Metacat

added support for the new abstract class) to handle data file uploading through metnewninterface (actuall

newninterface (actually an abstract class) to handle data file uploading through met

fixed error in logic. sending the filesize as an int terminated by a zero failed when the file size was greater than 256. The filesize is now sent as a string followed by a zero.

added support for the server to return the docid of the newly added data file. Also, now the data stream must be prefixed with the following information: filename <0> sessionid <0> filesize <0>. where the <0> symbol is a byte containing zero. The filesize is sent as an int and the filename and sessionid are sent as strings (in byte form).

unified getting "docid" value from metacat URLs specified in <resourceIdentifier> tags
by using MetaCatUtil.parseQuery() method in the same way as in MetaCatServlet.

fixed error handling and time out issues. Also fixed thread syncronization problem

multiple files with the same names no longer overwrite each other.

fixed minor error

added support for data file upload via a random, authenticated socket.
The action added to metacatServlet is getdataport. A client can send a getdataport request. the server returns a random port number and starts a thread to handle the request on the given random port. The thread will only accept requests from the client that has the session id equal to that of the user that made the getdataport request. the user must be authenticated and logged in to make a getdataport request. Once the port number is received by the client the connection can be made. The first two parameters in the data stream must be the filename, followed by session id. (note that it only wants the session id not the whole cookie and the "JSESSION=" must be stripped from the text.) If the session number is correct, the upload is allowed, the file is written and xml_documents is updated with the new data file information.

added support for data file upload via a random, authenticated socket.

included "accessfileid" attr in the unique constraint

- appling acl through access files only
- clearing around update/delete of access files - to delete the related records from both xml_access and xml_relation tables
- included check for user permissions for setting acl on the resources in the access file
- moved RelationHandler.deleteRelations(docid) down in RelationHandler obj just before the write of the new relations

new attr "accessfileid" in xml_access table

fixed url error in data download function

Use the new property "accessdoctype" to check if this is access file.
If it is, run AccessControlList for wirting into xml_access and xml_relation tables.
This is the thread that writes XML Index but afterwards.

included support for submition of access files
+ storing the acl info in xml_access table and relationship records in xml_relation table like
<aclfile, "isaclfilefor", recourceIdentifier> for every resource included in the access file

added support to download a data file through metacat using a metacat url or a docid parameter.

updating a package file now updates the relations that are created from it. Also, if you delete a package file, it will delete the relations associated with it.

added docid field to xml_relation to track which package document creates the relation.

fixed package location

changed replication interface to make it easier to see the servers.

added support for one way replication

added revision support to the delete function

added resource to the dtd script

fixed sql error

cleaned up code

removed package doctype filtering from the resultset.

added comments to top of class

changed accession number schema to <sitecode>.<serialnumber>.<revisionnumber>. DocumentIdentifier.java parses docids into parts for easy manipulation. You can now specifiy a specific revision by asking for document <sitecode>.<serialnumber>.<revisionnumber> or you can get the newest revision by asking for <sitecode>.<serialnumber> or by asking for <sitecode>.<serialnumber>.newest.

fixed typos around acltext and dtdtext params

changed acc# generation to use db sequence instead of timestamp

fixed typo

added db sequence accnum_uniqueid_seq needed for uniqueID part of Accession#
get rid of xml_acc_numbers table

fixed typo

fixed typo

fixed typo

commented out print statement

fixed typo

problems with passing the connection between objects

problems with passing the connection between objects

overwrite Jivka's overwrite

problem with passing the connection between objects

problem with passing the connection between objects

removed print statement

fixed bug where some relations were not returned because the sql did not take into account an sql that had a directory structure in it.

changed connection handling in the constructor

one more write() method was needed

constructor DBEntityResolver(Connection conn) is needed; cleared by mistake

DBEntityResolver:
- added new parameter dtd for upload on Metacat file system; optional; Reader
- new routines to upload dtd if provided
- systemID of metadata document is used to exctract the filename of new dtd;
- the new dtd is uploaded on Metacat file system as specified by dtdPath property;...

enable the UNIQUE constraint in xml_catalog table

added additional actions to facilitate a web based interface to the replication servlet.

removed the initial connection pool creation step in init()

comment out line in read action that sends a copy of docid prior to returning doc
in XML format (apparent debug statement left in by mistake)

added replication for the xml_catalog table. Right now it is only replicated when the deltaT handler is called. I will add more replication times later.

parser for the xml encoding of the xml_catalog tableing (fi

added a logging system for replication. The log file location is set in metacat.properties

cleaned up code, added more complete documentation of replication algorithms and datastructures.

made replication on insert or update us. Also made a method in AccessionNumber public so that you can tell if an accession number has already been used.e place be

An asynchronous replication handler. This class creates a new thread to handle replication so that the user does not have to wait for replication to take place before he/she sees the result of an insert or update.

updated file locking mechanisms to use revisions instead of dates.

fixed typo, removed extra print statements

a script for clearing the contents of a metadat DB without destroying the table structures and without erasing the DTD and XSL entries in xml_catalog.

added replication_on_insert handling. Changed replication from date_updated base replication to revision number replication.

change the generation of Accession# in the form of <sidecode>.<createdate>

added revision number tracking support to xml_documents and xml_revisions. Also added an updated flag as a replacement for date based replication handling.

replication on insert functionality is now working.

added more replication/file locking functionality.

added more locking support and support for remote update of files.

added more locking functionality for replication

AccessControlList
- methods for parsing and loading acl file
- checkup method for permission for given principal on given resource
DBQuery
- checkup for READ permission using AccessControlList.hasPermission()
DocumentImpl
- using AccessControlList object to parse and load an acl file into xml_access table...

cleaned up code, added new static methods to MetacatReplication for handling static queries and getting URL content.

updated insert handling and added an action to request the time from a remote server.

created locking action in replication servlet and added semi-support to documentImpl to handle documents updatedee by replicatio

Modified stylesheets and server code to use a single "read" action instead
of the earlier "getdocument" and "getrelateddocument" actions. In the
process, developed and started to utilize a new suite of URL "protocol
handlers" that are in the package "edu.ucsb.nceas.protocols" and handle...

Fixed typo in create table syntax (added missing comma) for xml_access table.

Added a new URL protocol handler class for the metacat protocol. This
class will replace the current MetacatURL class in a much simpler and
extensible manner. See the documentation for java.net.URLConnection and
java.net.URL for details.

this class allows the parsing of xml data sent by a replication server to assertain document info about a remote document.

added functionality to allow the replication servlet to assertain and insert user and group info into the local database. started implementation of insert replication handler (it is commented out in this commit)

CREATE TABLE xml_access (
docid VARCHAR2, -- the document id #
principal_name VARCHAR2, -- name of user, group, etc.
permission NUMBER, -- "read", "write", "all"
perm_type VARCHAR2, -- "allowed" or "denied"
perm_order VARCHAR2 -- "allow first" or "deny first"...

fixed compatibility problem by overloading the write method. Jivka's new parameter (Reader acl) is now defaulted to null if it is not explicitly declared.

added interface to handle a new "acl" parameter bringing the text of a xml access file for a given metadata document
the "acl" parameter is optional and should be sent on INSERT or UPDATE action.

added ticket_counter and deny attributes to xml_access table

new class for parsing ACL XML file and loading acl data into metacat db

added delete functionality

added support for server_location in the xml_revisions table.

removed inadvertent comments

added needed table fields for storing document location in xml_documents

removed dependence on a server code in the accession number of each document. the file's resident server is now located in xml_documents.server_location.

no message

added new server_location field to xml_documents with a foreign key to xml_replication.serverid

now updates the xml_replication last_checked field

updated javadoc documentation

fixed bug/typo

no message

no message

changed action scheme

Modified pathquery.dtd and QuerySpecification to allow some new constraints
on queries submitted to Metacat. Now, queries can contain an optional
set of <owner> tags that identify the users for which documents should be
returned, and a list of <site> tags constraining the site identifiers...

can now send a file to another server

can now download a document from a server

no message