Commits on Sep 9, 2012
  1. Fixes #87 - Maven artifact now is jdom2

    Welcome to the onslaught of fire&brimstone ....
    rolfl committed Sep 9, 2012
  2. Issue #91 - AttributeList double-check of duplicate fix....

    The fix for this check results in some checks being done in a few places, and not all of these places were covered in the test-harness, resulting in uncovered code. This commit ensures that all places that check attributes now have appropriate test cases.
    rolfl committed Sep 9, 2012
  3. Fix copy/paste/typo in message

    rolfl committed Sep 9, 2012
  4. Issue #92 - Fix off-by-one bug introduced in character verification.

    The code will now verify the second character in a name.
    Additionally, the error messages for broken names has been improved to indicate what both the broken name and character are, instead of just the character.
    Also, change the checkNamespacePrefix code to leverage the common code from checkElementName and checkAttributeName.
    Also improve the test harness to cover test cases that are now more complicated because the name-check code is in two places now instead of one....
    rolfl committed Sep 9, 2012
  5. Remove ill-conceived +2 logic for ensureCapacity which does not in fa…

    …ct ensure even byte allocations.... what was I thinking.
    rolfl committed Sep 9, 2012
Commits on Sep 3, 2012
  1. Fixes #92 - Performance improvements in Verifier.

    Attribution for Wilf's contributions. Close off the issue.
    rolfl committed Sep 3, 2012
  2. Issue #92 - Update the performance benchmark tool to:

    - delay for 5 seconds to allow a  profiler tool to be attached.
    - run 200 iterations, of which the best 50 are reported.
    - report the average time for each loop rather than the sum time.
    rolfl committed Sep 3, 2012
  3. Issue #92 - Improve performance of checkCharacterData ...

    restructured loop is much faster (runs in half the time).
    Add check in to VerifierBuilder to confirm logic about CharacterData being a superset of other character roles.
    rolfl committed Sep 3, 2012
  4. Issue #92 - Verifier performance

    Use byte-based arithmatic everywhere.
    rolfl committed Sep 3, 2012
  5. Issue #92 - Verifie performance.

    Change the bitmask to not contain ':' as part of the valid name characters.
    Also improve the performance test to average the best times.
    Centralize the JDOM Name checking to  anew method.
    rolfl committed Sep 3, 2012
Commits on Sep 2, 2012
  1. Issue #92 - Performance of Verifier.

    Initial commit with Byte-per-Character system for bitmask-based lookup tables.
    See details at
    rolfl committed Sep 2, 2012
  2. Fixes #90 - false-positive error validating namespace collisions when…

    … 'setting' an Attribute.
    rolfl committed Sep 2, 2012
Commits on Sep 1, 2012
Commits on Aug 26, 2012
  1. Fixes #89 - Add specific JAXP-factory mechanisms to JAXP-based pillar…

    …s for SAX parsing.
    Also update the sax-based parsing descriptions/documentation.
    rolfl committed Aug 26, 2012
Commits on Aug 11, 2012
  1. Fixes #88 - broken serialization of subclasses of core JDOM classes (…

    …like custom subclasses of Element).
    Upgrade the serialiazion test-code to report exceptions better.
    Add a test-case that checks that all subclasses are serializable.
    rolfl committed Aug 11, 2012
Commits on Jun 9, 2012
  1. Revert change for jdom jar name

    rolfl committed Jun 9, 2012
  2. Fixes #84 - Message for IllegalArgumentException in re-declare of def…

    …ault namespace for XPath expression.
    rolfl committed Jun 9, 2012
Commits on May 28, 2012
  1. Accommodate new maven artifactID... change name of jars from jdom-x.y…

    ….z to jdom2-x.y.z
    with different maven artifact id's for JDOM 1.x and 2.x it is possible to have both jars in the same maven project.
    rolfl committed May 28, 2012
Commits on May 12, 2012
  1. Fixes #82 - multiple active FilterLists need to resynchronize on remo…

    This fix changes the tracking of dataModCount so that it is updated on remove(). Additionally, it centralizes the processing to b consistent with the size-mod-count (and data-mod-count).
    rolfl committed May 12, 2012
Commits on Apr 23, 2012
  1. Fixes #75 - Android Compatibility.

    Implements some workarounds for AssertionError.
    adds a new layer in the XMLReader*Factory that makes it much easier to implement a new custom Reader factory.
    rolfl committed Apr 23, 2012
Commits on Apr 22, 2012
  1. Issue #75 - change Resource paths to be absolute.

    change the android tests to use simpler process (no longer needs to use android Assets), instead us this.getClass().getResource(...). The same mechanism now works in Java an Android.
    rolfl committed Apr 22, 2012
Commits on Apr 20, 2012
Commits on Apr 19, 2012
  1. Issue #75 - Create URLs at runtime, not class construct time. This fi…

    …xes a broken context in Android where the base directory is not populated correctly.
    rolfl committed Apr 19, 2012
  2. Issue #75 - Android's e.printStackTrace() is not useful because there…

    … is no accessible STDERR (or STDOUT). Instead 'chain' the problem.
    rolfl committed Apr 19, 2012
  3. Issue #75 - Replace JUnit 'Assume' directives with plain code to hand…

    …le these conditions in Android test layer
    rolfl committed Apr 19, 2012
  4. Issue #75 - Set the android File resource path to absolute which make…

    …s more URLs resolve
    rolfl committed Apr 19, 2012
  5. Fixes #73 - DOM getElementsByTagName() was returning 'self' as well.

    This was broken in getElementsByTagNameNS() too.
    This is resolved. Also the '*' concept has been added to getElementsByTagName().
    This must have been 'over-thought' when I put it together. Not sure why the code was going out of the way to add in the 'self' element.
    rolfl committed Apr 19, 2012
  6. Issue #75 - Android

    Update the build process to include the 'android' target which loads an Android main and test project with the JDOM unit tests. See
    rolfl committed Apr 19, 2012
  7. Issue #75 - Android.

    Use new FidoFetch mechanism for getting test resources. This 'abstracts' the process so that it can be different between regular Java testing and Android testing.
    rolfl committed Apr 19, 2012
  8. Issue #75 - Android support.

    The class-build order appears to be subtly different for Android, which, in this enum, causes the items to become available at a different time (compated to regular Java). By restructring the 'DEFAULT' enum build process we end up with a better solution for both regular Java and Android.
    rolfl committed Apr 19, 2012
  9. Issue #75 - Android support

    Android does not implement the (mandatory) XMLSchema (XSD) specification. This is 'broken', and it throws IllegalArgumentException.
    This fix catches the IllegalArgumentException which means that the XMLReaders enum is able to initialize. (without this fix, the entire enum crashes, and any external references throw ClassNotFoundErrors in Android).
    rolfl committed Apr 19, 2012
Commits on Apr 11, 2012