Metacat: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362002-04-08T21:21:34ZEcoinformatics Redmine
Redmine Bug #465 (Resolved): Data diretoryhttps://projects.ecoinformatics.org/ecoinfo/issues/4652002-04-08T21:21:34ZJing Taotao@nceas.ucsb.edu
<p>Now if the Metacat doesn't have permssion, it could not create the data <br />direcotry automaticly. Uer should create the data directory and change it <br />permission manually.</p>
<p>We should let it created automaticly.</p> Bug #452 (Resolved): Configure ssl in dev and ecoinfohttps://projects.ecoinformatics.org/ecoinfo/issues/4522002-04-02T19:00:43ZJing Taotao@nceas.ucsb.edu
<p>We already configured ssl in local machine (Tomcat standalone). Now we need to <br />configure dev and ecoinfo. But in these machines, apache is the web server and <br />need more configuration than tomcat standalone.</p> Bug #449 (Resolved): Enable ssl for metacat and morphohttps://projects.ecoinformatics.org/ecoinfo/issues/4492002-03-26T21:27:15ZJing Taotao@nceas.ucsb.edu
<p>Now communication between Metacat and Morpho uses plain text. In order to <br />increase security, we want to enable SSL.</p> Bug #444 (Resolved): Create a JUnit test framework for MetaCathttps://projects.ecoinformatics.org/ecoinfo/issues/4442002-03-14T17:56:54ZJing Taotao@nceas.ucsb.edu
<p>JUint is powerful test tool and we need apply it to MetaCat. The first step is <br />to create a framework for it. The tasks include:<br />1. Revise build.xml and add a target named "test". When devopler type "ant <br />test", the JUnit test classes will be excuted.<br />2. Create a file stucture to store the test classes.<br />3. Write a simple test class to run.</p> Bug #442 (Resolved): There is no entry in xml_access table for access document itselfhttps://projects.ecoinformatics.org/ecoinfo/issues/4422002-03-07T22:50:36ZJing Taotao@nceas.ucsb.edu
<p>When we create a data package in Morpho and save it into metacat. But there is <br />no permission in xml_access table for access document itself.</p>
<p>Now we discard a permission rule - if there is no entry in xml_access table <br />for a document, this document will have same permission to data set permission <br />to the user. The rule now we are using is: if the there is no entry in <br />xml_access table for a document, if the user is owner, it has permission. <br />Otherwise, it doesn't have permission.</p>
<p>So we need to create entries for access document itself in xml_access table.</p> Bug #441 (Resolved): Unable to increment data id revision numbershttps://projects.ecoinformatics.org/ecoinfo/issues/4412002-03-05T23:28:31ZDan Higginshiggins@nceas.ucsb.edu
<p>In an attempt to let the Morpho user update a data file (NOT a metadata doc), it<br />was discovered that Metacat returns the following when an attempt is made to<br />insert a revised data document, eg 801.1 --> 801.2</p>
<p>Same code works when a new id is used rather than one with incremented version #.</p>
<p><error><br />ORA-00001: unique constraint (TAO.XML_DOCUMENTS_PK) violated</p> Bug #437 (Resolved): Document Access Control Ruleshttps://projects.ecoinformatics.org/ecoinfo/issues/4372002-02-26T19:28:17ZJing Taotao@nceas.ucsb.edu
<p>Now the rules for document access control in Morpho and MetaCat are:<br />1. If a data package doesn't have access control list, the owner of <br />datapackage will be the only full control.<br />2. If a data package does have access control list, apply it.<br />3. If a member document of data package doesn't have access control list, <br />access control of data set document will apply to the member document.</p>
<p>The third rule is no good and need to get rid of it.</p>
<p>The reason is:<br />If a package smith.23.1 has a triple to point smith.34.1 and smith.34.1 <br />doesn't have access control list. It is okay because they all belong to a <br />owner - Smith. If another package named henry.2.1 has a triple to point <br />smith.34.1 too. Because smith.34.1 doesn't access control list and henry can <br />access it. Maybe henry doesn't allow to access it.</p> Bug #436 (Resolved): Bug of MetaCat in handling permission issuehttps://projects.ecoinformatics.org/ecoinfo/issues/4362002-02-22T21:05:57ZJing Taotao@nceas.ucsb.edu
<p>After login the metacat from that web page by using <br />"uid=tao,o=NCEAS,dc=ecoinformatics,dc=org" as username, the data packages were <br />gotten correctly. If delect the url and type <br />"dev.nceas.ucsb.edu/tao/servlet/metacat?action=read&docid=mike.4.12". Though <br />we are not the data package's owner, we can read the document in the browser <br />even the principal value in document access permission is not public.</p>
<p>It seems Metacat have some bug in access control in document level. For <br />exmaple, read the document directly.</p> Bug #417 (Resolved): Is sequential numbering of versions needed?https://projects.ecoinformatics.org/ecoinfo/issues/4172002-02-08T17:06:50ZDan Higginshiggins@nceas.ucsb.edu
<p>Currently, version changes of documents must be sequentially numbered. This<br />means that someone working locally in Morpho might make 6 changes to a document<br />and 6 versions must be sent to Metacat in order, even though only the last one<br />is meant to be submitted. Also, if any of the early version is invalid then the<br />whole submission fails. Likewise, a download of any datapackage requires the<br />download of all version, even if only the most recent is needed.</p>
<p>Therefore, I suggest that submission of new versions only require that the new<br />version number be greater than the current one, not exactly one greater.</p> Bug #411 (Resolved): package export featurehttps://projects.ecoinformatics.org/ecoinfo/issues/4112002-02-05T00:57:51ZMatt Jonesjones@nceas.ucsb.edu
<p>Metacat needs to be able to produce an exported version of a data package in the<br />same ZIP format that morpho creates data package exports. This involves getting<br />all of the metadata and data files associated with the package, arranging them<br />in the directory correctly, and producing an html summary. The code for this is<br />essentially complete in morpho and should usable in metacat almost directly.</p>
<p>Note that the morpho export dows some funny things, like puts all of the data<br />files inthe metadata directory. We need to fix this as well. I suggest an<br />export structure like this:</p>
<p>package.dir
|
|---- data
| |
| |------d1.1
| |------d2.1
|
|---- metadata
| |
| |------m3.1
| |------m4.1
|
|---- metadata.html</p> Bug #408 (Resolved): Authentication is by-passed when a client follows an LDAP referralhttps://projects.ecoinformatics.org/ecoinfo/issues/4082002-02-01T22:29:21ZChris Jonescjones@nceas.ucsb.edu
<p>The ldap.ecoinformatics.org LDAP database services the dc=ecoinformatics,dc=org <br />tree, but is extended by LDAP referral objects that point to external LDAP <br />servers (e.g. dc=lter,dc=ecoinformatics,dc=org and dc=piscoweb,dc=org). When a <br />client (Morpho or web client) attempts to authenticate against ldap (via <br />metacat), if the user exists in one of the referred trees, authentication will <br />succeed, whether the password is correct or not. If the user entry is within <br />the main dc=ecoinformatics,dc=org tree, authentication will correctly fail if <br />the password is incorrect.</p> Bug #407 (Resolved): ldapweb.cgi confusing wrt 'other' organizationshttps://projects.ecoinformatics.org/ecoinfo/issues/4072002-01-29T16:43:56ZMatt Jonesjones@nceas.ucsb.edu
<p>From Dan Higgins:<br /> I discovered that I need to set the organization to 'unaffiliated' in order<br />to login after creating an account with 'Other' as the organization on the KNB<br />web page! This is likely to be pretty confusing for many potential users.<br />Perhaps the web page should have 'unaffiliated' rather than 'other'?</p>
<p>Comments from Matt Jones:<br />I agree. Will change on web forms.</p> Bug #403 (Resolved): LDAP and LTER personnel password synchronization problemshttps://projects.ecoinformatics.org/ecoinfo/issues/4032002-01-23T17:53:32ZDavid Blankmandblankman@lternet.edu
<p>Discovered that the the random password generator function in the new person entry script (entry <br />into sql server database and LDAP database) is actually being called twice. Need to fix this. <br />Currently the password entered into the database and the one emailed to users don't match.</p> Bug #398 (Resolved): error in handling '&' in response to querieshttps://projects.ecoinformatics.org/ecoinfo/issues/3982002-01-15T04:49:23ZDan Higginshiggins@nceas.ucsb.edu
<p>There appears to be a problem in how Metacat handles entities like '&' in<br />responding to queries. If a data file name contains a '&' that gets encoded into<br />the triple linking the datafile to the datapackage. It get encoded as '&' in<br />the text, but when one looks at the response to an 'squery' the triple shows up as</p>
<p>"><subject>higgins.498</subject><relationship>isDataFileFor(96l</relationship><object>higgins.497</object></triple><triple><subject>higgins.498</subject><relationship>&</relationship><object>higgins.497</object></triple>"</p>
<p>actually, the relationship should have & after the "(961..." <br />Apparently, entities are not created when characters like '&' are pulled from<br />the database as the result of an 'squery'</p> Bug #366 (Resolved): Install & Test LDAP SSL Componenet for o=lter serverhttps://projects.ecoinformatics.org/ecoinfo/issues/3662001-12-03T18:00:38ZDavid Blankmandblankman@lternet.edu
<p>Need to reconfigure o=lter LDAP server to support SSL, also implement Hashed password for <br />password creation.</p>