Fedora Commons Code

Microsoft word documents served from Tomcat are getting the MIME-type of application/msword changes to application/vnd.ms-word so that mapping was added. Also the demo pdf was not operating well with the current Adobe reader. This is connected to fcrepo-497 changes.

pdfbox-1.1.0.jar instead of PDFBox-0.7.2.jar
When exceptions are thrown during text extraction, they will be caught and empty index field returned with a log warning.
Before, exceptions were sent back to the indexing stylesheet, which would cancel the indexing,
now the index document will just lack the full text index field.
Text extraction from PDF documents now puts spaces instead of characters 00-31, because they have caused exception during indexing.

Revert changes made in r8644: not all tests are correctly annotated (eg @Test), so this changeset was in fact causing a large number of tests not to be run.

FCREPO-64: Datastream SIZE attribute not fully implemented
- size now computed for "M" when datastreams are modified, SIZE written to FOXML and desearialized when reading
- will need some kind of rebuild tool for pre-existing M datastream versions

Bugfix - lack of RELS-INT causing problems with implementation of FCREPO-703 - fixed.

BeanDefinitions from fcfg data

These are generated files, so removed from subversion.

JUnit cleanup; removed unused "All" test suites. These are no longer
necessary/useful since we have migrated to use maven to invoke the various
sets of tests. By default, the maven test runner will execute all tests
for a given subproject when necessary. Also, when specific subsets
need to be run, these should be specified directly the poms
(e.g. configA/B/Q in fcrepo-integrationtest's pom)

TestRESTAPI wasn't being included in automated ConfigA tests - added this to POM (follow-up from FCREPO-705)

FCREPO-703: REST API calls to getDatastreamDissemination fails when XACML enabled, API-A auth disabled
Changed to use API-A methods instead (listDatastreams, getDatastreamDissemination(RELS-INT))