Cutting edge technologies - About Java, Lotus Notes and iPhonehttp://www.mindoo.de/web/blog.nsf/en-usThu, 26 Aug 2010 09:53:51 AM +0200 Thu, 26 Aug 2010 09:53:51 AM +0200 John Our Lotusphere 2010 slides: BP203 - Leverage the New Java APIs in IBM Lotus Notes 8.5.1! http://www.mindoo.de/web/blog.nsf/dx/25.01.2010004124KLEVSF.htm?opendocument&comments#2010-08-26095351SERBA8.htm Fri, 20 Aug 2010 05:16:05 PM +0200 Karsten Lehmann Fast Notes view reading via C API Part 2: How to read views 99.7% faster than Notes.jar Link } ), we only get 91.

You may try to create your own buffer to optimize the throughput. We haven't played with that, because we need a way to let the user stop the read operation. If we read the view in one single step, there no way to cancel the operation AFAIK.]]> Link } ), we only get 91.

You may try to create your own buffer to optimize the throughput. We haven't played with that, because we need a way to let the user stop the read operation. If we read the view in one single step, there no way to cancel the operation AFAIK.]]> http://www.mindoo.de/web/blog.nsf/dx/25.04.2010235255KLETNN.htm?opendocument&comments#20.08.2010171605SERKWL.htm Fri, 20 Aug 2010 04:21:25 PM +0200 Mikael Paul Fast Notes view reading via C API Part 2: How to read views 99.7% faster than Notes.jar
I try the 6 th parameters ( ReturnCount ) with the value 3FF for a total of 9000 entries, and the method return me 219 lines each times.]]>
I try the 6 th parameters ( ReturnCount ) with the value 3FF for a total of 9000 entries, and the method return me 219 lines each times.]]> http://www.mindoo.de/web/blog.nsf/dx/25.04.2010235255KLETNN.htm?opendocument&comments#20082010162125SERJUF.htm Mon, 9 Aug 2010 09:55:30 PM +0200 Karsten Lehmann XPages series #9: How to debug an XPages application
Thanks a lot for the link! I did not know the Notes.ini options

JavaEnableDebug=1

JavaDebugOptions=transport=dt_socket,server=y,suspend=y,address=*portnumber*

to activate the server JVM's debug mode. That's an even better solution for some situations.

Best regards,

Karsten]]>
Thanks a lot for the link! I did not know the Notes.ini options

JavaEnableDebug=1

JavaDebugOptions=transport=dt_socket,server=y,suspend=y,address=*portnumber*

to activate the server JVM's debug mode. That's an even better solution for some situations.

Best regards,

Karsten]]> http://www.mindoo.de/web/blog.nsf/dx/20.07.2010094952KLEB7R.htm?opendocument&comments#09.08.2010215530SERRD9.htm Mon, 9 Aug 2010 08:23:50 PM +0200 Matt Vargish XPages series #9: How to debug an XPages application
{ Link }]]>
{ Link }]]> http://www.mindoo.de/web/blog.nsf/dx/20.07.2010094952KLEB7R.htm?opendocument&comments#08092010082350PMSERPKX.htm Mon, 9 Aug 2010 08:22:31 PM +0200 Matt Vargish XPages series #9: How to debug an XPages application
I've just posted about my own experience trying to debug Java in DDE and linked here to your excellent content on XPages ({ Link } I was able to step into XPages on the server, a bit messy -- and perhaps you've tried the approach -- but I thought I'd mention it.

Regards, Matt]]>
I've just posted about my own experience trying to debug Java in DDE and linked here to your excellent content on XPages ({ Link } I was able to step into XPages on the server, a bit messy -- and perhaps you've tried the approach -- but I thought I'd mention it.

Regards, Matt]]> http://www.mindoo.de/web/blog.nsf/dx/20.07.2010094952KLEB7R.htm?opendocument&comments#08092010082231PMSERPK5.htm Tue, 3 Aug 2010 05:03:54 PM +0200 Olle Thalén XPages series #4: Backing Bean Management with XPages
I'll think try the first advice with nested repeaters as a start. I also read your article number 5 about detecting data changes in java, I think I'll try to use that approach if I want to detect changes a user has made to the UI component.

Thanks again,

Olle]]>
I'll think try the first advice with nested repeaters as a start. I also read your article number 5 about detecting data changes in java, I think I'll try to use that approach if I want to detect changes a user has made to the UI component.

Thanks again,

Olle]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#08032010050354PMSERKNZ.htm Sat, 31 Jul 2010 04:08:14 PM +0200 Karsten Lehmann XPages series #4: Backing Bean Management with XPages
For a completely flexible data table, you could try three things:

1. use two nested repeater controls

The outer repeater iterates over the rows of your data (e.g. a Notes view). You could implement your own Java List interface that lazily loads data from a Notes view when someone calls List.get(index). I did this in a project. The List implementation contains a fixed cache of entries, so for example, when you call List.get(0), it not only fetches the view entry in row 0, but the entries with index 0 to 50 to improve performance.

In your inner repeater, you can then render the code for the table columns. Each row of your outer List could return another List with the column values.

2. use a Dojo grid component and feed it with JSON

There are already some pretty nice components that can display data lists, so why not use one of these and let it read its data via Ajax call from your XPage as a JSON string? Stephan Wissel from IBM demonstrated in his blog how to return any kind of data in an XPage:

{ Link }

2. use a completely dynamic approach

For another project, I created something more flexible. By using the XPages extensibility APIs, you can develop your own controls. I created a tree control that visualizes a Swing tree model on the web. The whole XPages UI is created in backend Java code when the XPages component tree is build at page load.

All value bindings and method bindings for expand/collapse links and partial refreshs are created programmatically as well. You can get some hints on how to do that when you take a closer look at the OpenNTF project Medusa and I think there are also some articles in the Designer wiki.

A sample control for this technique could look like this:

public class MyControl extends UIComponentBase implements FacesComponent {

public MyControl() {

super();

}

public String getFamily() {

return "com.company.MyControl";

}

public void buildContents(FacesContext ctx, FacesComponentBuilder builder) throws FacesException {

Application application=ctx.getApplication();

List children=getChildren();

XspOutputText txt=new XspOutputText();

ValueBinding binding=application.createValueBinding("#{javascript:return \"Hello World\"}");

txt.setValueBinding("title", binding);

children.add(txt);

}

}

You see, you can create whatever you want in your XPages component tree.

But be aware that this approach can be very hard to debug.]]>
For a completely flexible data table, you could try three things:

1. use two nested repeater controls

The outer repeater iterates over the rows of your data (e.g. a Notes view). You could implement your own Java List interface that lazily loads data from a Notes view when someone calls List.get(index). I did this in a project. The List implementation contains a fixed cache of entries, so for example, when you call List.get(0), it not only fetches the view entry in row 0, but the entries with index 0 to 50 to improve performance.

In your inner repeater, you can then render the code for the table columns. Each row of your outer List could return another List with the column values.

2. use a Dojo grid component and feed it with JSON

There are already some pretty nice components that can display data lists, so why not use one of these and let it read its data via Ajax call from your XPage as a JSON string? Stephan Wissel from IBM demonstrated in his blog how to return any kind of data in an XPage:

{ Link }

2. use a completely dynamic approach

For another project, I created something more flexible. By using the XPages extensibility APIs, you can develop your own controls. I created a tree control that visualizes a Swing tree model on the web. The whole XPages UI is created in backend Java code when the XPages component tree is build at page load.

All value bindings and method bindings for expand/collapse links and partial refreshs are created programmatically as well. You can get some hints on how to do that when you take a closer look at the OpenNTF project Medusa and I think there are also some articles in the Designer wiki.

A sample control for this technique could look like this:

public class MyControl extends UIComponentBase implements FacesComponent {

public MyControl() {

super();

}

public String getFamily() {

return "com.company.MyControl";

}

public void buildContents(FacesContext ctx, FacesComponentBuilder builder) throws FacesException {

Application application=ctx.getApplication();

List children=getChildren();

XspOutputText txt=new XspOutputText();

ValueBinding binding=application.createValueBinding("#{javascript:return \"Hello World\"}");

txt.setValueBinding("title", binding);

children.add(txt);

}

}

You see, you can create whatever you want in your XPages component tree.

But be aware that this approach can be very hard to debug.]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#31.07.2010160814SERJL7.htm Sat, 31 Jul 2010 03:19:49 PM +0200 Olle Thalén XPages series #4: Backing Bean Management with XPages
I was thinking about one thing though, in your example the columns and their names are fixed. So it is difficult to swap between different views. I was thinking about creating a dynamic view control where you can set the view name as a parameter and then load the content of that view in a custom control. My idea was to use a backing bean to read which view the user wants to use, then using the standard domino api to access that view and populate a backing bean with data from that view.

But in order to achieve this you must be able to render the columns in the table in a dynamic way, maybe using a repeat control? And instead of mapping the column value to a string property like you do, e.g value = "#{currentActor.firstname}", I would use value = "#{myBean.columns[indexVar]}". Where indexVar is the counter of the repeat control and columns is a java.util.List in the java bean.

Do you think this would be possible? I know it is technically possible to access a list in a java bean with the above syntax in E.L, I'm not sure however if you can index one element in the list using the counter from the repeat control.

I've seen example of dynamic views before, but those solutions have been based on SSJS and not the JSF technology. I think it breaks the MVC design pattern when you use too much SSJS, and I would prefer to have all my buisiness logic in beans, and use E.L to access data in the UI.]]>
I was thinking about one thing though, in your example the columns and their names are fixed. So it is difficult to swap between different views. I was thinking about creating a dynamic view control where you can set the view name as a parameter and then load the content of that view in a custom control. My idea was to use a backing bean to read which view the user wants to use, then using the standard domino api to access that view and populate a backing bean with data from that view.

But in order to achieve this you must be able to render the columns in the table in a dynamic way, maybe using a repeat control? And instead of mapping the column value to a string property like you do, e.g value = "#{currentActor.firstname}", I would use value = "#{myBean.columns[indexVar]}". Where indexVar is the counter of the repeat control and columns is a java.util.List in the java bean.

Do you think this would be possible? I know it is technically possible to access a list in a java bean with the above syntax in E.L, I'm not sure however if you can index one element in the list using the counter from the repeat control.

I've seen example of dynamic views before, but those solutions have been based on SSJS and not the JSF technology. I think it breaks the MVC design pattern when you use too much SSJS, and I would prefer to have all my buisiness logic in beans, and use E.L to access data in the UI.]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#2010-07-31151949SERHMX.htm Fri, 23 Jul 2010 03:37:49 AM +0200 Stephan H. WIssel The View article: Add LSDoc Support to Domino Designer on Eclipse with Custom Plug-ins (download link included)
Suggestion: Put the installable on an Update site (OpenNTF could accommodate you there), so future updates are automatically deployed.

One idea for the presentation: if you do an XPages UI it will work in a Local Notes client too.]]>
Suggestion: Put the installable on an Update site (OpenNTF could accommodate you there), so future updates are automatically deployed.

One idea for the presentation: if you do an XPages UI it will work in a Local Notes client too.]]> http://www.mindoo.de/web/blog.nsf/dx/22.07.2010223240KLES4H.htm?opendocument&comments#07232010033749AMSER3X8.htm Wed, 21 Jul 2010 02:14:04 PM +0200 Stephan H. Wissel XPages series #9: How to debug an XPages application http://www.mindoo.de/web/blog.nsf/dx/20.07.2010094952KLEB7R.htm?opendocument&comments#07212010021404PMSERGCV.htm Thu, 15 Jul 2010 11:16:05 AM +0200 Karsten Lehmann XPages series #4: Backing Bean Management with XPages
In 8.5.1, you may have to tweak with the Java security settings to make it work in the client, but IBM has worked on that issue for 8.5.2.

See this blog posting for details on running XPages in the Notes client with code in Java beans:

{ Link }

It should work in an XPages app on a Domino server without security issues.]]>
In 8.5.1, you may have to tweak with the Java security settings to make it work in the client, but IBM has worked on that issue for 8.5.2.

See this blog posting for details on running XPages in the Notes client with code in Java beans:

{ Link }

It should work in an XPages app on a Domino server without security issues.]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#15.07.2010111605SERCVL.htm Thu, 15 Jul 2010 10:43:29 AM +0200 Fedor Belov XPages series #4: Backing Bean Management with XPages
I've got a crazy idea, pls comment it:

1. Java application server (tomcat) is responsible for background logic. It doesn't know about LN objects because we store content in SQL DB (mySQL, for example)

2. xPages is used to display/update content in LN Client from/to SQL DB (using beans)]]>
I've got a crazy idea, pls comment it:

1. Java application server (tomcat) is responsible for background logic. It doesn't know about LN objects because we store content in SQL DB (mySQL, for example)

2. xPages is used to display/update content in LN Client from/to SQL DB (using beans)]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#15.07.2010104329SERC98.htm Thu, 15 Jul 2010 02:25:14 AM +0200 Karsten Lehmann XPages series #4: Backing Bean Management with XPages
Regarding the memory leaks: Well, it depends on your code. Compared to executing an agent, using beans might have a higher risk of producing memory leaks, because the Notes agents are loaded with their own classloader and unloaded after their execution (which makes it slower to call the agent compared to call Java methods directly).

You should avoid using the singleton design pattern in your bean code, because all XPages code is executed in the same JVM instance and that JVM only shuts down when you shut down the whole server. By using singletons, the garbage collector is not able to release the memory allocated by your bean code.

A better approach is to store data and object references either in one of the scope maps (e.g. session scope). The Domino server flushes those maps after a timeout. I think the default session duration is 30 minutes.]]>
Regarding the memory leaks: Well, it depends on your code. Compared to executing an agent, using beans might have a higher risk of producing memory leaks, because the Notes agents are loaded with their own classloader and unloaded after their execution (which makes it slower to call the agent compared to call Java methods directly).

You should avoid using the singleton design pattern in your bean code, because all XPages code is executed in the same JVM instance and that JVM only shuts down when you shut down the whole server. By using singletons, the garbage collector is not able to release the memory allocated by your bean code.

A better approach is to store data and object references either in one of the scope maps (e.g. session scope). The Domino server flushes those maps after a timeout. I think the default session duration is 30 minutes.]]> http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#15.07.2010022514SER2HT.htm Wed, 14 Jul 2010 12:31:38 PM +0200 Fedor Belov XPages series #4: Backing Bean Management with XPages http://www.mindoo.de/web/blog.nsf/dx/16.07.2009095816KLEBCY.htm?opendocument&comments#14.07.2010123138SERECT.htm Wed, 30 Jun 2010 04:28:33 PM +0200 Karsten Lehmann Lotusphere announcement: LS2Eclipse - Leverage Eclipse/Expeditor features in your Lotusscript code!
at the moment it's temporarily suspended due to very high project workload. I'm planning to work on it again in about 2 months from now.

We had some trouble with the tool after Lotusphere, because we tried for a long time to use an alternative approach that we got from our fellows at Teamstudio, which would have been the best bridging solution between both worlds so far and supported on multiple platforms. Unfortunately, it turned out to be too instable to be used, so I focused again on two different approaches for Windows/Linux and Mac that I had created (As heavy Mac users, our goal is of course to be Mac compatible from the beginning)

The Mac client was a bit tricky, because threading is different than on Windows and we also worked with IBM on some issues in the client and waited long time whether they will be fixed in 8.5.2 to make our Windows approach also work on the Mac.

IBM now told us that they will look into the issues for 8.5.3.

So finally, the last time I worked on LS2Eclipse, I was unifying the code of my different approaches for Windows/Linux and Mac to make deployment easier.

Well, then a few projects crossed our road...

If you're interested in the solution, please feel free to send me an email.

Best regards,

Karsten]]>
at the moment it's temporarily suspended due to very high project workload. I'm planning to work on it again in about 2 months from now.

We had some trouble with the tool after Lotusphere, because we tried for a long time to use an alternative approach that we got from our fellows at Teamstudio, which would have been the best bridging solution between both worlds so far and supported on multiple platforms. Unfortunately, it turned out to be too instable to be used, so I focused again on two different approaches for Windows/Linux and Mac that I had created (As heavy Mac users, our goal is of course to be Mac compatible from the beginning)

The Mac client was a bit tricky, because threading is different than on Windows and we also worked with IBM on some issues in the client and waited long time whether they will be fixed in 8.5.2 to make our Windows approach also work on the Mac.

IBM now told us that they will look into the issues for 8.5.3.

So finally, the last time I worked on LS2Eclipse, I was unifying the code of my different approaches for Windows/Linux and Mac to make deployment easier.

Well, then a few projects crossed our road...

If you're interested in the solution, please feel free to send me an email.

Best regards,

Karsten]]> http://www.mindoo.de/web/blog.nsf/dx/11.01.2010135317KLEH5D.htm?opendocument&comments#30.06.2010162833SERJYW.htm Wed, 30 Jun 2010 04:05:32 PM +0200 Johnny Jiang XPages series #5: Helper classes to access the JSF environment and Lotus Notes data
Your blog has been really helpful for me, thanks a lot.

Johnny]]>
Your blog has been really helpful for me, thanks a lot.

Johnny]]> http://www.mindoo.de/web/blog.nsf/dx/18.07.2009191738KLENAL.htm?opendocument&comments#06302010040532PMSERJJH.htm Wed, 30 Jun 2010 04:04:04 PM +0200 Joerg S. Lotusphere announcement: LS2Eclipse - Leverage Eclipse/Expeditor features in your Lotusscript code!
what's the current state of your LS2Eclipse project?

Joerg]]>
what's the current state of your LS2Eclipse project?

Joerg]]> http://www.mindoo.de/web/blog.nsf/dx/11.01.2010135317KLEH5D.htm?opendocument&comments#06302010040404PMSERJHL.htm Sat, 8 May 2010 08:53:30 PM +0200 Bill Wheaton Fast Notes view reading via C API Part 2: How to read views 99.7% faster than Notes.jar http://www.mindoo.de/web/blog.nsf/dx/25.04.2010235255KLETNN.htm?opendocument&comments#05082010085330PMSERQ6H.htm Mon, 26 Apr 2010 16:41:27 +0200 Nathan T. Freeman Fast Notes view reading via C API Part 2: How to read views 99.7% faster than Notes.jar
4 overall loops: JNI ViewEntry walk (FastVE), notes.jar ViewEntry walk (SlowVE), JNI Document walk (FastDoc), notes.jar Document walk (SlowDoc).

Running against 4 identical NSFs, each with a single view that shows 15,000 documents.

Each test run starts 8 jobs of that loop styles, 2 jobs per NSF. The timing number is the average run time for each job.

Local NSF, FastVE, avg ms: 1700

Local NSF, SlowVE avg ms: 15000

Local NSF FastDoc, avg ms: 2000

Local NSF SlowDoc avg ms: 24000

Remote server: 100Mbps server about 30 feet from my desk.

Remote NSF, FastVE, avg ms: 6100

Remote NSF, SlowVE, avg ms: 132000

Remote NSF, FastDoc, avg ms: 6200

Remote NSF, SlowDoc avg ms: 261000

This is using a ReadEntry block size of 1000.

Remember, each average is determined by running 8 jobs with the same algorithm concurrently against 4 separate instances of an identical NSF.]]>
4 overall loops: JNI ViewEntry walk (FastVE), notes.jar ViewEntry walk (SlowVE), JNI Document walk (FastDoc), notes.jar Document walk (SlowDoc).

Running against 4 identical NSFs, each with a single view that shows 15,000 documents.

Each test run starts 8 jobs of that loop styles, 2 jobs per NSF. The timing number is the average run time for each job.

Local NSF, FastVE, avg ms: 1700

Local NSF, SlowVE avg ms: 15000

Local NSF FastDoc, avg ms: 2000

Local NSF SlowDoc avg ms: 24000

Remote server: 100Mbps server about 30 feet from my desk.

Remote NSF, FastVE, avg ms: 6100

Remote NSF, SlowVE, avg ms: 132000

Remote NSF, FastDoc, avg ms: 6200

Remote NSF, SlowDoc avg ms: 261000

This is using a ReadEntry block size of 1000.

Remember, each average is determined by running 8 jobs with the same algorithm concurrently against 4 separate instances of an identical NSF.]]> http://www.mindoo.de/web/blog.nsf/dx/25.04.2010235255KLETNN.htm?opendocument&comments#04262010044127PMSERK8Y.htm