Metacat: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362017-11-17T21:19:00ZEcoinformatics Redmine
Redmine Bug #7229 (New): Mis-Formatting of Data Package Contentshttps://projects.ecoinformatics.org/ecoinfo/issues/72292017-11-17T21:19:00ZThomas Thelen
<p>In MetacatUI we're getting a slight mis-formatting when displaying data package contents. This can be seen in the attached images. The issue was initially reported in MetacatUI as issue 379.</p>
<p><a class="external" href="https://github.com/NCEAS/metacatui/issues/379">https://github.com/NCEAS/metacatui/issues/379</a></p>
<p>From Bryce,</p>
<p>"The HTML in question is actually produced by Metacat and MetcatUI is just rendering it without modification from Metacat's View Service. ... The fix would involve changing the underlying eml-2 XSLT."</p> Bug #7182 (New): Allow partial package downloads when some of the objects are private https://projects.ecoinformatics.org/ecoinfo/issues/71822017-04-13T15:10:57ZLauren Walkerwalker@nceas.ucsb.edu
<p>When you try to download a package that has at least one private object, you get a 401 - Unauthorized response. When I am authorized to read at least one object in a package, I would expect to still be able to download the .zip package with those objects.</p>
<p>It's difficult for MetacatUI to tell when the "Download All" will fail, since it would need to check the /isAuthorized/{pid}?action=read result for every single object in the package, which can sometimes be >100. So right now we have an issue where users are getting a failed package download.</p> Bug #6959 (New): http response charset not includedhttps://projects.ecoinformatics.org/ecoinfo/issues/69592016-02-11T00:29:11ZPeter Slaughterslaughter@nceas.ucsb.edu
<p>Metacat doesn't include the content type charset in the http response headers:</p>
<p>curl -o test.dat -v "https://knb.ecoinformatics.org/knb/d1/mn/v2/object/solson.25.5" returns</p>
<p>Content-Type: text/csv</p>
<p>but should be something like:</p>
<p>Content-Type: text/xml;charset=UTF-8</p>
<p>so that the client doesn't have to assume the charset</p> Feature #6848 (New): Add link to send another verification emailhttps://projects.ecoinformatics.org/ecoinfo/issues/68482015-09-30T16:29:26ZLauren Walkerwalker@nceas.ucsb.edu
<p>This has been something suggested by a few users when they don't get their original verification email</p> Feature #6034 (New): Simplify configuration for certificate delegationhttps://projects.ecoinformatics.org/ecoinfo/issues/60342013-07-09T21:32:35Zben leinfelderleinfelder@nceas.ucsb.edu
<p>So we don't loose this in the shuffle - we could probably fill out some of the MyProxy configuration values based on Metacat properties. But I don't think we can really generate provate keys and register nodes with CILogon...</p> Bug #5836 (New): Logshttps://projects.ecoinformatics.org/ecoinfo/issues/58362013-01-30T18:16:16ZBrendan Hahnhahn@nceas.ucsb.edu
<p>Logs are kind of a mess, mixing multiple logging object types along with System.* output. Metacat has its own log directory, but only replication events are recorded there. Default build/distribution has uneven log levels set.</p> Bug #5835 (New): Leakshttps://projects.ecoinformatics.org/ecoinfo/issues/58352013-01-30T17:44:53ZBrendan Hahnhahn@nceas.ucsb.edu
<p>There are significant memory leaks that will eventually hang or kill the application server if metacat is restarted without terminating the entire server process. E.g. use of log4j dynamic configuration leaves orphan threads about (not really metacat's fault -- I'm amazed this is <strong>still</strong> a problem). Other threads, classloaders, etc along with the associated resources get left behind as well.</p>
<p>In short: should clean up better.</p> Bug #5830 (New): Deployment directory incorrecthttps://projects.ecoinformatics.org/ecoinfo/issues/58302013-01-29T23:05:21ZBrendan Hahnhahn@nceas.ucsb.edu
<p>Metacat stores its deployment directory in metacat.properties, where the value can lose sync with the actual deployment. The deployment directory is a discoverable feature of the environment and should be dynamically determined rather than configured as a property.</p> Bug #5828 (New): docids not validatedhttps://projects.ecoinformatics.org/ecoinfo/issues/58282013-01-29T22:58:48ZBrendan Hahnhahn@nceas.ucsb.edu
<p>Externally provided document identifiers are not always checked for syntax (e.g. multipart-form submission) and produce strange errors (e.g. string index out of range) if invalid.</p> Bug #5821 (New): Allow certificate-based Metacat administrationhttps://projects.ecoinformatics.org/ecoinfo/issues/58212013-01-24T22:12:04Zben leinfelderleinfelder@nceas.ucsb.edu
<p>As we move toward the DataONE API where the MN does not provide identity and authorization services, perhaps the Metacat administrative functions should also follow suit. This would be a pretty large change for our users, but ultimately will simplify things so that we are not using two different identity/auth schemes to manage a single server.</p>
<p>In cases where the Metacat administrator did not have a useable (CILogon) identity we cold provide a utility to generate a client certificate for administrative use (or something akin to this). Ultimately this would need to be available in a browser UI where the bulk of our admin/config is performed.</p> Bug #5608 (New): Enable all FK constraints in Metacat production [copies]https://projects.ecoinformatics.org/ecoinfo/issues/56082012-05-14T19:14:31Zben leinfelderleinfelder@nceas.ucsb.edu
<p>Looks like the FK constraints have been removed from the production knb database.</p>
<blockquote>
<p>select conname, contype, conkey, confkey from pg_constraint;</p>
</blockquote>
<p>Need to see what FKs are no long satisfied and potentially fix them as best we can so that the constraints can be re enabled.</p> Bug #5599 (New): absence of line feeds in eml causes pathQuery to not find some elementshttps://projects.ecoinformatics.org/ecoinfo/issues/55992012-05-07T22:54:42Zgastil gastilmarygastil@yahoo.com
<p>Presence of line feeds seems to be needed for an eml doc to get loaded properly so pathQuery can find attributeList or attribute. Not just one line feed at the end.<br />We detected this on metacat 1.9.5 at metacat.lternet but tested it on metacat 2.0 (lava.lternet)</p>
<p>Evidence:<br />in lava.lternet.edu</p>
<p>knb-lter-kbs.10.19 has no line feeds at all in the document.</p>
<p>revision 20 is same as 19 except stmml-1.1 is spelled right.<br />revision 21 is same as 20 except it has one line feed at the end of the file.<br />(so revision 21 has one line)</p>
<p>revisions 19 thru 21, while they were the last revision, did not have their attributeList found by pathQuery.</p>
<p>revision 22, with 165 lines feeds DOES have its attributeList seen by pathQuery.</p>
<p>wc -l knb-lter-kbs.10.*<br /> 0 knb-lter-kbs.10.19.xml<br /> 0 knb-lter-kbs.10.20.xml<br /> 1 knb-lter-kbs.10.21.xml<br /> 165 knb-lter-kbs.10.22.xml</p>
<p>pathQuery result snippets from two separate queries (when two different revisions were the last revision):</p>
<p><document><br /><docid>knb-lter-kbs.10.22</docid><br /><docname>eml</docname><br /><doctype>eml://ecoinformatics.org/eml-2.1.0</doctype><br /><createdate>2012-05-07</createdate><br /><updatedate>2012-05-07</updatedate><br /><param name="attributeList"></param><br /><param name="@packageId">knb-lter-kbs.10.22</param><br /></document></p>
<p>older query:</p>
<p><document><br /><docid>knb-lter-kbs.10.21</docid><br /><docname>eml</docname><br /><doctype>eml://ecoinformatics.org/eml-2.1.0</doctype><br /><createdate>2012-05-07</createdate><br /><updatedate>2012-05-07</updatedate><br /><param name="@packageId">knb-lter-kbs.10.22</param><br /></document></p> Bug #5553 (New): setaccess action may have deleting access rule functionalityhttps://projects.ecoinformatics.org/ecoinfo/issues/55532011-11-23T22:55:24ZJing Taotao@nceas.ucsb.edu
<p>Currently, setaccess action can only add access rules to the metacat. There is a limitation.</p>
<p>Sometimes, we want to keep granting or revoking a allow public read access rules for a document.</p>
<p>If we choose "denyFirst" as the order type(metacat can't change order type when a document set it), we can't revoke the public readable access if we granted it.</p>
<p>If we choose "allowFirst" as the order type, we can't regrant it if we denied it.</p>
<p>So if setaccess action (or we can have another action deleteaccess), this scenario can be avoided.</p> Bug #5548 (New): Replace COS Mime multipart libraryhttps://projects.ecoinformatics.org/ecoinfo/issues/55482011-11-21T22:22:38ZJing Taotao@nceas.ucsb.edu
<p>Here is a comment from Matt in bug 5543:</p>
<p>The COS Mime multipart library is incredibly old, and we should<br />seriously consider replacing it. On the client side for DataONE, we've<br />switched to using the nicely maintained Apache HttpComponents implementation,<br />particularly the MIME MultipartEntity class for representing our mime messages<br />(<a class="external" href="http://hc.apache.org/httpcomponents-client-ga/httpmime/apidocs/org/apache/http/entity/mime/MultipartEntity.html">http://hc.apache.org/httpcomponents-client-ga/httpmime/apidocs/org/apache/http/entity/mime/MultipartEntity.html</a>).</p> Bug #5535 (New): The 611 CDR documents replicated to KNB need to be repaired by fixing their inli...https://projects.ecoinformatics.org/ecoinfo/issues/55352011-11-09T18:31:39ZJing Taotao@nceas.ucsb.edu
<p>This bug is the section 4 of the ongoing concerns:<br /><a class="external" href="http://bugzilla.ecoinformatics.org/show_bug.cgi?id=3296">http://bugzilla.ecoinformatics.org/show_bug.cgi?id=3296</a></p>
<p>During the replication, the "&" and "<" in the inline data section of the 611 CDR document in LTER were changed to "&" and "<" in knb site.</p>
<p>We need to fix it. Duane proposed a solution on the bug 3296.</p>