RSS

Category Archives: XPages

Reading and writing Resource Bundles – an alternative approach

Dan Soares explains at NotesIn9 a nice way to read and write resource bundles. It’s a good read so enjoy his work. I was also faced with the problem of reading and a specially writing properties files, because Switzerland is multilingual and my idea was to give the power of translation to the hands of the business project leaders. But how to write properties files, and how to do it in am more productive way? I did go the same way as Dan, I did also find some sample from Sven Hasselbach, but the I asked my self, how can I provide this in an easy to consume way to my developers and to the community. This was the birth of a new part in the XPages Toolkit.

Btw. we started with the development of the XPages Tookit one year before we mad it public available on OpenNTF. And we want to be shure that we provide something useful and approved by our own developer team.

The result is a very easy to use API (requirement: Install the XPages Toolkit):

SSJS:
to get the properties as java.util.Properties object:
var properties = xptPropertiesBean.getProperties( databaseFilePath,
                               fileNameOfyourProperties );
to save the properties:
var success = xptPropertiesBean.saveProperties( databaseFilePath, 
                               fileNameOfYourProperties, properties);
Or to use in Java:
XPTPropertiesBean bean = org.openntf.xpt.properties.beans.XPTPropertiesBean.get();
Properties props = bean.getProperties( databaseFilePath, 
                                fileNameOfyourProperties ); 
int success = bean.saveProperties( databaseFilePath, 
                                fileNameOfYourProperties, properties);

But be aware of the following. Each change to properties seems to reload your application on the server.

 

To stay informed about the recent development on the XPages Toolkit, please register here and discuss ideas and give feedback.

 
Leave a comment

Posted by on September 6, 2014 in Domino, Java, OpenNTF, XPages

 

org.openntf.junit.xsp – now with EasyMock support

Frank van der Linden asked in his blog post (http://elstarit.nl/?p=157) if EasyMock or PowerMock can be integrated into the org.openntf.junit.xsp and Ryan J. Baxter mentioned also that a testing framework without a good mocking solution is only the half worth.

Only the half worth is a good reason to complete the effort. So I implemented also EasyMock and PowerMock to the org.openntf.junit.xsp. With version 1.1.0 you get the capabilty to mock objects. You may ask how does EasyMock work. Her a small example:

Imagine this. You have this class called Share. The class can be initialized with a builder method called initFromDocument( Document doc) and you want to test this method.

package org.openntf.junit.example.bo;

import lotus.domino.Document;

public class Share {

    private String m_ShareName;
    private int m_Count;
    private int m_PricePerShare;
    
    public static Share initFromDocument( Document doc) {
        Share share = new Share();
        try {
            share.m_ShareName = doc.getItemValueString("ShareName");
            share.m_Count = doc.getItemValueInteger("Count");
            share.m_PricePerShare = doc.getItemValueInteger("PricePerShare");
        } catch (Exception ex) {
            ex.printStackTrace();
            return null;
        }
        return share;
    }
    
    public String getShareName() {
        return m_ShareName;
    }
    public int getCount() {
        return m_Count;
    }
    public int getPricePerShare() {
        return m_PricePerShare;
    }

    public Object getValue() {
        return m_Count * m_PricePerShare;
    }
    
    
}

Your intention is now to write a test for that (or better you write first the test). But you wont initialize a Notes Document, because one thing that unit testing is about has to do with isolation. At this point comes EasyMock to the game:

package org.openntf.junit.example;

import lotus.domino.Document;

import org.junit.Test;
import org.openntf.junit.example.bo.Share;
import org.openntf.junit.xsp.easymock.EasyMockWrapper;

import static org.easymock.EasyMock.*;
import static org.junit.Assert.*;

public class ShareTest {

	@Test
	public void testShareWithDocumentMock() {
		Document docMock = EasyMockWrapper.createNiceMock(Document.class);
		try {
			expect(docMock.getItemValueString("ShareName")).andReturn("WebGate");
			expect(docMock.getItemValueInteger("Count")).andReturn(5);
			expect(docMock.getItemValueInteger("PricePerShare")).andReturn(2870);
			replay(docMock);
			Share shareWebGate = Share.initFromDocument(docMock);
			assertEquals("WebGate", shareWebGate.getShareName());
			assertEquals(5 * 2870, shareWebGate.getValue());
		} catch (Exception e) {
			e.printStackTrace();
			assertFalse(true);
		}
	}
}

With

Document docMock = EasyMockWrapper.createNiceMock(Document.class);

initialize EasyMock a new document object. You may recognize that instead of calling EasyMock.createNiceMock(), I’m calling EasyMockWrapper.createNiceMock(). This is needed because of security issues.

Now we can “record” what will happen to the docMock object. We have 1 call of getItemValueString(“ShareName”) and 2 calls to getItemValueInteger(). For each call can we define what value should be returned.

expect(docMock.getItemValueString("ShareName")).andReturn("WebGate");
expect(docMock.getItemValueInteger("Count")).andReturn(5);
expect(docMock.getItemValueInteger("PricePerShare")).andReturn(2870);

Like on the tape recorder. It’s time to go back to start. You do this with

 
 replay(docMock);

Now our mocked document object is ready and you can test the initFromDocument method.

There is a lot more to cove about mocking and this only the start.

 

Have fun (and don’t forget to use the EasyMockWrapper)

Christian

 

 
1 Comment

Posted by on September 2, 2014 in Domino, Java, OpenNTF, XPages

 

Performance, some findings!

It’s a long time since my last post. But after my sons heart surgery, my time was consumed by one of the biggest challenge that we as WebGate ever has to face. I will definitely write about this, after the successful closing of the project.

But in the mean time, I have done some experiments with the XPagesToolkit and large datasets. The question that I’ve to answer is very easy. How to process a large dataset like 100’000 documents to a nice List of contact objects, or in java spoken: List contacts = ContactSorageServer.getInstance().getAll().

My first question was of course: Is it really mandatory to have 100’000 Contacts in one List? To be honest, it is mostly no mandatory. Having 100’000 Contacts in a List is a typical “Notes View” case. Most agile Application will not present you a tabular list with 100 results per page and a pager control to browse thru 1000 pages, that you can find the page with the correct dataset. But having such a large dataset helps to find the best solution to read a dataset.

So our test scenario is very easy:

  1. 100’000 Documents to transfer to Contact Object add to an ArrayList.
  2. The transfer from document to object is fix, no changes will make there, we try only to manipulate the way we get the 100’000 Documents and how we can add them to the ArrayList
  3. All tests are instrumented and measured with the XPages Toolbox (see Mastering XPages Second Edition, Chapter 20)

1. Using the current XPT Approach:

The XPagesToolkit uses currently a view and browse the document by getNextDocument(). Each object is added to the ArrayList(). This has the effect that the ArrayList has to extend and realign the internal array each time when we achieve the internal capacity. This is the base value.

2. Using LinkedList instead of ArrayList

The only change to the XPT Approach is, that we use a LinkedList instead of the ArrayList and copy the LinkedList as last opration to an ArrayList. We expect to see a significant speed boost, because the LinkedList can growth in a “one Operation” step and has no reassigning of an internal array.

3. Using a NoteCollection based on the view’s selection formula (suggested by Nathan T. Freeman)

Instead of looping thru a view, we use the views selection formula to build a NoteCollection and access the document directly by “getDocumentById()”. I was personally a bit sceptic about building a selection of NoteID ad hoc. Would this work in such a big dataset?

Each approach was tested 10 times (using the code at the end of the entry). Here the results:

XPT LinkedList NoteCollection
Averag Time (sec) 40.5 39.9 32.7
Min 39 39.2 31.7
Mas 43.8 40.8 33.8
Convesion doc to object (sec) 28.8 29 25.8
getNextDocument() (100’000 calls) (sec) 10.3 10.3
Copy LinkeList to ArrayList (ms) 54
Building NoteCollection (ms) 726
getDocumentByID (100’000 calls) (ms) 2294
getNextNotesID (100’000 calls) (ms) 580

My first finding was, that building a LinkedList and then converting to an ArrayList could be a bit faster than the permanent resizing of the ArrayList, but no so much as expected. That was also the reason to see what was happening in the backend. And as you see here, the time to loop thru the we is long, 10.3 sec, means 0.1ms per document.

But the big surprise was the result with the NoteCollection. It needs only 726ms to build the Collection and other 2.9 sec to browse thru the collection. That’s very fast. This means, even accessing the View to get the selection formula and then building the NoteCollection and browse thru this collection is much faster!

I think that will definitely go in to the XPagesToolkit. But this will also need testing. If you wanna be informed about new SNAPSHOT build of the XPagesTookit, please register here http://mm.wgcdev.ch/mailman/listinfo/xpagestoolkit!

And then, the child in me told me. How will this scale, if you set in the view selection formula a @Today function. We all know that such views has to be rebuilt permanently. And the child wants to play, so I did. Here the same stuff with a view with the following selection formula: SELECT ((Form = “Contact”)) & @Created > @Adjust(@Today;0;0;-10;0;0;0)

The XPT had an average time of 50 sec (that’s what I excepted to see), to get all the document (also 100’000). BUT……..

….. the NoteCollection approach has done this in 32.8 sec! Means near the same speed! The building of the NoteCollection has a time vor 844 ms. That’s definitely great!

Let see what this has for en impact for, small datasets. And a lot of stuff to discuss with the XPagesToolkit Team 🙂

 

Here my test code:

/*
 * © Copyright WebGate Consulting AG, 2014
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 * http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
 * implied. See the License for the specific language governing
 * permissions and limitations under the License.
 */
 package org.openntf.xpt.demo.bo;
import java.util.ArrayList;
 import java.util.LinkedList;
 import java.util.List;
import lotus.domino.Document;
 import lotus.domino.NotesException;
 import lotus.domino.View;
 import lotus.domino.NoteCollection;
import org.openntf.xpt.core.dss.AbstractStorageService;
 import org.openntf.xpt.core.dss.DominoStorageService;
import com.ibm.commons.util.StringUtil;
 import com.ibm.commons.util.profiler.Profiler;
 import com.ibm.commons.util.profiler.ProfilerAggregator;
 import com.ibm.commons.util.profiler.ProfilerType;
 import com.ibm.xsp.extlib.util.ExtLibUtil;
public class ContactStorageService extends AbstractStorageService {
private static ContactStorageService m_Service;
private static final ProfilerType pt = new ProfilerType("ListPerformance");
private ContactStorageService() {
}
public static ContactStorageService getInstance() {
 if (m_Service == null) {
 m_Service = new ContactStorageService();
 }
 return m_Service;
 }
@Override
 protected Contact createObject() {
 return new Contact();
 }
// Performance TestCode
 public List getAllXPT(String viewName) {
 List lstRC;
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "getAllXPT -> " + viewName);
 long startTime = Profiler.getCurrentTime();
 try {
 lstRC = getAll(viewName);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 lstRC = getAll(viewName);
 }
 return lstRC;
 }
public List getAllLinkedList() {
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "getAllLinkedList");
 long startTime = Profiler.getCurrentTime();
 try {
 return _getAllLinkedList();
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 return _getAllLinkedList();
 }
 }
public List<Contact> _getAllLinkedList() {
 List lstRC = new LinkedList();
 try {
 View viw = ExtLibUtil.getCurrentDatabase().getView("AllContacts");
 viw.setAutoUpdate(false);
 Document docNext = viw.getFirstDocument();
 while (docNext != null) {
 Document docProcess = docNext;
 docNext = viw.getNextDocument(docNext);
 Contact con = createObject();
 if (DominoStorageService.getInstance().getObjectWithDocument(con, docProcess)) {
 lstRC.add(con);
 }
 docProcess.recycle();
 }
 viw.recycle();
 } catch (Exception e) {
 e.printStackTrace();
 }
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "copyLinkedList");
 long startTime = Profiler.getCurrentTime();
 try {
 return _copyList(lstRC);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 return _copyList(lstRC);
 }
 }
private List _copyList(List linkedList) {
 return new ArrayList(linkedList);
 }
public List getAllNC(String viewName) {
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "getAllNC " + viewName);
 long startTime = Profiler.getCurrentTime();
 try {
 return _getAllNC(viewName);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 return _getAllNC(viewName);
 }
}
public List getAllNCLinkedList() {
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "getAllNC-LinkedList");
 long startTime = Profiler.getCurrentTime();
 try {
 return _getAllNCLL();
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 return _getAllNCLL();
 }
}
private List _getAllNC(String viewName) {
 List lstRC = new ArrayList();
 try {
 View viw = ExtLibUtil.getCurrentDatabase().getView(viewName);
 String formula = viw.getSelectionFormula();
 NoteCollection nc;
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "buildNoteCollection " + viewName);
 long startTime = Profiler.getCurrentTime();
 try {
 nc = _buildNoteCollection(formula);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 nc = _buildNoteCollection(formula);
 }
lstRC = new ArrayList(nc.getCount());
 String notesIDNext = nc.getFirstNoteID();
 while (!StringUtil.isEmpty(notesIDNext)) {
 Document processDoc = ExtLibUtil.getCurrentDatabase().getDocumentByID(notesIDNext);
 notesIDNext = nc.getNextNoteID(notesIDNext);
 if (processDoc.isValid()) {
 Contact con = createObject();
 if (DominoStorageService.getInstance().getObjectWithDocument(con, processDoc)) {
 lstRC.add(con);
 }
 processDoc.recycle();
 }
 }
 nc.recycle();
 viw.recycle();
 } catch (Exception e) {
 e.printStackTrace();
 }
 return lstRC;
 }
private List<Contact> _getAllNCLL() {
 List lstRC = new LinkedList();
 try {
 View viw = ExtLibUtil.getCurrentDatabase().getView("AllContacts");
 String formula = viw.getSelectionFormula();
 NoteCollection nc;
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "buildNoteCollection");
 long startTime = Profiler.getCurrentTime();
 try {
 nc = _buildNoteCollection(formula);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 nc = _buildNoteCollection(formula);
 }
 String notesIDNext = nc.getFirstNoteID();
 while (!StringUtil.isEmpty(notesIDNext)) {
 Document processDoc = ExtLibUtil.getCurrentDatabase().getDocumentByID(notesIDNext);
 notesIDNext = nc.getNextNoteID(notesIDNext);
 if (processDoc.isValid()) {
 Contact con = createObject();
 if (DominoStorageService.getInstance().getObjectWithDocument(con, processDoc)) {
 lstRC.add(con);
 }
 processDoc.recycle();
 }
 }
 nc.recycle();
 viw.recycle();
 } catch (Exception e) {
 e.printStackTrace();
 }
 if (Profiler.isEnabled()) {
 ProfilerAggregator pa = Profiler.startProfileBlock(pt, "copyLinkedList");
 long startTime = Profiler.getCurrentTime();
 try {
 return _copyList(lstRC);
 } finally {
 Profiler.endProfileBlock(pa, startTime);
 }
 } else {
 return _copyList(lstRC);
 }
 }
private NoteCollection _buildNoteCollection(String formula) throws NotesException {
 NoteCollection nc = ExtLibUtil.getCurrentDatabase().createNoteCollection(false);
 nc.setSelectDocuments(true);
 nc.setSelectionFormula(formula);
 nc.buildCollection();
 return nc;
 }
 }
 
4 Comments

Posted by on July 19, 2014 in Domino, Java, OpenNTF, XPages, XPT

 

Tags: ,

Automated building and testing of XPages Applications

You know that situation? You have done a roll out of a new version of an application and a short while after being in production, there is this call from an angry customer? A bug, fixed in a previous release did re appear in the new release!

How annoying is that. And you ask yourself DID THEY NOT TEST IT? So if you walk to the QA department and ask nicely about “Testing”, what answer would you get? I thinks that’s the moment when you start to think about “Automated Testing”. And automated testing, without an automated build of your application is nice, but only the half way!

To collect all the pieces to make this happen was a part of my mission to go to IBM Ireland and meet the XPages Team in Dublin. And a day after 3 days of hard work with them, it’s time to talk about automated build and testing of XPages application. First the story line and pre requirement:

  • You are working on a XPages Application.
  • The Application is available as on disk project in a SCM (Source Control Managementsysten)
  • You have build some selenium UI Tests

Each night or on every change to your application that is commited to the SCM, you want to have a automated build of the application and the selenium tests should run.

Thanks to the Headless Designer (see this Wiki about), you can create with a command line directly a new NSF. But there is more to do then only executing a command line. You want to be sure that the designer has all required plugins (like OpenNTF Essentials) during the build process. And all of this plugins should also be removed after a build. After the build you want to have the application moved to a test server and the selenium tests should be executed against this application. The whole build process is now covered by a maven plugin. You can see the progress on that on this GitHub project.

The headlessdesigner-maven-plugin covers now the following steps:

  1. Installing defined features to the designer
  2. Enabling defined features in the designer
  3. Build the application from on disk project
  4. Disabling defined features in the designer
  5. Uninstalling of the features
  6. Copy the build NSF to the target directory of the maven project

The following steps are planed for the next 2-3 weeks:

  • Distributing the NSF to the DominoServer
  • Building a archetype plugin to build per command line a new maven project
  • Thinking about distribution of the plugins to a official maven repository

 

 

 
1 Comment

Posted by on May 1, 2014 in IBM, Java, OpenNTF, Uncategorized, XPages

 

Tags: , ,

My Slides from Engage.ug

 
Leave a comment

Posted by on March 24, 2014 in Domino, OpenNTF, SocialBusiness, XPages

 

Tags:

Video

Message to the Participant of the German Developer Camp

A short message to the german developer camp (in german).

 
Leave a comment

Posted by on March 19, 2014 in OpenNTF, SocialBusiness, XPages

 

First Swiss XPages Developer Crowd Day

Yes, it was a huge success. The Event-Room of WebGate was full of XPages Developers from Switzerland. One of the reason was, that we had the pleasure to meet Martin Donnelly and Eamon Muldon from the IBM XPages Team. The other reason is the DNA of Domino / XPages Developers. It seems that collaboration and sharing of knowledge is a key factor in this community. So we could have an open discussion across the border and boundaries of company and personal interest. During this special moment we were all only XPages Developers with a passion for good Applications.

Is it the right time to establish an event similar to XPages and Beer, which is a success around the world? Yes we think so. I’m working with several people from the Swiss XPages Developer Community to make this happen. The open discussion and the sharing of knowledge and experience was in the past a success factor and will be in the future a success factor.

Stay tuned for a next update.

I’m currently on travel. I was this week in Geneva and will be next week at the engage.ug and the bccon.de. I would love to meet you there.

And I’m still living without my E-Mail Folders! My life is much easier.

See you!

 
1 Comment

Posted by on March 14, 2014 in OpenNTF, SocialBusiness, XPages

 

Tags: , ,