fixing columns in create form wizard to not move around wildy once files are uploaded
fixing some bugs in repeats as a result of the chiba1.3 upgrade.
- xpaths were not being resolved for newly created instances (and shouldn't be)
- xpaths for nested repeats weren't behaving properly
fixed caching issue in edit which was causing the wrong form to be loaded for instance data
using alfresco namespace as defined in NamespaceService
passing namespace uris and prefixes to js to avoid errors if they should change
fixed a bug where the renditions property of forminstancedata was getting duplicate renditions (and thus regenerating way too much stuff when doing an edit)
using an italic 'description not set' when description isn't set in several screens.
using the avm service to set properties in renderingenginetemplateimpl.
upgrading to xalan-2.7.0 in order to be able to use bsf for extension functions.
adding the path that was not found to AVMNotFoundExceptions. very helpful when debugging.
substantial refactoring of rendering engines in preparation for integration with TemplateService.
- implementing a common model as a hash of QNames to objects. for method, providing a simple method wrapper called TemplateProcessorMethod which takes an array of Objects as a parameter, and returns an object. it is up to the template processors to properly convert arguments. a QName is used for the variable name rather than a string in order to include a namespace prefix (needed for xsl, and generally better looking).
- for xsl, using javascript bindings for formdatafunctions, which using liveconnect within rhino to call into the xsl rendering engine to evaluate the function. it ends up generating this js block into a xalan:component within the xsl tempalte:
// gets the handle to the backing java object which can invoke the function based on id
var _xsltp_invoke = java.lang.Class.forName('org.alfresco.web.forms.XSLTRenderingEngine$ProcessorMethodInvoker').newInstance();
// utility to convert js arrays into java
function _xsltp_to_java_array(js_array) {
var java_array = java.lang.reflect.Array.newInstance(java.lang.Object, js_array.length);
for (var i = 0; i < js_array.length; i++) { java_array[i] = js_array[i]; }
return java_array; }
// js handles to each of the form data functions which uses _xsltp_invoke to call the actual method
function _getAVMPath() { return _xsltp_invoke.invokeMethod('_getAVMPath8829055', _xsltp_to_java_array(arguments)); }
function parseXMLDocuments() { return _xsltp_invoke.invokeMethod('parseXMLDocuments12235190', _xsltp_to_java_array(arguments)); }
function parseXMLDocument() { return _xsltp_invoke.invokeMethod('parseXMLDocument15280968', _xsltp_to_java_array(arguments)); }
xml model data is inferred as a root namespace document within the model hash provided.
- for freemarker, things pretty much work as they did before, just i now need to convert values by hand to TemplateModels
fixed a bug with hidden iframe upload. seems like the complexity of actually cloning the file input element is unnecessary and that simply attaching the node in two places within the dom works just fine.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@4764 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Reworked handling of path lookup failures to not throw exceptions
internally, to improve performance of certain layered directory
operations. Unfortunately there remains at least one scenario,
handling of bulk loads, and promotions of deeply nested directories in
layered contexts, in which performance is considerably less than
ideal.
Made AVMService.createBranch() and AVMSyncService.update() perform
implicit snapshots of source tree's stores before proceeding.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3812 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
without the bad aftertaste. It's needed for the AVMSyncService's
update method which will move new versions of nodes from one AVM
tree to another. Also a checkpoint for AVMSyncService. One can safely
call compare and get back an empty List of AVMDifferences.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3784 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
node that has been deleted in a version. This makes tree comparisons easier by
resolving the cases in which one wants to know whether a deletions is more or less
recent than another version of a node. Along the way got rid of DeletedChild and
friends including a table in the schema, since DeletedNode takes on all the work
that it performed.
Deleted two illegally named files in Virgin content.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3748 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
now derived from AbstractNodeServiceImpl to make policy awareness easier.
Moved addDefaultPropertyValues from DbNodeServiceImpl to
AbstractNodeServiceImpl so AVMNodeService can use it. A fix for an
NPE that showed up in stress test.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3603 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Schema has been updated, DAO written, garbage collection updated to clean
out aspects. Also some seemingly unnecessary changes in visibility
declarations to deal with strange intermittent Spring wiring failures in
one of my tests.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3560 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
wasn't it) is now AVMRepository. What used to be Repository is now AVMStore as it
more closely matches what is meant by a store in Alfresco. Many adjustments
in ancillary class names, references, and comments followed.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/WCM-DEV2/root@3329 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
