Daffodil 2.0.0 - detailed changelog

Core

• support for unparse, with one-to-one feature parity with parse
• support for unparse specific properties like dfdl:outputValueCalc, dfdl:filleByte, dfdl:truncateSpecifiedLengthString
• now requires Java 8 and Scala 2.11
• DFDL expression changes:
  • Support added for dfdl:contentLength(), dfdl:valueLength(), fn:error(), and math:pow() functions
  • fix bugs with equality comparisons and converting differing types
  • fix bugs with div/idiv not using the appropriate kindof division
  • fix bugs with daf:trace on complex types
• creation of a new data input/output stream layer
• add support for 3-, 4-, 5-, 6-, and 7-bit ASCII encodings
• add property daf:parseUnparsePolicy and tunable to enable only parsing, unparsing, or both
• add tunable unqualifiedPathStepProperty to specify how unqualified DFDL expression path steps are resolved
• allow dfdl:lengthKind="pattern" when type="xs:hexBinary"
• support for xs:choice direct dispatch
• support for xs:boolean type
• support for xs:hexBinary type with bit lengths
• support for xs:ignoreCase="yes"
• support for xs:union restrictions
• support for runtime valued properties
• improved diagnostic messages
• numerous bug fixes and performance improvements

API

• new parse and unparse API with InfosetInputter's and InfosetOutputter's
• native infoset support for JDOM, w3c DOM, Scala XML Nodes, XML Text, and json

TDML

• created an XSL stylesheet for creating DFDL tutorials from TDML files

Daffodil 1.1.0 - detailed changelog

Core

• initial support for unparsing, including support for unparsing
  • initiators, terminators, and separators
  • escape characters
  • padding/fillByte
  • delimited, fixed length, and pattern length strings
  • arrays with occursCountKind 'implicit', 'parsed', and 'fixed'
  • nillable simple and complex elements
  • sequences and choices
  • variables
• support for stream infosets via XMLEventReader
• no longer pretty print large infoset to decrease memory and time to print. Set the new prettyPrintElementLimit tunable to change the size at which infosets will not be pretty printed
• improve support including/importing absolute Windows file paths
• improve parser performance with calendar types
• fixed bug with delimiters and escape characters in which scoping could be incorrect, resulting in the use of the wrong delimiter/escape char
• improved OOLAG to make debugging easier
• support for save/reload of unparsers

CLI

• variables set on the command line using -D when saving a parser are saved, and can be overridden when reloading the parser
• add --unparse option to performance subcommand to run unparser performance tests

TDML

• enhancements to support unparser tests and round tripping parser tests
• no longer trim resulting infoset, requiring expected infoset to be more exact, but give less false positives
• detect and warn about duplicate test names

Daffodil 1.0.0 - detailed changelog

Core

• schemas inside jars on the classpath can now be linked via include/import statements
• updates to DFDL expression parser to fix numerous bugs and add features
  • add date, time, constructors
  • add timeZoneFromDateTime, Time, and Date functions
  • add comparison operators for booleans
  • improve error messages
• support empty strings in long-form properties
• throw SDE when testPattern is empty but testKind="pattern"
• throw SDE when inputValueCalc is used on optional or non-scalar elements 
• improve numerous diagnostic error messages
• improve performance when writing hexBinary to the infoset
• support %ES; in delimiters to make a delimiter optional
• fix non-thread-safe code related to text numbers
• fix numerous typos and grammatical errors
• fix delimiter scanning with %WSP*;
• add tests to validate Java 8 support
• check for disallows regular expression constructs
• add ability to specify a temporary directory, currently only useful as part of TDML tests
• improve performance when processing errors are heavily used for backtracking

CLI

• allow the --schema option to use schemas inside jar files on the classpath
• only allow the --schema option to be provided once; all other schemas must come from import/include statements
• prevent the eval command from aborting if a DFDL expression failed to compile
• re-enable support for breakpoints
• add validation and auto-completion to debugger display command

API

• add new API functions to compile schema's in either File or URI form
• only allow a single schema to be provided to the API; all other schemas must come from import/include statements
• add scala API and scaladoc

TDML

...

Daffodil 0.15.0 - detailed changelog

Infrastructure

• license information for dependencies is now distributed in the released files
• remove test dependencies from released files
• use a plugin to create release files rather than our own custom packager

Core

• replace Saxon with custom written DFDL Expression compiler/evaluator. This causes large performance improvements when parsing files that rely on DFDL expressions.
• temporarily remove support for unordered sequences
• remove support for query-style paths in DFDL expressions
• add dfdl namespace XPath constructor functions
• add support for the ability to serialize and deserialize parsers

CLI

• support the -P/--parser option to use a saved parser in the parse and performance subcommands
• support the save-parser subcommand

API

...

Daffodil 0.14.0 - detailed changelog

Core

• support for bitOrder property
• support for encodingErrorPolicy="replace"
• fix bug where schemas using import/include were not properly validated, causing various error messages
• fix bug where lengthKind="explicit" in some cases incorrectly included initiators and terminators in the content length
• various performance improvements, with some parses having increased speed of ~30%
• improve error messages when invalid regular expressions are used in properties like testPattern and lengthPattern
• fix error with when runtime byteOrder expressions fail
• fix potential thread safety bugs in delimiter scanning

CLI

...

Daffodil 0.13.0 - detailed changelog

Dependencies

• upgrade to scala-2.10.4
• upgrade to sbt-0.13.2

Core

• fix non-thread-safe code in date/time/dateTime types, text numbers, delimiters, occursCountKind='expression', and expression evaluation
• use correct namespace for external variables
• improve performance for data containing only ASCII encoding
• remove broken support for lengthUnits=bytes with specified lengthKind and variable-width encodings
• replace much of the parser combinator code with DFAs specifically tailored towards DFDL requirements, giving a performance increase on delimited data

CLI

...

Daffodil 0.12.0 - detailed changelog

Dependencies

• upgrade to Saxon-HE-9.5.1-1
• upgrade to jdom2-2.0.5
• upgrade to scallop-0.9.4

Core

• check for unescaped curly braces in DFDL literals
• ignore dfdl:representation when type is xs:hexBinary or xs:string. Representation is assumed to be binary with xs:hexBinary, and text with xs:string
• remove defaulting of dfdl:sequenceKind to ordered. The property must now be specified
• fix bug causing the use of the wrong root element when not provided via the API or CLI
• report schema definition error when no terminator is supplied, but lengthKind='delimited' and tralingSkip > 0
• add check for alignment ambiguity
• require xs:string and xs:hexBinary to defined minOccurs/maxOccurs when lengthKind='implicit'
• when a delimiter is not found, display what was found in the error message
• check for distinctness of delimiters
• disallow whitespace in some properties. Instead the DFDL character entites such as %SP; must be used instead
• add support for DFDL expressions in escapeCharacter and escapeEscapeCharacter 
• fix incorrect parse failures when escapeEscapeCharacter was set to the empty string

API

• add new function to the Compiler setValidateDFDLSchemas to enable/disable DFDL schema validation, such as unique particle attribute and various facet related checks

CLI

...

Daffodil 0.11.0 - detailed changelog

Core

• add support for runtime expressions in byteOrder
• add support for sequenceKind="unordered"
• add support for infoset validation,  three modes
  1. "on": check constraints and Xerces validation
  2. "limited": check constraints, no Xerces validation
  3. "off"" no validation performed
• add support for external variables
• add support for advanced text number properties: 
  
  • textStandardDecimalSeparator (static & dynamic)
  • textStandardGroupingDeparator (static & dynamic)
  • textStandardExponentRep (static & dynamic) (renamed from textStandardExponentCharacter)
  • textStandardInfinityRep 
  • textStandardNaNRep 
  • textStandardZeroRep 
  • textNumberPattern 
  • textNumberCheckPolicy 
  • textNumberRounding 
  • textNumberRoundingMode 
  • textNumberRoundingIncrement 
• add new DFDL expression compiler to better detect invalid expressions, primarily  those related to unordered sequences
• add min/maxOccurs validation check
• require appinfo source attribute to be "http://www.ogf.org/dfdl/"
• add support for binary and text xs:decimal type
• improve diagnostics when a variable is used but not set  
• fix bug with dynamic separators using variables
• fix SDE detection on last required element when maxOccurs="unbounded" and {{ occursCountKind="implicit"}} 
• fix lengthKind='pattern' to not error when a match is not found, but instead  to return a length of zero
• fix bug where nonNegativeInteger range was not checked correctly
• fix bug iso-8859-1 support

Java API

• fix bug in JAPI default log writer causing logs to fail to write 
  

CLI

• new options to enable external variables
  • -D, --variables to define one or more external variables
  • -c, --config to define an XML file containing external variables
• new option for validation
  • --validation-mode=<mode>, where mode can be "on", "off", or "limited" 
    
• new option for the test subcommand
  • -i, --info to display more information about what a test failed. May be supplied multiple times to increase the amount of information

TDML

• ability to enable validation via a new attribute, validation="<mode>", placed on a parserTestCase 

API

...

Daffodil 0.10.1 - detailed changelog

Core/ Java API

• fix bug causing duplicate elements to be added to the infoset
• fix bug where delimiters containing DFDL character entities but not starting with a DFDL character entity caused a schema definition error

Debugger

...

Daffodil 0.10.0 - detailed changelog

Core

• xs:include and xs:import now work with relative file schemaLocations
• respect minOccurs when occursCountKind="implicit"
• diagnostics improvements
• fix 7-bit USASCII with implicit alignment
• add schema definition warning if xs:appinfo source is incorrect
• fix unsignedXXX values
• improve schema definition error detection
• eliminate duplicate errors
• require that a complexType must have exactly one child, which must be a sequence, choice, or group reference
• allow spaces in pathnames to input files and schemas
• calculated values (e.g. inputValueCalc) maintain their precision in error messages
• check for invalid DFDL entities
• support for escaping percent sign

Java API

• fix character entity infoset output when using the Java API
• add support for controlling the debugger, including a trace debugger
• fix issue where multiple retrievals of the parser result caused error

TDML Runner

• can now use CDATA tags in dfdl:infoset for cases where characters like < and > are in the expected output
• replaceDFDLEntities attribute for documentPart element now defaults to false

CLI

• new --trace option to display verbose output of Daffodil parsing
• the --debug option can now take a file containing a list of debugger commands (e.g. --debug=cmds.txt)
• add check for not yet implemented features
• fix bug where order of multiple --schema options was not maintained

Debugger

...

this page has moved to https://daffodil.apache.org/releases/