %hash{$object.WHICH}` with `my %h{Any}; ...; %hash{$object}`
(See http://doc.perl6.org/type/Hash#Object_hashes_and_type_constraints)
- [PDF::DAO::Doc]
-- added $doc.ast method. factored out from $doc.save-as
-- Added $doc.Str method; for in-memory serialization
-- $doc.save-as($target); # can now be file-name, io-path, or io-handle
-- added $.update( :diffs(...) ) option.
This allows the update body to be saved to a separate 'differences' file, as an
alternative to appending to the original PDF.
Useful in a mail-merge scenario where multiple documents are being produced
from a common template and each can be reconstituted by appending the
update file to the original template.
Possible further work:
-- JSON output of 'to' files
-- add `$doc.open( $file, :diff(...))` option - load a base pdf, then apply
differences.
- [PDF::Reader] bug-fixes:
-- avoid encrypting the encryption dictionary
-- fixed de-Linearization. Now handled by PDF::Storage::Serializer
- [PDF::Storage::Filter] bug fixes:
-- fixed ascii-85, ascii-hex and run-length encoders to add
end-of-data markers
-- withdrew LZW encoding
Release 0.0.5
- added some reader exception classes, e.g. X::PDF::BadXRef
- renamed META.info => META6.json
- changed $.delegator.find-delegate method
- removed PDF::DAO::Type
- [PDF::DAO::Delegator]
-- simplified and more general coercement to roles
-- added X::PDF::Coerce warning
- [PDF::DAO::Tie]
-- removed support for method aliases
- [PDF::Storage::Filter::Predictors]
-- fixed multi-channel (color) PNG prediction
Release 0.0.4
- encrypted document creation is now supported. RC4 (Encryption revisions 1-4)
only at this stage.
- [PDF::Writer]
-- add extra new-line after each indirect object - readability improvement
Release 0.0.3
- [PDF::DAO::Util] from-ast() sub now handles cyclical hashes and arrays
Release 0.0.2
- [META.info] Correct PDF::Storage::Filter::Predictors entry
- [PDF::Writer]
-- improved formatting of PDF output. Reduced number of new-lines in output.
--- '>> stream' are now cuddled together on a line
--- as are 'endstream endobj'
-- added $.make-body and $.make-xref methods
Release 0.0.1
- Initial release