RC1

Released 9 May 2011

• New metadata support
  • Added support for XMP metadata
  • Added support for Photoshop metadata embedded in Jpeg files
  • Added support for ICC colour profiles
  • Added support for JFIF metadata
  • Added many more tags to IPTC directory
• Makernotes
  • Improved support for Panasonic makernotes, including decoding of detected and recognised faces in images
  • Wider support for Canon makernote data
  • Renamed existing SonyMakernote classes to SonyType1Makernote
  • Added tags to the previously empty SonyType1MakernoteDirectory
  • Added support for Sony Type 6 makernotes
  • Added support for Sigma/Foveon camera makernotes
  • Added support for many more Panasonic makernote tags
  • Considerable additions to Nikon Type 2 makernote handling (covering modern Nikon cameras)
• Reliability
  • Converted all access to byte[] in ExifReader to be via BufferReader, ensuring proper bounds checking during processing
  • JpegMetadataReader no longer assumes that certain segments always hold one type of metadata, cycling through the segments available and opportunistically detecting segments that can be processed
  • Now stopping IFD processing when an invalid format type is encountered, helping to avoid populating directories with lots of rubbish from corrupt files
  • Can now decode version numbers of form [0 1 0 0] as well as [48 49 48 48]
  • Introduced new checked BufferBoundsException which is thrown by all BufferReader methods, ensuring that calling code properly handles cases when we run out of data
  • Applied @Nullable and @NotNull attributes across entire codebase
• Unit Tests
  • Significantly improved IPTC unit testing.
  • Added unit tests for Sony Type 1 and Type 6 makernotes
  • Updated all unit tests to jUnit 4
  • AllTests class no longer needed and removed
  • Effort to make unit tests work in non-English locales
• Bugs
  • Fixed a potential bug in ExifReader around handling of float32/64 values. This may never have been seen as it doesn't seem that these types are ever used in Exif data
  • Fixed a potential exception in Directory.getString(int) when the underlying value is an array of primitives other than int
  • Fixed typo in GpsDirectory which incorrectly confused the TAG_GPS_IMG_DIRECTION_REF and TAG_GPS_IMG_DIRECTION tag names
  • Corrected integer value assigned to tag TAG_GPS_AREA_INFORMATION
  • Added proper decoding of GpsDirectory.TAG_GPS_VERSION_ID
  • Introduced Yuri's fix for the various SOF0-SOFF Jpeg segments
  • Improved Exif directory 'user comment' decoding for various character sets.
  • Avoid OutOfMemoryError if the advertised size of the image thumbnail is larger than the available bytes in the Exif segment
• Public API changes
  • Metadata and Directory are now both iterable, so they support foreach now
  • Removed checked MetadataException from all description getters
  • Refactored all Reader classes to take their data in the extract method, rather than in their constructors, offering thread safety and allowing reuse
  • Created method Metadata.hasErrors() that reports whether there exists an error in any directory
  • Separated Exif IFD1 (thumbnail) tags out from the IFD0 tags and into the new ExifThumbnailDirectory with ExifThumbnailDescriptor
  • Split ExifDescriptor into ExifIFD0Descriptor and ExifSubIFDDescriptor to mirror actual IFD structure in physical metadata, preventing against tag overwrites due to merging, and making later support for writing metadata easier
  • Added an overload of Directory.getString that accepts a charset
  • Use of generics
    • Replaced several instances of ArrayList and HashMap with their generic counterparts
    • Genericised methods on Metadata for obtaining Directory objects, removing several casts
    • Made TagDescriptor generic
    • Added generic constraint upon Metadata.getDirectory and Metadata.containsDirectory methods and removed cast and check that are subsequently unnecessary
    • Made descriptor constructors take more specific directory types
  • Added overload of Directory.getDate that accepts a TimeZone to be used if the underlying value is a string and must be parsed
  • Renamed ExifDirectory.TAG_COMPRESSION to ExifDirectory.TAG_THUMBNAIL_COMPRESSION
  • Removed some incomplete and unused constants on ExifDirectory
  • Removed dependency upon com.sun.image.codec.jpeg in JpegMetadataReader
  • Created hashCode method for Rational as it overrides equals
  • Removed old deprecated constructors for reader classes
• Internal API changes
  • Rational now backed by long, not int
  • Considerable simplification of interface MetadataReader and all its implementations
  • Overhauled means of reading different types from byte[]}} across allReaderclasses via use of newBufferReader` class
  • Created methods to add arrays of float/double to a Directory
  • Renamed Metadata.getDirectory as getOrCreateDirectory, and made a new method getDirectory that doesn't lazily create a directory
  • Metadata readers only accept non-null byte[] as a data source
  • Added overload of BufferReader.getString that takes a charset
  • Simplified ExifReader by removing some fields so that state is now mostly maintained on the stack
  • JpegSegmentReader.readSegment() no longer throws JpegProcessingException, simplifying some call sites too
• Findbugs support
  • Addressed a few more minor Findbugs warnings
  • Disabled Findbugs inspections on XMPCore module
• Serialisation
  • Directory and Descriptor are no longer serialisable
  • Set serialVersionUID on all serialisable exception types, and Rational
• Miscellaneous
  • Added Javadoc to SVN for viewing via Google Code
  • Updated source code headers with Apache 2 License banner
  • Excluded Output folder from IntelliJ project
  • ImageMetadataReader.main now has an option to output Wiki format tables to more easily show sample output information on the project's wiki
  • ImageMetadataReader.main now accept multiple file paths
  • Several spelling mistakes corrected
  • Removed commented code
  • Removed compiler warnings
  • Added some more sample images

RC2

Released 27 May 2011

• Created new Age class for more meaningful representation of recognised people's ages in Panasonic camera photos.
• Created descriptor methods for additional Panasonic makernote tags (for baby ages and text stamps)
• Fixed a bug where PanasonicMakernoteDescriptor.getTitleDescription was returning the text associated with TAG_LANDMARK
• Created PanasonicMakernoteDirectory.getAge method for raw access to the Age object associated with certain tags
• Added missing hashCode() method to Face class
• Added missing getters to Face class
• Fixed logic that checks for the end of a printable string in PanasonicMakernoteDescriptor
• Fixed annotation issue
• Made the XMPMeta object instance available on XmpDirectory for general consumption
• Added support for AdobeJpegDirectory and related classes, which comes from the APPE JPEG segment, when present
• Additional API documentation
• Dealt with some incorrect code analysis warnings

RC 3 (released)

Released 18 Nov 2011

• Fixed a bug where adding an offset and length resulting in integer overflow was incorrectly passing a bounds check
• Added unit test for IccReader
• Added exception handling for parsing of numeric types from strings in XmpReader
• Fixed bug in parsing of JFIF data
• Added JFIF parsing unit test
• Removed ISO description check that multiplied ISO values less than 50 by 200
• Created utility class ExtractAppSegmentBytesToFileUtility for writing specific JPG segments to files, and reading them again into byte[] for convenience in unit testing (and keeping source code file size down)