Metacat: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362001-07-10T22:55:51ZEcoinformatics Redmine
Redmine Bug #247 (Resolved): metacat inserts ids into some fileshttps://projects.ecoinformatics.org/ecoinfo/issues/2472001-07-10T22:55:51ZChad Berkleyberkley@nceas.ucsb.edu
<p>Metacat sometimes inserts ids into files with the old "meta_file_id" tags. it<br />inserts them incorrectly because it does not include the rev number. This may<br />not be needed now since morpho inserts ids before it inserts the document into<br />metacat. The problem right now is that metacat overwrites the correct id that<br />morpho puts in (an id with a rev number) and writes an incorrect id into the<br />file. This only happens with metadata files that have the old style id tags.</p> Bug #242 (Resolved): delete does not work properly when a file has relationshttps://projects.ecoinformatics.org/ecoinfo/issues/2422001-06-21T18:04:15ZChad Berkleyberkley@nceas.ucsb.edu
<p>Metacat will not delete a file if it has relations in the xml_relations tables<br />because of a constraint error between xml_documents and xml_relations. <br />Relations need to be removed before deleting the document from xml_documents.</p> Bug #240 (Resolved): data file upload does not allow the client to specify an idhttps://projects.ecoinformatics.org/ecoinfo/issues/2402001-06-20T18:29:16ZChad Berkleyberkley@nceas.ucsb.edu
<p>The current file upload utility in metacat needs to be altered so that a client<br />can specify the id of the file being uploaded. We should consider just ignoring<br />the http problem and using the standard PUT method for this. it would make<br />uploading much easier and much more uniform.</p> Bug #235 (Resolved): ssl support for metacat (https)https://projects.ecoinformatics.org/ecoinfo/issues/2352001-06-05T22:40:22ZMatt Jonesjones@nceas.ucsb.edu
<p>Need support for clients to connect to metacat using ssl. Need to determine a<br />reasonable way of generating and assigning certificates (probably we generate<br />them and provide a central registry of metacat servers where clients can get the<br />public keys).</p>
<p>Need support for metacat to contact LDAP over SSL to make this a secure connection.</p> Bug #234 (Resolved): metacat accession number handlinghttps://projects.ecoinformatics.org/ecoinfo/issues/2342001-06-05T17:27:11ZMatt Jonesjones@nceas.ucsb.edu
<p>Metacat currently generates an accession number if one is not provided by the<br />client on insert. We decided last March that this behavior should not be<br />allowed. Thus, metacat needs to be changes so that it requires a valid<br />accession number on insert -- should fail if no accession number is provided, or<br />if the accession number is not unique, or if the revision portion of the<br />accession number is not "1". Thus, "mbj.145.1" would be valid for an insert<br />action, but "knb.189.3" would not be (because the revision is not 1).</p>
<p>On update actions, we need to require an accession number and ensure that the<br />new accession number has a revision number that is one greater than the current<br />accession number revision. This will be some indication that the document being<br />updated "came from" the most recent version. For example, if someone downloads<br />"mbj.1.1" and tries to update it with new accession number "mbj.1.2", it would<br />succeed. If they don't provide a new accession number, it would fail. If they<br />download "knb.1.3", someone else updates it first to create "knb.1.4", and the<br />original user tries to update with "knb.1.4", it would fail, telling them to get<br />the most recent revision and try again. This is basically a simple mechanism<br />for asynchronous update conflicts.</p> Bug #233 (Resolved): authentication always succeeds with empty passwordhttps://projects.ecoinformatics.org/ecoinfo/issues/2332001-05-09T00:57:47ZMatt Jonesjones@nceas.ucsb.edu
<p>If logging into metacat, providing an empty string for a password always results<br />in successful authentication. This is a bug and security hole.</p> Bug #225 (Resolved): finish SSDBM paperhttps://projects.ecoinformatics.org/ecoinfo/issues/2252001-04-10T20:47:13ZChad Berkleyberkley@nceas.ucsb.edu
<p>The paper for the SSDBM conference which is 7/18-7/20/2001 needs to be complete<br />and camera ready by 4/30/2001. Note the reviewer comments attached to the<br />acceptance email.</p> Bug #197 (Resolved): revisionid should be generated on client, not metacathttps://projects.ecoinformatics.org/ecoinfo/issues/1972001-04-09T19:59:32ZMatt Jonesjones@nceas.ucsb.edu
<p>Currently metacat assigns a revisionid when allowing updates of metadata<br />documents. This should be changed so that the client specifies the right<br />revisionid. If a client specifies a revision that is less than the revision<br />stored on metacat, the operation must fail with an appropriate error message<br />(the client can then retrieve the recent revision and resubmit). This requires<br />parallel changes in Morpho.</p> Bug #196 (Resolved): Metacat READ function should set mime typehttps://projects.ecoinformatics.org/ecoinfo/issues/1962001-04-09T19:56:21ZMatt Jonesjones@nceas.ucsb.edu
<p>Metacat read fucntion should set the Content-header: to contain the appropriate<br />mime-type for the data being returned. This will allow clients to distinguish<br />metadata downloads (text/xml) from data downloads (application/octet-stream) and<br />allow any object from metacat to be requested using only the accession #.</p> Bug #194 (Resolved): manage files by accession# on file systemhttps://projects.ecoinformatics.org/ecoinfo/issues/1942001-04-09T19:49:32ZMatt Jonesjones@nceas.ucsb.edu
<p>Need to manage files by accession # on the filesystem when using the file system<br />as a storage system for data.</p> Bug #191 (Resolved): Override protocol handling for http URLshttps://projects.ecoinformatics.org/ecoinfo/issues/1912001-04-09T19:41:36ZMatt Jonesjones@nceas.ucsb.edu
<p>Metacat currently uses a socket system for uploading data files because of an<br />inefficiency in how http urls are handled by the default protocol handler. Need<br />to investigate overrriding that protocol handler so that this problem is not<br />encountered.</p> Bug #190 (Resolved): modify resultset to return all package informationhttps://projects.ecoinformatics.org/ecoinfo/issues/1902001-04-09T19:39:17ZMatt Jonesjones@nceas.ucsb.edu
<p>Need to accomodate changes to EML packages that we have discussed, and to fix<br />the incorrect treatment of packages in Metacat. Currently, Metacat returns a<br />list of related objects, some of which are "transitively" related. this is<br />incorrect behavior. Instead, we need to return the whol elist of related items<br />that are in the package list. Because package has been moved into "resource",<br />this may mean returning the whole resource document might be appropriate. This<br />needs a lot of thought wrt the changes we decided for EML 2.</p> Bug #187 (Resolved): test metacat on Windows NT/2000https://projects.ecoinformatics.org/ecoinfo/issues/1872001-04-09T19:26:12ZMatt Jonesjones@nceas.ucsb.edu
<p>Need to test the Metacat servlet on NT/2000, possibly using an Oracle DB that<br />runs on another host for the time being. Then test with Oracle on NT as well. <br />Eventually, when SQLServer support is enabled, will need to test against SQL<br />Server as well.</p> Bug #183 (Resolved): Port metacat to PostgresSQL and other databaseshttps://projects.ecoinformatics.org/ecoinfo/issues/1832001-02-14T19:59:21ZMatt Jonesjones@nceas.ucsb.edu
<p>To install Metacat broadly at research stations, we need to eliminate the<br />requirement for Oracle as a backend DB. Thus, we need to port Metacat so that<br />it can use PostgresSQL and probably MS SQL Server. To do this, we should create<br />a database interface class that implements the SQL calls needed for each<br />database system. At runtime, Metacat would then load the proper database<br />adapter (that implements the interface) and thus have access to the specific SQL<br />cals needed to access that database. The changes to Metacat involved in this<br />will mainly be limited to two classes (DocumentImpl.java and<br />QuerySpecification.java) as these are the main classes that execute SQL calls,<br />although there will also be some work needed in the AccessControlList.java,<br />DBDTDHandler, DBEntityResolver, DBSAXHandler, DBSAXNode, DBUtil,<br />ForceReplication and ReplicationHandler classes. Owen has agreed to work on this<br />as well.</p> Bug #181 (Resolved): possible security risks in data file uploadhttps://projects.ecoinformatics.org/ecoinfo/issues/1812001-01-24T18:15:00ZChad Berkleyberkley@nceas.ucsb.edu
<p>The data file upload utility could have security holes in it. Ports may not be<br />closed when a client does not send all of the data that it says it is going to. <br />The data file upload utility needs to be thoroughly tested.</p>