I Know Some Guys

Tue Feb 26 13:39:33 EST 2013

I've been a bit quiet lately, but that's mostly because I've been pretty busy lately. After my old company began closing down, I started going whole-hog in my consulting company, I Know Some Guys. Naturally, I can't go TOO much into it, but the general gist is "so far, so good." We have a couple clients so far and they've been keeping me busy indeed. A lot of that has involved classic Notes client and web development, but I guess that serves me right for snickering at people working on large "modernization" projects. Nonetheless, if you're looking for someone to build an app for you, consult, or so forth, drop me a line.

One nice side effect of having so many tasks flying around early is that it's really giving the project-tracking database I built a trial by fire. I built it to be pretty similar to the one I made at my old place, but boiled down to just the necessary components. Putting it under heavy use has done a fantastic job figuring out which parts I don't really use, which parts I sorely need to add, and which UI decisions I made turned out to be... questionable. Moreover, we're using Freshbooks for our time tracking and billing, and I'm really itching to integrate with their API. I hear tell that Domino has some REST- and OAuth-consuming capabilities and this will finally give me a chance to put them to use in a real situation.

Now, with any luck, I'll have some time shortly to knock out a few posts I've been developing notes on. In the mean time, if you somehow haven't already, go read Tim Tripcony's post about SSJS and the ensuing comment discussion. I promise it's worth your time.

The Bean-Backed Table Design Pattern

Tue Jan 22 17:54:14 EST 2013

First off, I don't like the name of this that I came up with, but it'll have to do.

One of the design problems that comes up all the time in Notes/Domino development is the "arbitrary table" idea. In classic Notes, you could solve this with an embedded view, generated HTML (if you didn't want it to be good), or a fixed-size table with a bunch of hide-whens. With XPages, everything is much more flexible, but there's still the question of the actual implementation.

The route I've been taking lately involves xp:dataTables, Lists of Maps, MIMEBean, and controller classes. In a recent game-related database, I wanted to store a list of components that go into a recipe. There can be an arbitrary number of them (well, technically, 4 in the game, but "arbitrary" to the code) and each has just two fields: an item ID and a count. The simplified data table code looks like this:

<xp:dataTable value="#{pageController.componentInfo}" var="component" indexVar="componentIndex">
	<xp:column styleClass="addRemove">
		<xp:this.facets>
			<xp:div xp:key="footer">
				<xp:button value="+" id="addComponent">
					<xp:eventHandler event="onclick" submit="true" refreshMode="partial" refreshId="componentInfo"
						action="#{pageController.addComponent}"/>
				</xp:button>
			</xp:div>
		</xp:this.facets>
						
		<xp:button value="-" id="removeComponent">
			<xp:eventHandler event="onclick" submit="true" refreshMode="partial" refreshId="componentInfo"
				action="#{pageController.removeComponent}"/>
		</xp:button>
	</xp:column>
					
	<xp:column styleClass="itemName">
		<xp:this.facets><xp:text xp:key="header" value="Item"/></xp:this.facets>
						
		<xp:inputText value="#{component.ItemID}"/>
	</xp:column>
					
	<xp:column styleClass="count">
		<xp:this.facets><xp:text xp:key="header" value="Count"/></xp:this.facets>
		<xp:inputText id="inputText1" value="#{component.Count}" defaultValue="1">
			<xp:this.converter><xp:convertNumber type="number" integerOnly="true"/></xp:this.converter>
		</xp:inputText>
	</xp:column>
</xp:dataTable>

The two data columns are pretty normal - they could easily point to an array/List in a data context or a view/collection of documents. The specific implementation is that it's an ArrayList stored in the view scope - either created new or deserialized from the document as appropriate:

public void postNewDocument() {
	Map<String, Object> viewScope = ExtLibUtil.getViewScope();
	viewScope.put("ComponentInfo", new ArrayList<Map<String, String>>());
}

public void postOpenDocument() throws Exception {
	Map<String, Object> viewScope = ExtLibUtil.getViewScope();
	Document doc = this.getDoc().getDocument();
	viewScope.put("ComponentInfo", JSFUtil.restoreState(doc, "ComponentInfo"));
}

public List<Map<String, Serializable>> getComponentInfo() {
	return (List<Map<String, Serializable>>)ExtLibUtil.getViewScope().get("ComponentInfo");
}

Those two "addComponent" and "removeComponent" actions are very simple as well: they just fetch the list from the view scope and manipulate it:

public void addComponent() {
	this.getComponentInfo().add(new HashMap<String, Serializable>());
}
public void removeComponent() {
	int index = (Integer)ExtLibUtil.resolveVariable(FacesContext.getCurrentInstance(), "componentIndex");
	this.getComponentInfo().remove(index);
}

One key part to notice is that the "componentIndex" variable is indeed the value you'd want. In the context of the clicked button (say, in the second row of the table), the componentIndex variable is set to 1, and the resolver and FacesContext make that work.

On save, I just grab the modified Document object, serialize the object back into it, and save the data source (that's what super.save() from my superclass does):

public String save() throws Exception {
	List<Map<String, Serializable>> componentInfo = this.getComponentInfo();

	Document doc = this.getDoc().getDocument(true);
	JSFUtil.saveState((Serializable)componentInfo, doc, "ComponentInfo");

	return super.save();
}

The end result is that I have a user-expandable/collapsible table that can hold arbitrary fields (since it's a Map - it could just as easily be a list of any other serializable objects):

If I wanted to access the data from non-Java contexts, I could replace the MIMEBean bits with another method, like a text-list-based format or response documents, but the final visual result (and the XSP code) would be unchanged.

More On "Controller" Classes

Mon Jan 07 19:39:09 EST 2013

Tags: xpages mvc java
  1. Dec 26 2012 - "Controller" Classes Have Been Helping Me Greatly
  2. Jan 07 2013 - More On "Controller" Classes

Since my last post on the matter , I've been using this "controller" class organization method in a couple other projects (including a refresh of the back-end of this blog), and it's proven to be a pretty great way to go about XPages development.

As I mentioned before, the "controller" term comes from the rough equivalent in Rails. Rails is thoroughly MVC based, so a lot of your programming involves creating the UI of a page in HTML with a small sprinkling of Ruby (the "view"), backed by a class that is conceptually tied to it and stores all of the real business logic associated with that specific page (the "controller"). XPages don't have quite this same assumption built in, but the notion of pairing an XPage with a single backing Java class is a solid one. Alternatively, you can think of the "controller" class as being like a stylesheet or client JavaScript file: the HTML page handles the design, and only contains references to functionality defined elsewhere.

What this means in practice is that I've been pushing to eliminate all traces of non-EL bindings in my XSP markup in favor of writing the code in the associated Java class - this includes not only standard page events like beforeRenderResponse , but also anywhere else that Server JavaScript (or Ruby) would normally appear, like value and method bindings. Here's a simple example, from an XPage named Test and its backing class:

Test.xsp
<p><xp:button id="clickMe" value="Click Me">
	<xp:eventHandler event="onclick" submit="true" refreshMode="partial" refreshId="output"
		action="#{pageController.refreshOutput}"/>
</xp:button></p>
	
<p><xp:text id="output" value="#{pageController.output}"/></p>
controller/Test.java
package controller;

import frostillicus.controller.BasicXPageController;
import java.util.Date;

public class Test extends BasicXPageController {
	private static final long serialVersionUID = 1L;

	private String output = "default";
	public String getOutput() { return this.output; }

	public void refreshOutput() {
		this.output = new Date().toString();
	}
}

In a simple case like this, it doesn't buy you much, but imagine a more complicated case, say code to handle post-save document processing and notifications, or complicated code to generate the value for a container control. The separation between the two pays significant dividends when you stop having to worry scrolling through reams of code in the XSP markup when you're working on the page layout, while having a known location for the Java code associated with a page makes it easier to track down functionality. Plus, Server JavaScript is only mildly more expressive than Java, so even the business logic itself stays just about as clean (moving from Ruby to Java, on the other hand, imposes a grotesque LOC penalty).

It sounds strange, but the most important benefit I've derived from this hasn't been from performance or cleanliness (though those are great), but instead the discipline it provides. Now, there's no question where business logic associated with a page should go: in a class that implements XPageController with the same name as the page and put in the "controller" package. If I want pages to share functionality, that's handled either via a common method stored in another class or, as appropriate, via class inheritance. No inline code, no Server JavaScript script libraries. If something is going to be calculated, it's done in Java.

There is one area that's given me a bit of trouble: custom controls. For example, the linksbar on this blog requires computation to generate, but it's not associated with any specific page. For now, I put that in the top-level controller class, but that feels a bit wrong. The same solution also doesn't apply to the code that handles rendering each individual post in the list, which may be on the Home page, the Month page, or individually on the Post page, in which case it also has edit/save/delete actions associated with it. For now, I created a "helper" class that I instantiate (yes, with JavaScript, which is depressing) for each instance. I think the "right" way to do it with the architecture is to create my controls entirely in Java, with renderers and all. That would allow me to handle the properties passed in cleanly, but the code involved with creating those things is ugly as sin. Still, I'll give that a shot later to see if it's worth the tradeoff.

"Controller" Classes Have Been Helping Me Greatly

Wed Dec 26 16:54:16 EST 2012

Tags: xpages mvc java
  1. Dec 26 2012 - "Controller" Classes Have Been Helping Me Greatly
  2. Jan 07 2013 - More On "Controller" Classes

I mentioned a while ago that I've been using "controller"-type classes paired with specific XPages to make my code cleaner. They're not really controllers since they don't actually handle any server direction or page loading, but they do still hook into page events in a way somewhat similar to Rails controller classes. The basic idea is that each XPage gets an object to "back" it - I tie page events like beforePageLoad and afterRenderResponse to methods on the class that implements a standard interface:

package frostillicus.controller;

import java.io.Serializable;
import javax.faces.event.PhaseEvent;

public interface XPageController extends Serializable {
	public void beforePageLoad() throws Exception;
	public void afterPageLoad() throws Exception;

	public void afterRestoreView(PhaseEvent event) throws Exception;

	public void beforeRenderResponse(PhaseEvent event) throws Exception;
	public void afterRenderResponse(PhaseEvent event) throws Exception;
}

I have a basic stub class to implement that as well as an abstract class for "document-based" pages:

package frostillicus.controller;

import iksg.JSFUtil;

import javax.faces.context.FacesContext;

import com.ibm.xsp.extlib.util.ExtLibUtil;
import com.ibm.xsp.model.domino.wrapped.DominoDocument;

public class BasicDocumentController extends BasicXPageController implements DocumentController {
	private static final long serialVersionUID = 1L;

	public void queryNewDocument() throws Exception { }
	public void postNewDocument() throws Exception { }
	public void queryOpenDocument() throws Exception { }
	public void postOpenDocument() throws Exception { }
	public void querySaveDocument() throws Exception { }
	public void postSaveDocument() throws Exception { }

	public String save() throws Exception {
		DominoDocument doc = this.getDoc();
		boolean isNewNote = doc.isNewNote();
		if(doc.save()) {
			JSFUtil.addMessage("confirmation", doc.getValue("Form") + " " + (isNewNote ? "created" : "updated") + " successfully.");
			return "xsp-success";
		} else {
			JSFUtil.addMessage("error", "Save failed");
			return "xsp-failure";
		}
	}
	public String cancel() throws Exception {
		return "xsp-cancel";
	}
	public String delete() throws Exception {
		DominoDocument doc = this.getDoc();
		String formName = (String)doc.getValue("Form");
		doc.getDocument(true).remove(true);
		JSFUtil.addMessage("confirmation", formName + " deleted.");
		return "xsp-success";
	}

	public String getDocumentId() {
		try {
			return this.getDoc().getDocument().getUniversalID();
		} catch(Exception e) { return ""; }
	}

	public boolean isEditable() { return this.getDoc().isEditable(); }

	protected DominoDocument getDoc() {
		return (DominoDocument)ExtLibUtil.resolveVariable(FacesContext.getCurrentInstance(), "doc");
	}
}

I took it all one step further in the direction "convention over configuration" as well: I created a ViewHandler that looks for a class in the "controller" package with the same name as the current page's Java class (e.g. "/Some_Page.xsp" → "controller.Some_Page") - if it finds one, it instantiates it; otherwise, it uses the basic stub implementation. Once it has the class created, it plunks it into the viewScope and creates some MethodBindings to tie beforeRenderResponse, afterRenderResponse, and afterRestoreView to the object without having to have that code in the XPage (it proved necessary to still include code in the XPage for the before/after page-load and document-related events).

So on its own, the setup I have above doesn't necessarily buy you much. You save a bit of repetitive code for standard CRUD pages when using the document-controller class, but that's about it. The real value for me so far has been a clarification of what goes where. Previously, if I wanted to run some code on page load, or attached to a button, or to set a viewScope value, or so forth, it could go anywhere: in a page event, in a button action, in a dataContext, in a this.value property for a xp:repeat, or any number of other places. Now, if I want to evaluate something, it's going to happen in one place: the controller class. So if I have a bit of information that needs recalculating (say, the total cost of a shopping cart), I make a getTotalCost() method on the controller and set the value in the XPage to pageController.totalCost. Similarly, if I need to set some special values on a document on load or save, I have a clear, standard way to do it:

package controller;

import com.ibm.xsp.model.domino.wrapped.DominoDocument;
import frostillicus.controller.BasicDocumentController;

public class Projects_Contact extends BasicDocumentController {
	private static final long serialVersionUID = 1L;

	@Override
	public String save() throws Exception {
		DominoDocument doc = this.getDoc();
		doc.setValue("FullName", ("CN=" + doc.getValue("FirstName") + " " + doc.getValue("LastName")).trim() + "/O=IKSGClients");

		return super.save();
	}

	@Override
	public void postNewDocument() throws Exception {
		super.postNewDocument();

		DominoDocument doc = this.getDoc();
		doc.setValue("Type", "Person");
		doc.setValue("MailSystem", "5");
	}
}

It sounds like a small thing - after all, who cares if you have some SSJS in the postNewDocument event on an XPage? And, besides, isn't that where it's supposed to go? Well, sure, when you only have a small amount of code, doing it inline works fine. However, as I've been running into for a while, the flexibility of XPages makes them particularly vulnerable to becoming tangled blobs of repeated and messy code. By creating a clean, strict system from the start, I've made it so that the separation is never muddied: the XPage is about how things appear and the controller class is about how those things get to the XPage in the first place.

We'll see how it holds up as I use it more, but so far this "controller"-class method strikes a good balance between code cleanliness without getting too crazy on the backing framework (as opposed so some of the "model" systems I've tried making).

Moving From Ruby-in-XPages to Polyglot

Wed Dec 12 14:47:28 EST 2012

Tags: ruby polyglot

I've tweeted about this a couple times, but in some of my spare time lately I've put together a new, cleaner implementation of the generic-scripting-language support I first created back when I did my original work with Ruby. I named the new project "Polyglot" and it has two facets:

  1. The aforementioned generic-scripting-language support, done more cleanly and with better capabilities.
  2. A method for storing standalone page-generating scripts as documents in the database that are executed in the context of an XPage.
Generic Scripting Languages

With the usual when-I-get-around-to-it caveat, I plan to use Polyglot to entirely supplant Ruby-in-XPages, since its JSR-223-based features cover JRuby as well. The main thing I need to check on is proper memory usage - the embedding mechanism I originally used provides clear control over the lifetime of the scripting runtime, while JSR 223 does not. That may be fine, since it may properly manage itself, but it makes me a bit nervous.

I'll also have to work out proper context variable access (e.g. "database", "currentDocument", etc.). Some languages (Ruby, namely) seem to have gained a magic power since last I checked to access those variables without having to add special support, but I need to figure out if others work the same way or if I need to try an idea for an adapter I had.

In addition to the JSR 223 languages, I also just today added support for formula language via #{formula: ... } bindings, since I realized that that was exceedingly easy to do. For that one, I didn't bother binding it to the contextual variables, but instead just have it look for "currentDocument" - if it's present, it passes the lotus.domino.Document from that to the session.evaluate(...) call; if it's not, it forgoes the second parameter. I don't expect I'll use that binding much, but it was fun to add.

Standalone Scripts

The standalone script support is a little different - the idea is that, rather than working another language into an XPage, you write a script entirely in the other language, much like you might do with, say, PHP (don't use PHP). You can write a script using a basic editor, it's stored in a document in the database, and then you can execute it in the context of a ScriptRunner XPage, meaning you have the same Java environment as usual (with the same caveat as above that I need to sort out variable access).

Besides the usual suite of JSR 223 languages, I've also started working on integrating the XSP parser from the XPages Bazaar, both as a way to familiarize myself with the Bazaar and to see if there's any use in a setup where your XSP markup is stored in documents and evaluated at runtime rather than design elements compiled in Designer. Maybe, maybe not, but it's a fun test.

 

For now, it's still in an "experimental" state, but eventually I hope to cobble it into a proper releasable state and have it supersede Ruby-in-XPages entirely.

Programming Tips: Implied Booleans and the Ternary Operator

Tue Dec 11 08:40:52 EST 2012

Tags: programming

This isn't Domino- or XPages-specific per se, but I figured I'd make a post about some of my favorite stylistic bits in many programming languages: implied boolean values and the ternary operator, which are distinct but often used together.

What I mean by "implied boolean values" is when a language lets you use things that aren't strictly comparisons or boolean values in if/then tests. The best example for this is probably C (and no doubt the languages that came before it): for a long time, C didn't actually have a boolean data type. Instead, the convention was to use 1 for "true" and 0 for "false" - and many libraries would map constants to those values. That's well enough on its own, but the fun part was that the real rule behind it is that C treats ANY non-zero numeric value as true. Most of the time, that doesn't matter too much, but there are occasions where you can take advantage of that property to save a bit of code without sacrificing readability.

Many languages, such as JavaScript, continue this tradition, and also allow for implied null checking this way (I can't be bothered to look into C's deal with nulls). To demonstrate:

var foo = null
if(foo) { alert("foo") } else { alert("not foo") }
if(!foo) { alert("not foo") } else { alert("foo") }

That code will display "not foo" twice. JavaScript acts the same way for "undefined" values, such as uninitialized object properties:

var foo = { bar: 1 }
if(foo.bar) { alert("yep") } else { alert("nope") }
if(foo.baz) { alert("yep") } else { alert("nope") }

That will display "yep" followed by "nope". This can be handy if you want to test for, say, the presence of an attribute value in an HTML entity (it also treats "" as false).

Though JavaScript and, by extension, SSJS support thing kind of thing, Java is stricter about this kind of thing, forcing you to specifically test for null.

That brings me to the ternary operator, which is one of my favorite things in the world. If you're not familiar with it, the ternary operator is a shorthand way to do an if/then test that results in a value. As a demonstration:

var result = a > b ? option1 : option2

While you can get the same job done with a normal if/then/else block, that operator saves you some typing and, more importantly, often makes it easier to understand your intent. They can even be chained together, though you should be careful not to create too much of a mess:

var result = a == 1 ? option1 :
             a == 2 ? option2 :
             a == 3 ? option3 :
                      option4

In this case, Java DOES support the operator in pretty much the same way as every other non-PHP language (including Formula - @If(...) is basically this in different garb). EL in XPages supports this too, but only in ${}-bound values (due to Designer complaining about "invalid" syntax in #{}-bound values... presumably, you could write your own value bindings programmatically).

I use this kind of thing all the time. One pretty common use is to provide default values, either by checking the current state of a variable and re-assigning it as necessary or looking in, say, an options for overrides:

title = title ? title : "default title"
someVal = options.someVal ? options.someVal : "default someVal"

If you want to be EXTRA slick, you can take advantage of an extra property of JavaScript's "or" operator:

title = title || "default title"

That basically means "use the value of title unless it's null/false/0, in which case use 'default title'".

Hopefully, you can put some of these tricks to good use. Used properly, this kind of thing can drastically simplify your code while making it much easier to read and understand.

Putting Apache in Front of Domino

Sat Dec 08 13:59:11 EST 2012

Tags: apache admin
  1. Dec 08 2012 - Putting Apache in Front of Domino
  2. May 30 2013 - Better Living Through Reverse Proxies
  3. Nov 01 2015 - Domino's Server-Side User Security
  4. Jan 30 2021 - A Partially-Successful Venture Into Improving Reverse Proxies With Domino
  5. Jan 19 2022 - PSA: Reverse-Proxy Regression in Domino 12.0.1

The other day, for my side-project company, I wanted to set up hosting for a WordPress site, ideally without setting up another whole server. The first two ideas I had were pretty terrible:

  1. Hosting it on Domino directly with PHP via CGI. Even if this worked, I assume the performance would be pretty poor and I'd have no confidence in its general longevity.
  2. Hosting it on Apache on another port and using Domino to proxy through. While Domino does have some minor proxy capabilities, they didn't strike me as particularly thorough or suitable for the task.

Since the second option involved running two web servers anyway, I decided to flip it around and do the more-common thing: making Apache the main server and switching Domino to another port. Fortunately, even though it's been years since I ran an Apache server and I'd consider myself novice at best, the process has been exceptionally smooth and has brought with it a couple benefits:

  1. Apache's virtual-host support works perfectly, allowing me to host just the one named site and pass through all other requests to Domino.
  2. My crummy SSL setup works better with Apache, allowing for a poor-man's multi-host-name SSL with my one basic StartSSL certificate. Not only does Apache support SNI for down the line, but in the mean time I can use the same certificate for multiple names (with the usual "mis-matched name" warning) - since Apache handles all requests and funnels them over with the host name to Domino via HTTP, I don't run into the usual Domino problem of only the one SSL-enabled Web Site document being active.
  3. I'm now ready to add load-balancing servers at any time with just a one-line adjustment in my Apache config.

The actual configuration of Apache on my Linux machine was pretty straightforward, with the layout of the configuration directory making it fairly self-explanatory. I linked the modules I wanted from the /etc/apache2/mods-available directory to /etc/apache2/mods-enabled (namely, proxy, proxy_balancer, proxy_http, php5, rewrite, and ssl). Then, I set up a couple NameVirtualHost lines in ports.conf:

NameVirtualHost *:80
NameVirtualHost *:443

Then, I set up a new default site in /etc/apache2/sites-available and linked it to /etc/apache2/sites-enabled/000-domino:

<VirtualHost *:80>
        <Proxy balancer://frostillicus-cluster>
                BalancerMember http://ceres.frostillic.us:8088
                ProxySet stickysession=SessionID
        </Proxy>
        ProxyPass / balancer://frostillicus-cluster/ nocanon
        ProxyPassReverse / balancer://frostillicus-cluster/
        ProxyPreserveHost On
        AllowEncodedSlashes On
</VirtualHost>

That last directive is an important note, and I missed it at first. The "optimize CSS and JS" option in 8.5.3 creates URLs with encoded slashes and, by default, Apache's proxy breaks them, leading to 404 errors in apps that use it. If you turn on AllowEncodedSlashes, though, all is well. Note also the ProxySet line: if that's working correctly (I haven't tested it yet since I don't have a second host set up), that should make sure that browser sessions stick to the same server.

For SSL, I'm not sure what the absolute best way to do it is, but I set it up as another proxy just pointing to the HTTP version locally, so I don't have to set up multiple SSL sites for each virtual host (put into a new site document, 002-ssl):

<VirtualHost *:443>
        ProxyPass / http://127.0.0.1/
        ProxyPassReverse / http://127.0.0.1/
        ProxyPreserveHost On

        SSLEngine On
        SSLProtocol all -SSLv2
        SSLCipherSuite ALL:!ADH:!EXPORT:!SSLv2:RC4+RSA:+HIGH:+MEDIUM

        SSLCertificateFile /path/to/ssl.crt
        SSLCertificateKeyFile /path/to/ssl-decrypt.key
        SSLCertificateChainFile /path/to/sub.class1.server.ca.pem
        SSLCACertificateFile /path/to/ca.pem
</VirtualHost>

With that config, SSL seems to work exactly like I want: all my sites have an SSL counterpart that acts just like the normal one, much like with Domino normally.

It's only been running like this a couple days, so we'll see if I run into any more trouble, but so far this seems to be a solid win for my hosting, other than the increase in memory usage. I'm looking forward to having another clustermate in the same location so I can try out the load balancing down the line.

A Couple Blog Matters

Mon Nov 26 18:58:02 EST 2012

I've made a couple structural changes to the blog here. Normally, I wouldn't make a fuss over things like that, but they're pertinent to the overall theme.

First off, I changed the links over on the right, adding a link to the in-progress home page of my side-project company, I Know Some Guys. We're in the market for contracts - web site projects, internal apps, and the like. I may have more about that later.

I also promoted the link to my portfolio page to its own category. Since the aforementioned side-project company isn't a full-time gig, it can't hurt to keep that and the resume visible.

Finally, the blog itself is hopping on the ever-popular Bootstrap bandwagon. For the most part, the look of the site should be the same, though some widgets here and there are an awkward mix of the two. Many of my projects lately have been Bootstrap-based, and I figured I may as well move this over too, but not just leave it at the default style. Not only is the site (somewhat) more responsive now, but it and the portfolio have been giving me practice with building controls and Themes to make the application of Bootstrap to XPages apps easier. One huge advantage of OneUI is that IBM did a lot of work theming standard controls and building excellent new ones in the Extension Library to automatically gain OneUI styling as appropriate. Fortunately, I can piggyback on that: with a bit of work with Themes and Renderers, my goal is to make it so that an app can be almost-seamlessly switched between OneUI and Bootstrap as desired. They're not perfect matches for each other (OneUI has more structural geegaws), but they're close. That'll give me some much-needed Renderer practice, too.

My Current Data-Source Musings

Mon Nov 26 18:41:48 EST 2012

Tags: mvc

My quest to find the ideal way to access data in XPages continues. Each time I make a new project, I come up with a slight variation on my previous theme, improving in one or more of the various conflicting factors involved. The problem, as always, is that the standard xp:dominoView and xp:dominoDocument data sources alone are insufficient to properly separate business logic from presentation. On the other hand, accessing view data in a way that is both efficient and flexible via just the standard Java API is a non-trivial prospect. The advantage to using xp:dominoView in particular is that IBM already did an incredible amount of work making performance passable when dealing with cases of resorting, key-lookups, FT searches, response hierarchies, and the like.

The ideal method would build upon the efficient-access and serialization work already done while retaining that Rails-like collection-fetching and relation code from my forum attempts. Accordingly, I'm trying something a bit different with my latest project: I extended xp:dominoView to be able to take another parameter, modelClass, that is the class name of the objects it should return, rather than just DominoViewEntry. In turn, the objects I'm using wrap the real DominoViewEntry and allow for transparent access to the document (as needed, e.g. when a requested field is not a column) and the usual benefits of writing Java classes: namely, I can write code to handle field-change side effects without having to care about the logic behind that in the XPage itself.

This is sort of a half-formed idea so far, and I'm not 100% comfortable with it, but I think it's promising. Ideally, there wouldn't be any data sources written into the XPage itself at all (since it should be just a View), but that's sort of a necessity given the current state of things. This should be a reasonable next-best thing, and maybe a stepping stone to a new, faster, and cleaner data-access method.

The end goal of all this is to make it so that creating a new data "type" is very straightforward but still more declarative and structured than just making a new form name in the xp:dominoDocument data source. There should be classes to go with each type, but those classes should be very simple and easy to write. The quest continues!

The Ruby Builder for XPages

Wed Nov 07 13:50:14 EST 2012

Tags: xpages ruby

After I got Ruby in XPages to the point where it's generally working enough to power this blog, I set my sights on an even-more-important goal: being able to write backing Java classes in Ruby. While replacing SSJS is quite handy, my general use of inline scripting like that has declined significantly in favor of Java classes.

Fortunately, JRuby has a language cross-"compiler" and some hooks to write Java-compatible Ruby classes. Unfortunately, I was repeatedly stymied by a couple things:

  • Having the conversion happen automatically
  • Having the resultant "compiled" classes use the right class loader, allowing them to access other classes in the app
  • Implementing Java Interfaces

The first one was the most work but also the first to be solved: I wrote an Eclipse builder. The latter two were tough until I realized yesterday that the answers were sitting under my nose the whole time. For the classloader, I was able to switch the compiled output from using the global JRuby runtime to a runtime stored in the application scope and set to use the loader from facesContext.getContextClassLoader(). For the interfaces, it turned out that JRuby already had another annotation for declaring implemented interfaces, unsurprisingly called "java_implements".

So the upshot of this is: now I can write classes in Ruby and have Designer automatically convert them to Java and then compile those classes, ready to be used like any "normal" Java class in an XPage app. For demonstration purposes, I whipped up a useless-in-reality class to implement DataObject:

require "java"

java_package "frostillicus"

class TestDataObject
	java_implements "com.ibm.xsp.model.DataObject"
	
	
	def initialize
		@values = {}
	end
	
	java_signature "Object getValue(Object key)"
	def [](key)
		@values[key] or "#{key.to_s} not found"
	end
	
	java_signature "void setValue(Object key, Object value)"
	def []=(key, value)
		@values[key] = value
	end
	
	java_signature "Class<?> getType(Object key)"
	def get_type(key)
		java.lang.Class.class
	end
	
	java_signature "boolean isReadOnly(Object key)"
	def read_only?(key)
		false
	end
end

Once you have that file, it gets automatically converted to Java and compiled (once you have your build path set up correctly), and then it's available for use in other Java classes, in SSJS (or Ruby-in-XPages), and as a managed bean.

As you might expect with something like this, particularly for a first draft, there are a number of caveats:

  • The implementation is ugly as sin. There's a part that actually contains Java code that writes Ruby code that writes Java. It's very much a product of a series of "I wonder if this would work..." tests.
  • I haven't tested it to look for weird conflicts or leaks, though I suspect that my use of application-specific runtimes will help head off the worst of those potential problems.
  • You can't write classes that extend other classes. While I think you can do this within JRuby itself, the resultant objects are children of RubyObject, and there's no multiple inheritance. Interfaces will work in many cases, but that still limits the applicability.
  • Since the building happens in Designer, you need to have the feature installed client-side, which is not a problem with Ruby-in-XPages.
  • All those Java annotations really harsh the buzz of writing Ruby code. In that example above, the Ruby class isn't really any cleaner than a pure-Java equivalent. However, for larger, more complex classes, the expressiveness benefits of Ruby  would start to show.
  • There's basically no IDE help to be had. I haven't written anything into the builder to highlight syntax errors yet and non-syntax errors (like implementing an Interface but not all the methods) only show up with build errors on the resultant Java class. Plus, there's no autocomplete, and you only get syntax highlighting if you specifically install Ruby syntax support (I think I got it from the Eclipse Ganymede update site). Some of those may be fixable with minor effort, but the really big stuff would go far beyond the amount of time I have to dedicate to this kind of thing.

For now, I tossed the first draft up on GitHub for the curious. It will take more work to be be considered anywhere near finished, but getting it this far means I can start using it in personal/demo apps and really start finding the rough edges in practical use.