Seperated web API code, from plain java code, by creating seperate constructors and handling Http stuff in just two methods: "clientRequest" and "handleDownloadResponse".Results in using HttpServlet code in an isolated/thin layer, making it easier to test and reuse all other methods.
Added new functionality for action="Set Access". Includes methods handleChangeAccess, setPublicAccess and getNodeTextStack. Also, cleaned up some code as a result.
Modified string constant "SUB_DOCS_PATH" to be static and public accessible.
Added param to indicate if public has read access or not, in the method "readFromMetacat". It only does this if reading a metadata document, for a skin-format,and the user is not public. This is useful to the skin's metadata display/editing form, for indicating public accessiblity (or inaccessibility)....
Fixed bug where the Permission Order was always being set to 'allowFirst'.It was comparing against the constant AccessControlInterface.DENYFIRST with the variable name surrounded by quotes. Removed the quotes and updated comments.
Modified a sql command from "like" to "=". It dramatically improves the peformance of build index.
Grant the public read access to the data files, if it was done for the metadata file. Also cleaned-up some debug statements.
Find out if the document is a Member of an EML package, and if that's not the case, than skip the test to see if access is set through EML metadata.
use "within" rather than "intersects" as the spatial search criteria for dataset location.(see bugzilla #2972)
Fixed 3 bugs:1) If there is a call to this class without an "action" (i.e. the initial time the login form is loaded) a null pointer exception would be thrown.2) If the user logged out, and then tried to query, the server would send an error (invalidated session)...
View revisions
Also available in: Atom