Revision history for Perl module EBook::Ishmael.

2.05  Jul 07 2026
  - Fix cleaning up HTML for formats that use XML namespaces like EPUBs and
    FictionBook2.

2.04  Mar 27 2026
  - ishmael now attempts to clean up HTML prior to dumping by removing
    unnecessary elements like <img> and <style>.
  - Remove character encoding debug print when dumping text files.
  - Change test EPUB's date to fix test failures for systems with Time::Piece
    1.38.
  - Skip some tests that fail for Time::Piece <= 1.39.

2.03  Mar 10 2026
  - Fix MobiHuff and PDB::Record package names.

2.02  Mar 10 2026
  - ishmael now attempts to heuristically guess an ebook's character encoding
    when the encoding is not explicitely specified instead of blindly assuming
    UTF-8.
  - Imporve PDF date parsing.
  - Improve image and cover dumping.
  - Add support for dumping CHMs via hh.exe on Windows.
  - 'pjson' and 'pxml' --metadata options are completely removed. Please use
    'json' and 'xml' instead.

2.01  Feb 19 2026
  - Fix time parsing for different versions of Time::Piece.

2.00  Feb 17 2026
  - Overhauled how ishmael handles metadata. The main update this overhaul
    brings is that ishmael no longer allows multiple values for fields that
    would not make sense for having multiple different values.
    - The following fields are now treated as single item fields: title,
      software, format, created, modified, id, and description.
    - List fields are now always treated as arrays in -m json.
    - Singular fields are not treated as a list of <item> nodes in -m xml.
  - ishmael now tries to convert timestamps into a consistent format for the
    -m option.
  - Add support for using pandoc for formatting.
  - Fix executing programs with spaces in their path names.
  - ishmael can now identify EPUB version.
  - Improve image format identification.
    - Add support for identifying AVIF and JXL image files.

1.09  Sep 16 2025
  - Ported PalmDoc/Mobi decoding code to XS (C), vastly improving performance
    for reading some PalmDoc and Mobi ebooks.
    - If you wish to use the previous pure-Perl version of the decoding code,
      you can pass NO_XS=1 to perl Makefile.PL to disable the use of XS.
  - Switched from using JSON to JSON::PP for JSON serialization. This reduces
    ishmael's dependencies, as JSON::PP is a part of Perl core.

1.08  Sep 01 2025
  - Add support for reading ebooks from standard input when '-' is given as the
    input path.
  - -m json and -m xml now act the same as -m pjson and -m pxml.
  - The 'pjson' and 'pxml' metadata options are deprecated and will be removed
    sometime in a future release. Please just use 'json' and 'xml' instead.
  - ishmael can now dump PDF covers with ImageMagick's convert.
  - For ebook formats that ishmael uses system commands to process, file names
    with single quotes (') are properly quoted.
  - Using '.*' for cover image dumping has now been removed.
  - Slightly improve startup time.
  - Fix documentation typos.

1.07  May 23 2025
  - Add Fishmael: the new mascot/logo for ishmael.
  - Add -N|--no-network option: disable fetching of remote resources when
    reading an ebook.
  - Add support for generic Zip ebook archives.
  - Fix -I|--file-encoding.

1.06  Apr 15 2025
  - Added -I|--file-encoding option: Specify the character encoding of the ebook
    file's text. This option only applies to ebook formats that do not specify
    their character encoding.
    - The addition of this option means that non-UTF-8 PalmDocs, zTXTs, and
      plain text files are now better supported.
  - Added -t|--text option: Manually specify formatted text dump mode (the
    default mode).
  - Fix typos in documentation.

1.05  Apr 01 2025
  - Fix indefinite hang in formatted text dump for pre-5.34 perls.
  - Fix character mapping errors in -e option.
  - When using the -e option, characters that cannot be converted are silently
    discarded, rather than causing a warning or being substituted with a '?'
    character.
  - During ishmael tests, write output to temporary file rather than to stdout.
  - Add Encode to prerequisites.

1.04  Mar 28 2025
  - Added the -e|--encoding option: Specify the output encoding for outputted
    text.
    - Also added the ISHMAEL_ENCODING environment variable.
  - Improve UTF8-handling for plain text ebook formats (text, zTXT, PalmDoc).
  - Added support for the chawan web browser as a potential HTML formatter.
  - When specifying cover image output, '.-' should be used now instead of '.*'
    for suffix substitution. '.*' is now deprecated, but will still be
    supported. However, it will be removed at some point in a future release.
  - Die when stdout ('-') is given as output argument to --image|-g.
  - '<body>' tags are no longer included in HTML/XHTML html dump.
  - Added additional test environment variables to force enable/disable tests
    for optional ebook formats.
    - TEST_CBR, TEST_CB7, TEST_CHM
  - Fix typos in documentation.

1.03  Mar 21 2025
  - Add support for the KF8/AZW3 ebook format.
  - Fix raw text dumping in Mobis.
  - Fix Mobi HTML cleanup.
  - ishmael verifies Mobi image records actually contain image data.
  - Fix PalmDoc decoding errors.

1.02  Mar 14 2025
  - Do not try to unzip CBZ archives to symlinked temporary directories.
    Primarily an issue with Darwin platforms.
  - Ishmael-formatted metadata (--metadata=ishmael) is now dumped in UTF8.
  - Recognize DC metadata in X?HTML files.
  - Recognize "generator" metadata in X?HTML files as software metadata.
  - Removed dump tests.
  - Improved Mobi encoding handling.
  - ishmael dies if comic book archive does not have any images.
  - Fixed unescaped left brace error in queequeg regex.

1.01  Mar 11 2025
  - Force text browsers to dump output in UTF8.
  - Fix dumping X?HTML files with no title.
  - Add -e|--html-encoding option to queequeg: Force the use of a specified
    encoding when reading HTML input.
  - Remove unnecessary module loads.
  - Use "remove_tree" instead of "rmtree".
  - Fixed previous changelog date.

1.00  Mar 04 2025
  This is a major upgrade that introduces many changes that may be incompatible
  with previous versions of ishmael. Pay special attention to the Added,
  Changes, and Removed sections of this changelog if this concerns you.
  Added:
    - Added the -g/--image option: Dump all images present in a given ebook to a
      specified directory.
    - Added support for additional ebook formats.
      - Microsoft Compiled HTML (CHM)
      - Comic Book archives (cb7, cbr, cbz)
      - These new formats have introduced some new optional dependencies; chmlib
        for CHM, 7z for cb7, and unrar for cbr.
    - New metadata dump formats: pretty xml, normal xml, and normal JSON. See
      the changes section for information on how the new metadata dumping system
      works.
    - Added EXAMPLES section to manual.
  Changes:
    - Rebranding as a general ebook dumper as opposed to just a plain text
      converter.
    - The -m/--metadata option can now dump multiple different formats of
      metadata; ishmael (the original), json, pjson (the original --meta-json),
      xml, and pxml. These formats are specified by an optional argument to
      --metadata (--metadata=<form>). This has removed the need for --meta-json
      to be a separate option.
    - Changed default -c/--cover behavior. Output is now written to a file
      rather than stdout by default ("pick the right default"). You can also add
      a '.*' (dot asterisk) to the end of the output path name which ishmael
      will substitute for the image's format suffix.
    - The -c/--cover option no longer "dies" if a cover image is not present
      in an ebook.
  Removed:
    - Removed the -o/--output option. The new way to specify output is via a
      second command-line argument following the given ebook file.
    - Removed the -j/--meta-json option. Please use --metadata=pjson instead.
  Fixes:
    - When executing system commands via qx, ishmael now quotes arguments using
      single quotes instead of double quotes. This should mean that arguments
      with shell metacharacters should not cause unwanted behavior.
    - ishmael no longer relies on an EPUB's metadata file to specify the 'dc'
      namespace, which should fix reading some unconventionally formatted
      EPUBs.
    - ishmael now converts CP1252-encoded Mobis to UTF-8.
    - Unix time handling has been fixed for PDB-based formats (Mobi, AZW,
      PalmDoc, zTXT).
    - ishmael no longer recognizes unset creation/modification dates in
      PDB-based formats.
    - Fixed HTML/XHTML identification heuristics.
    - Fix documentation typos.
    - Fix test typos.
  Improvements:
    - Format identification heuristics have been optimized.

0.07  Feb 25 2025
  - Added -r/--raw option: Dumps the raw, unformatted text contents of a given
    ebook.
  - Added -c/--cover option: Dump the cover image of a given ebook if one is
    present.
    - As a result, pdftopng is an additional dependency if one wishes to dump
      PDF covers. pdftopng should be included with most versions of
      poppler-utils.
    - MIME::Base64 was also added as a dependency, although it should be
      included with Perl core.
  - XHTML is now considered a seperate format from HTML (although its class is
    derived from the HTML class, so it should act mostly the same except for
    being called XHTML rather than HTML).
  - Recognize some more FictionBook2 metadata.
  - Improve some format heuristics.
    - FictionBook2, HTML, XHTML
  - When reading EPUBs, try not to dump items that are not under the
    "application/xhtml+xml" media type.
  - Moved PDB modules out of EBook namespace.
  - Removed EBook::Ishmael::EBook::Skeleton.

0.06  Feb 22 2025
  - When ran with no arguments, queequeg reads input from stdin.
  - queequeg now accepts multiple files as argument.
  - --output option now works with --metadata and --meta-json options.
  - Removed extra newline from json output.
  - No longer bothering categorizing Changes entries.
  - AZW support.
  - Add support for Huff/CDIC MOBI compression scheme.

0.05  Feb 17 2025
  Added:
    - queequeg: ishmael's own HTML dumper script that acts as a fallback
      formatter if none are available.
  Changes:
    - Specify XML::LibXML version 1.70 as minimum required version.
  Improvements:
    - If temporary directory and working directory are not ok directories to
      unzip epub to, try HOME directory.
    - When reading HTML documents, ishmael attempts to recover malformed HTML.
  Fixes:
    - Check if working directory is writable instead of readable when creating
      temporary epub extract directory.

0.04  Feb 14 2025
  Added:
    - Support for TEST_PDF environment variable in tests. This allows you to
      force enable/disable tests for PDFs regardless of whether poppler utils
      are installed or not.
  Changes:
    - Introduced Cwd dependency.
    - Specified File::Temp version 0.10 as minimum required version.
  Improvements:
    - More accurate text heuristic.
  Fixes:
    - Try to unzip epub to working directory if temporary directory is
      symlinked, as Archive::Zip does not support unzipping to symlinked
      directories. This was primarily an issue with Darwin platforms.
    - Corrected some documentation errors.

0.03  Feb 11 2025
  Changes:
    - ishmael now uses a format-agnostic interface for handling ebook metadata.
      This means that although the metadata ishmael gathers will now be less
      detailed, it should at least be consistent across different formats.
  Fixes:
    - Modify PDF tests to account for differing output between different
      versions of pdfinfo.
    - Use '-T text/html' option for w3m, so that it recognizes given data as
      HTML and actually tries to format it.

0.02  Feb 10 2025
  Added:
    - --format can now recognize some alternative format names.
      - fb2 for FictionBook2
      - xhtml for HTML
  Changes:
    - Removed dumper list from help message.
  Fixes:
    - Fix the syntax of the width check in browser_dump() to make it compatible
      with older Perl versions.
    - Add missing developer documentation.

0.01  Feb 09 2025
  - Initial release.