File Release Notes and Changelog
Release Name: 0.6
Notes: This release includes several bugfixes and improvements. It also makes TapirLink compatible with the latest changes in the TAPIR protocol. Users are strongly encouraged to upgrade.
Changes: * Bugfix: When different configuration interfaces from different TapirLink installations were being accessed through the same browser, session caching was not making any distinction between them. * Users now need to explicitly choose the conceptual schema format when trying to load an additional conceptual schema in the configuration interface (mapping form). * When loading concepts from a CNS configuration file (old text format), if the concept_source alias was not specified then the last concept_source will be loaded. * Bugfix: Resuming the inclusion of new resources was probably not working correctly in previous versions. * Configuration interface can now make use of debug logging. * Bugfix: KVP filters were not working properly on PHP 4.x when they had nested parenthesis. * Many adjustments in the statistics module. * Fixed bug that was affecting KVP requests when parameter names were in uppercase. * Included missing "charset=utf-8" in the HTTP header. * Flush cache after any change in the local mapping. * Resources are not serialized in session anymore when running the web service. * Bugfix: Empty tagname parameters (inventory operation) were not being ignored. * Bugfix: Output models were failing to produce results when the indexing element had a simple type. * Bugfix: Searches were failing when the parameter "limit" was not specified (default was being set to zero instead of null). * Follow up protocol update: Filter condition will now be ignored when parameter is not passed. When a concept is not locally mapped, the error will now only be raised if there's no missing parameter. * Bugfix: KVP filters were not working if they contained any of the following operators: lessthan, lessthanorequals, greaterthan, greaterthanorequals. * Bugfix: Search XML requests referencing external templates were not working. * Updated local version of ADODB to 4.98. * Changed DiGIR configuration import procedure to overcome an XML parsing problem (or maybe it's a feature?) that sometimes splits a string and calls the character data handler for each individual piece of the string. This was happening when DiGIR metadata content contained diacriticals. * Implemented driver to load concepts in the old DarwinCore format (tied to DiGIR). * Created separate library for Xs classes. * Included new settings TP_LOCAL_REPOSITORY (to indicate where local copies of remote files may be found if manually copied there) and TP_FILE_RETRIEVAL_BEHAVIOUR prefer_original, prefer_local or only_local). * Implemented driver to load concepts in the new CNS XML format. * Added support to case sensitive dabatabase schemas. * Bugfix: Now each resource has its separate cache directory. * Bugfix: When a node contained a concatenated mapping, it was being rejected if one of the mappings evaluated to NULL. * Added alias support for output models. * Capabilities response now includes the local datatype for each mapped concept. * Removed ConceptNameServers element from capabilities response. * Bugfix: log-only requests were not working for search and inventory operations. * Bugfix: "next" attribute was not being included in the search summary when no records were returned. * Added xsd:dateTime to db timestamp conversion in filters. * Added db timestamp to xsd:dateTime conversion in search and inventory responses. * Bugfix: Fixed value mappings using numeric type zero were not working.
