[GitHub] stevedlawrence opened a new pull request #92: Update dependencies to their latest versions

[GitBox <gi...@apache.org>]

stevedlawrence opened a new pull request #92: Update dependencies to their latest versions
URL: https://github.com/apache/incubator-daffodil/pull/92
 
 
   Updates that required changes:
   
   - ICU4J
     - License has switched to the Unicode license, LICENSE and NOTICE
       updated.
     - No longer requires that a positive number pattern exists. That is
       required by the DFDL specification, so the check is done manually
       (i.e. the pattern cannot start with a semi-colon)
     - ICU now correctly checks roundUnnecsssary and throws an exception
       if rounding is needed on unparsing--before it would just round with
       roundHalfEven. We now catch this exception and return an
       UnparseError. To maintain backwards compatibility, the
       textNumberRoundingMode property in built-in-formats.xsd and
       DFDLGeneralFormat.dfdl.xsd is changed from roundUnnecessary to
       roundHalfEven.
     - ICU's DecimalFormat.parse() method now returns an ICU BigDecimal
       instead of a Java BigInt/BigDecimal.
     - There is a bug in ICU that if a text number pattern uses scientific
       notation but the number is +Inf, -Inf, or NaN, ICU will add the
       exponent (e.g. INFx10^0). This is wrong, so this now manually
       detects if the number is one of those values and outputs the
       appropriate inf/nanRep along with pattern prefix/suffix instead of
       using the DecimalFormat.format method.
     - Lax number parsing no longer trims whitespace, so we now manually
       perform the trim if dfdl:checkPolicy is lax.
     - Various tests are changed or moved out of scala-debug due to ICU bug
       fixes. For example ICU now handles MMM (abbreviated month) for a
       Russian calendar language. Some tests are changed just due to
       different error messages. Some just need to use twoPass round trip.
     - A few tests now fail due to new bugs in ICU. The failing features are:
         1) ICU no longer trims padding characters after a suffix,
            resulting in failures when trying to parse such numbers leading
            to ParseErrors. Padding before prefix, after prefix, and
            before suffix still work as expected.
         2) ICU Fails to parse scientific numbers where the exponent rep
            is the empty string, e.g. 12.34+2 is now invalid.
       Tests (total of 3) that require these features are disabled and
       moved to scala-debug.
   
   - Xerces
     - Some error messages changed to now expand namespace prefixes or
       display slightly different information, requiring updates to some
       tests.
   
   - JLine
     - Deprecated shutdown() in favor of close()
   
   - Jackson
     - Deprecated JsonStringReader.getInstance() replaced with
       BufferRecyclers.getJsonStringEncoder()
   
   DAFFODIL-1973, DAFFODIL-851, DAFFODIL-845

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services