Remove all bsh libraries and bsh remaining fonctionnalities.
Bsh interpreter is replaced by GroovyShell evaluation.
To avoid groovyShell undefined var exception, the evaluated expression is analysed, detecting variables used and setting its to null if undefined in context.
Today only alphabetic variable with '_' are detected (i.e. my_variable==null).
Add in example a test replacement of " syntax to simple quote.
Remove all found bsh references
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1752231 13f79535-47bb-0310-9956-ffa450edef68
This is the first patch that takes OFBiz to the point of no return
in terms of switching from Ant to Gradle build system. Specifically
this patch applies the following items:
- Delete all old Ant and Maven build scripts
- Delete 169 Jar libraries and replace them with remote ones in the
gradle build script
- update the gradle build scripts for the dependencies and other minor
changes needed for the script work correctly
All tests pass with the command ./gradlew cleanAll loadDefault testIntegration
Again, thank you for everyone who contributed valueble time and effort in this
process.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1752033 13f79535-47bb-0310-9956-ffa450edef68
* removed the code that implemented the deprecated "call-bsh" operation in Minilang
* removed the BeanShellContainer (it has been disabled a long time ago)
* removed the BeanShell service engine, no more used
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1751947 13f79535-47bb-0310-9956-ffa450edef68
This feature has been contributed a long time ago as an experimental feature but never really matured into a production ready artifact: since it was contributed, no users asked about how to use it, or tested it, or reported bugs or enhancement requests.
More importantly, the feature has always been disabled in the OFBiz framework (i.e. the widget code and the entity engine code were not using it).
This removal should not impact our user base; however, after the removal, an entry to the Attic will be created to keep track of it and to make it easier, to interested users, to get this code, improve it (e.g. by making it dependent on a live external jar, for example MapDB instead of JDBM) and possibly contribute it back to the project.
The most compelling reason for its removal is that the feature was based on a snapshot, unreleased version of JDBM, which is now a mostly dead project: for licensing reasons we are no more allowed to bundle jars in our releases. Additionally, having code that relies an old jar from a no more active project represents a maintenance issue for the community.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1751817 13f79535-47bb-0310-9956-ffa450edef68
(resource == null || resource.length() <= 0)
From Java spec. a String, Collection, Map or CharSequence can't have a negative length, so the pattern above can be reduced to
(resource == null || resource.length() == 0)
which can be replaced using
UtilValidate.isEmpty()
I checked, there are several other occurrences of this pattern (some very old, I guess most were routinely copied from an initial occurrence). I decided to replace all of them, here they are.
I did that already long ago for the
(resource == null || resource.length() == 0)
pattern. It remains 4 of them. Three are justified (because of the dependency on base component or because using isEmpty() on Object is disputable for performance reasons). I took care of the remaining one right aways (in SplashScreen.java)
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1747976 13f79535-47bb-0310-9956-ffa450edef68
I removed duplicate paths were componentname was uselessly repeated after component://componentname/script/org/ofbiz (projectmngr was the only one using there project instead)
I had few typos, easily fixed thanks to the tests. There could be still typos but they are unexpected and anyway no-brainer to fix
I'm still wondering about specialpurpose/example/minilang/README.txt.
I changed its content to
"The minilang directory is for static resources that are interpreted at run time rather than being compiled. This goes on the classpath, but does not get built or put into a JAR file."
not sure it makes any sense, I don't see how this relates to the classpath.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1747712 13f79535-47bb-0310-9956-ffa450edef68
1. The following xsd are duplicated to use namespaces
site-conf.xsd -> site-conf-ns.xsd
widget-form.xsd -> widget-form-ns.xsd
widget-screen.xsd -> widget-screen-ns.xsd
widget-menu.xsd -> widget-menu-ns.xsd
simple-methods.xsd-> simple-methods-ns.xsd
For example, in site-conf-ns.xsd, we have the following document level attribute
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified" xmlns="http://ofbiz.apache.org/Site-Conf" targetNamespace="http://ofbiz.apache.org/Site-Conf">
instead of
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified">
2. The above schema are used into compound-widgets.xsd so that compound widgets use only one consolidated schema.
3. Update ExampleCompoundWidgets.xml to use the new compound-widgets.xsd. For example
<compound-widgets xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:sc="http://ofbiz.apache.org/Site-Conf"
xmlns:sm="http://ofbiz.apache.org/Simple-Method"
xmlns:wf="http://ofbiz.apache.org/Widget-Form"
xmlns:wt="http://ofbiz.apache.org/Widget-Tree"
xmlns:wm="http://ofbiz.apache.org/Widget-Menu"
xmlns:ws="http://ofbiz.apache.org/Widget-Screen"
xsi:noNamespaceSchemaLocation="../../../../framework/widget/dtd/compound-widgets.xsd">
<sc:site-conf>
<sc:request-map uri="CompoundWidgets1">
<sc:security https="true" auth="true"/>
<sc:event type="simple" invoke="CompoundWidgetsFunc" path="component://example/widget/example/ExampleCompoundWidgets.xml"/>
<sc:response name="success" type="view" value="CompoundWidgets1"/>
</sc:request-map>
<sc:request-map uri="CompoundWidgets2"><sc:security https="true" auth="true"/><sc:response name="success" type="view" value="CompoundWidgets2"/></sc:request-map>
<sc:view-map name="CompoundWidgets1" type="screen" page="component://example/widget/example/ExampleCompoundWidgets.xml#CompoundWidgets1"/>
<sc:view-map name="CompoundWidgets2" type="screen" page="component://example/widget/example/ExampleCompoundWidgets.xml#CompoundWidgets2"/>
</sc:site-conf>
...... the rest
4. Change java code to support reading xml with namespace (i.e. xml for compound widgets)
jleroux: I simply replaced tabs by spaces in Java files and removed this commented out line
<compound-widgets xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://ofbiz.apache.org/dtds/compound-widgets.xsd">
in ExampleCompoundWidgets.xml
We decided to duplicate the simple-methods.xsd rather than updating all concerned files. Else we would have to update the attributes at document level for rest of the controllers, menus, forms, simple methods and screens. For example, in controller.xml, we would have to change
xsi:noNamespaceSchemaLocation="http://ofbiz.apache.org/dtds/site-conf.xsd”
to
xmlns="http://ofbiz.apache.org/sc” xsi:schemaLocation="http://ofbiz.apache.org/dtds/site-conf-ns.xsd”>
But especially we would have to use the same syntax with namespaces prefixes than in ExampleCompoundWidgets.xml (see example above). This can be discussed because with auto-completion it's not a big deal, but could be overwhelming for existing custom projects...
James mentionned that "the schema location at compound-widgets.xsd is relative to the project...". I don't see any problems with that. We can keep as is as long as we maintain the relation (ie if the main XSDs are moved). Moreover with the previous addition of compound-widgets.xsd in widget-catalog.xml it seems all things work correctly (eg w/o the new *-ns.xsd yet committed, the auto-completion works well in ExampleCompoundWidgets.xml
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1746302 13f79535-47bb-0310-9956-ffa450edef68
See CVE-2016-2175: Apache PDFBox XML External Entity vulnerability
I did not try to update to version 2.0.1.
I only tested by using https://localhost:8443/example/control/ExampleReportPdfOptions?exampleId=EX01 but I got nothing, so I tried with R15.12 before backporting with the same issue so I guess it's unrelated with this update moreover with both branches I get an error in log for the barcode PDF I will open a Jira
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1745751 13f79535-47bb-0310-9956-ffa450edef68
Because using SecureRandom comes with a cost, I have identified the places where it's reasonable to keep the non secured Random (like tests, internal sequences, etc.).
Ant to use UUID.randomUUID to generate the external link id
Also, though there are no real proven vulnerabilities, I decided to backport as much as possible since it's now public.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1741684 13f79535-47bb-0310-9956-ffa450edef68
Currently the following is in the ofbiz-component.xsd:
<xs:attribute name="menu-name" default="main">
<xs:simpleType>
<xs:restriction base="xs:token">
<xs:enumeration value="main"/>
<xs:enumeration value="secondary"/>
</xs:restriction>
</xs:simpleType>
</xs:attribute>
However, this is limiting the options that can be applied generates unnecessary errors when developers opt for another methodology.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1739968 13f79535-47bb-0310-9956-ffa450edef68
tika-parsers-1.7.jar -> tika-parsers-1.12.jar
tika-core-1.7.jar -> tika-core-1.12.jar
This update induce updating of :
pdfbox-1.8.5.jar -> pdfbox-1.8.11.jar
jempbox-1.8.5.jar -> jempbox-1.8.11.jar
fontbox-1.8.5.jar -> fontbox-1.8.11.jar
commons-compress-1.9.jar -> commons-compress-1.11.jar
And pdfbox-1.8.11.jar need the introduction of :
jce-jdk13-154.jar under MIT licence (www.bouncycastle.org)
Apache Tika is needed to test the screenfop on issue OFBIZ-6497. So I move all library from applications/content/lib/ to framework/base/lib/ to didn't create a dependency between widget and content
Related issue OFBIZ-7015
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1739551 13f79535-47bb-0310-9956-ffa450edef68
Implements a PagedList object (stores list of data items, resultset size, viewSize, viewIndex, lowIndex (called startIndex) and highIndex (called endIndex), also a utils class that returns the PagedList object based on a EntityListIterator, viewSize and viewIndex.
And implements a queryPagedList method in EntityQuery, used as an example in findOrders service Java implementation
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1738900 13f79535-47bb-0310-9956-ffa450edef68
Since commons.httpclient.HttpClient is now deprecated/not-maintained an vulnerable I prefer to simply put a warning at the top of the class. So people can pick what they prefer and is handy for their problem at than
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1737695 13f79535-47bb-0310-9956-ffa450edef68
Fixes "Comment out RMI related code because of the Java deserialization issue" - https://issues.apache.org/jira/browse/OFBIZ-6942
We decided to comment out as less as possible because once, in the start and both properties; the rmi part is off and the related test services are off there is no RMI related danger left (test services are not a danger but would fail during tests run).
It's then easier for users who need RMI in their projects to have only to uncomment those and not digg everywhere.
[CVE-2016-2170] The infamous Java serialization vulnerability
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1736083 13f79535-47bb-0310-9956-ffa450edef68
I decided to comment out as less as possible because once the RMI loaders, the RMI dispatcher and the related test services are off there is no RMI related danger left (test services are not a danger but would fail during tests run). It's then easier for users who need RMI in their projects to have only to uncomment those and not digg everywhere. Because the naming (JNDI) server relies on the rmi loader it will also be commented out.
[CVE-2016-2170] The infamous Java serialization vulnerability
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1735569 13f79535-47bb-0310-9956-ffa450edef68
Log:
==========================================================
- Make use of ResourceLoader.readXmlRootElement, which is
non-caching; this fixes a deprecation warning.
- Use ResourceLoader.readXmlRootElement, which is a
non-caching version. This fixes a deprecation warning.
- Provide method variants (readXml{Document,RootElement}) that contain the body of the deprecated methods (getXml{Document,RootElement}).
==========================================================
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1731399 13f79535-47bb-0310-9956-ffa450edef68
Log:
============================================================
UtilDateTime implements deprecated methods, so add the
@Deprecated tag to stop javac from producing a warning.
============================================================
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1731385 13f79535-47bb-0310-9956-ffa450edef68
Log:
============================================================
new Configuration() is deprecated; instead, the desired
freemarker version should be passed in. To make it easier to control
the version centrally, add a helper method on FreeMarkerWorker that
wraps the instantiation of the Configuration object.
============================================================
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1731380 13f79535-47bb-0310-9956-ffa450edef68
This implies to update also Xerces from 2.9.1 to 2.11.0 and also xml-apis from 2.9.1 to 1.4.01 (2.9.1 was a wrong version number. It was actually part of the Xerces 2.9.1 package but I was unable to find the real version number then at https://xerces.apache.org/xerces2-j/releases.html)
Also updates Xalan from 27(?) to 2.7.2 in cmssite/template/docbook/extensions. I rendered https://localhost:8443/cmssite/cms/APACHE_OFBIZ_HTML w/o issues
Note: According to the DOM Level 3 specification and DOM Level 2 errata the createElementNS and createAttributeNS methods convert empty string namespaceURI to null.
jleroux: though the tests pass I'm not sure all is covered...
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1730882 13f79535-47bb-0310-9956-ffa450edef68
Since OFBIZ-6867 is now done, it will no longer be used OOTB and anyway should not be needed because we should preferably always use sessionIds in cookies and newer have sessionsIds in URLs.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1728121 13f79535-47bb-0310-9956-ffa450edef68