Welcome!

Java IoT Authors: Elizabeth White, Matt Davis, Yeshim Deniz, Liz McMillan, Sematext Blog

Related Topics: Java IoT

Java IoT: Article

Exception Chaining Simplifies Debugging

Exception Chaining Simplifies Debugging

Exception chaining (also known as "nesting exceptions"), is a technique for handling exceptions. A list is built of all the exceptions thrown as a result of a single originating exception as it's converted from lower to higher levels of abstraction. It can be used in both client and server environments to greatly simplify software debugging without adding undue complexity. This article discusses good exception-handling techniques and shows how to implement and use exception chaining.

Exceptions Are Part of the Interface
When our professors taught us about encapsulation and modular programming, we were told that modules should hide the details of their implementation. For example, an employee lookup service shouldn't advertise how or where it finds employees; only the methods required to access this service should be publicly available. This reduces complexity by isolating the implementation of one component from others that use it.

Most programmers realize that exceptions form an important part of the interface of any class, but aren't sure what to do with low-level exceptions that can't be handled directly by their code. Throwing a low-level exception from a higher-level class isn't a good idea because it exposes details of that class's implementation. The correct solution is for the class to catch the low-level exceptions, and to rethrow exceptions of a higher level of abstraction. If the getEmployee() method retrieves employee objects from a database via JDBC, for instance, a SQLException might be caught inside the method. This exception would be converted to an EmployeeLookupException and rethrown to the calling method.

Converting an Exception Object
There's no way, of course, to simply change the type of an exception. While it might be argued that we're converting from a more to a less specific exception, we can't cast one to the other because they don't have a parentÐchild relationship. For example, SQLException certainly doesn't extend from EmployeeLookupException or vice versa.

In this article "converting" an exception means to catch one exception type and throw a brand new exception of a different class. It's a conversion in the sense that the new exception is a direct result of the earlier one. When the EmployeeLookupException is created and thrown, it doesn't mean that a second error has occurred; it's just new packaging to represent the original error at a higher level of abstraction.

try {
stmt.execute(sql);
} catch (SQLException ex) {
throw new
EmployeeLookupException();
}

It quickly becomes obvious to many programmers that exception conversion (without chaining) has a serious deficiency: the root cause of the exception is lost. Is the EmployeeLookupException a result of a logon failure or a SQL query error? Exception conversion can make debugging more difficult.

I've argued in the paragraphs above that details of an implementation should be hidden from the user of the method, but we don't want to make debugging more difficult than necessary. The distinction to be made is that the internal implementation details should be hidden from the public API used at compile time, but this isn't a reason to obscure the cause and location of internal exceptions at runtime. We want to make these easy to discern. Correct exception handling with exception chaining achieves both these goals.

Upcasting and Other Poor Exception- Handling Techniques
One common solution to the problem of losing information with exception conversion is to convert through upcasting (i.e., casting "up" the tree to a superclass). Usually, the up-cast is to Exception and is programmed in an implicit manner by simply declaring that each method throws Exception. This undermines the intent of requiring exceptions to be caught in Java. The compiler doesn't require the programmer to catch any exceptions by this method.

public Employee getEmployee()
throws Exception
{
... database query code ...
}

While proper exception conversion adds context, making it more specific to the problem at hand, upcasting makes it more difficult to determine how to handle the exceptions. The desired exception conversions are from lower to higher levels of abstraction. This rarely matches the inheritance hierarchy of exceptions (which changes in amount of detail rather than abstraction - a very subtle difference). Remember that the abstraction a programmer seeks in designing a class should relate to the problem area he or she is trying to solve. For example, Exception is more generic than SQLException (a subclass of Exception), but it's probably not a higher level of abstraction for an employee lookup service. EmployeeLookupException and NoPermissionException are good examples of a higher level of abstraction in this case.

An even worse strategy for exception handling is to catch and log each exception (or not) before ignoring it. As long as the exception is logged, debugging is relatively straightforward. Unfortunately, this leaves no way for the calling method to detect or handle the error, which again eliminates the value of having exceptions.

What Is an Exception Chain?
An exception chain is a list of all the exceptions generated in response to a single root exception (say, a SQLException). As each exception is caught and converted to a higher-level exception for rethrowing, it's added to the chain. This provides a complete record of how an exception is handled (see Figure 1).

try {
stmt.executeUpdate(sql);
} catch (SQLException ex) {
throw new
EmployeeLookupException(
"Query failure",ex);
}

The Implementation
The exception chain is implemented as a linked list of exceptions in reverse order from last exception thrown to first. Each exception is a link in the chain. The first exception thrown that begins the chain can be of any type - there's no need for any special functionality - but all subsequent exceptions must have a way of referring to the previous exception so that the chain of exceptions can be maintained. In addition, there must be some means for accessing/examining the exception chain. I refer to exceptions that provide these things as being "chainable" - that is, an exception chain can be constructed using them.

For my purposes a "chainable" exception must do two things:

  1. Provide at least one constructor taking another (previous) exception as a parameter and storing it.
  2. Override each of the printStackTrace() methods to first print their own stack trace, then invoke the corresponding printStackTrace() method of the previous exception (given as a parameter in the constructor).
The first requirement enables the exception chain to be built. The second provides a simple means of displaying the results by invoking printStackTrace().

In practice, these simple rules allow chaining to be added to exception classes without forcing undue requirements on the users of those classes. A programmer doesn't need to know about exception chaining, even when using chainable exceptions. If chaining is used (by passing the previous exception to the new exception when converting), then the benefits are seen in the stack traces. If it's not used, there's no learning curve or overhead for the developer, and he or she is no worse off than if using nonchainable exceptions.

Any exception class can be written to meet the requirements described above, regardless of what its superclass is.

Let's examine a simple implementation of a chainable exception that extends Exception. The ChainedException (see Listing 1) can be used as a template when writing other chainable exceptions (e.g., copy the code overriding the printStackTrace() methods), or simply extended by those exceptions that would normally extend the Exception class.

Notice that the constructors take Throwables as parameters rather than Exceptions. This allows the "exception" chain to be made up of any classes that can be thrown in Java, including both compile time and runtime exceptions as well as Errors. The implementation of these constructors is straightforward; they save the reference to the given Throwable instance with the _previousThrowable member variable. This reference is used only by the printStackTrace() methods.

Also note that some constructors don't take any Exceptions or Throwables as a parameter. These constructors are used to create the root exception.

Server (and Non-GUI) Applications
Using exception chaining in server applications is simple. Build the chain as normal, and when logging exceptions be sure to use one of the printStackTrace() methods. There's no need to log each exception as they're caught. Only the last exception of the chain has to be logged, since it will display the message and call stack of each exception in the chain.

There's no harm if intermediate exceptions are logged individually, however; the worst that can happen is a slightly more cluttered log (see Figure 2). Server administrators are familiar with reading logs to determine errors, and are grateful for the extra information provided by using exception chaining.

The next section describes how a specialized dialog class can make exception chaining accessible in client programs.

Client Applications
One enormous benefit of the Java language is its resilience to errors. Java programs aren't generally "auto-terminated" as a result of an error; instead, the error is thrown (in the form of an exception) up the call stack to be handled. In a Java client application, if a runtime exception isn't handled, the main event dispatcher will print the exception's stack trace to standard out (or a Java console). This simple action is actually quite useful. While the user's button click won't appear to have done anything, at least a record of the error is available (if the Java console is visible and the user knows to look at it). Therefore, many programmers decide to handle all exceptions this way: log the exception and abort the action.

A better approach is to inform the user of any errors, usually via a dialog box. The best dialogs clearly indicate two things: what action failed, and why. I've implemented a dialog box (called ExceptionDialog) for use with exceptions that provide that information. It can be used equally with exception chains and single exceptions; it doesn't distinguish between the two.

To use the ExceptionDialog, a "try/catch" block is added to the individual GUI event handlers the programmer has registered. When an exception is caught by an event handler, an instance of ExceptionDialog is created (passing the exception as a parameter) and displayed. The title of the dialog (also passed in as a parameter to the constructor) indicates to the user what requested action has failed to occur because of the exception being thrown. The dialog message string is populated with the result of calling the exception's getLocalizedMessage() method, providing the second important piece of information - why that action couldn't be completed. Figure 3 shows an ExceptionDialog with details hidden.

try {
... event handler code ...
} catch (Exception ex) {
ExceptionDialog dlg =
new ExceptionDialog(
frame,ex,
"Look-up failed.");
dlg.show();
}

To examine the cause of the exception in more detail, the user clicks the "Show Details" button to expand the dialog. The expanded dialog (see Figure 4) displays the call stack of the exception and, if it's a chain of exceptions, the call stack of each exception in the chain. The information can be printed, or copied into an e-mail and sent to the user's help desk.

Listing 2 demonstrates exception chaining and use of the ExceptionDialog class in a GUI program. The BonusCalcException and EmployeeLookupException classes (not shown) both inherit from ChainedException.

The complete source code for all classes can be downloaded from www.JavaDevelopersJournal.com.

Where Have I Seen This Before?
The RemoteException in RMI and the ServletException from the Java Servlets API are both chainable classes. Sun refers to exceptions as being "nested" rather than "chained," but the intentions are the same.

Working on the Chain Gang
My own experience with exception chaining has shown it to be very valuable, and I tend to use it in most projects I'm involved with. However, I've found one common scenario that presents new hazards: distributed Java programs using RMI (remote method invocation).

The benefit of using RMI is that the distinction between local and remote objects is greatly reduced. It's not much more difficult to invoke a method on a remote object than a local object, except that the calling method must catch or throw RemoteException. Using exception chaining with exceptions that may be thrown during a remote method invocation offers the same benefits as with regular method invocations. It's easy to determine the root cause of errors even if they occurred in a remote object.

Unfortunately, while the sets of classes available to client and server applications overlap, they're not usually identical. Server programs have no need for GUI classes, and client programs have no need for database classes. The set of classes (or at least JAR files) available to the client is often intentionally stripped down to the minimum number required for its functionality to minimize its footprint. Therefore, if an exception chain containing a database-specific exception is thrown from the server to a remote client, it may not be possible to deal with it because the client application doesn't know what that database exception is. A ClassNotFoundException will be thrown when there's an attempt to unmarshal the unknown exception class.

Exception chaining can still be valuable in distributed applications. The easiest way to prevent the problem described above is for each application to be responsible for reporting its own exceptions. Methods that can be invoked remotely should log their internal exceptions and create a new exception for throwing remotely. The new exception should not chain the previous exception. This guarantees that no unknown exception types will be passed remotely by an exception chain.

Summary
Exception handling is often seen as a hindrance and treated as an afterthought. In fact, good exception-handling techniques can greatly simplify debugging. The exception-chaining technique is useful for both client and server components, making it easier to determine the root cause of exceptions without increasing the complexity of work required to handle them.

More Stories By Barry Mosher

Barry Mosher is a Java consultant in the San Francisco Bay Area. A Sun-certified Java programmer, Barry has more than seven years of programming experience, including three in Java.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


@ThingsExpo Stories
"My role is working with customers, helping them go through this digital transformation. I spend a lot of time talking to banks, big industries, manufacturers working through how they are integrating and transforming their IT platforms and moving them forward," explained William Morrish, General Manager Product Sales at Interoute, in this SYS-CON.tv interview at 18th Cloud Expo, held June 7-9, 2016, at the Javits Center in New York City, NY.
Apache Hadoop is emerging as a distributed platform for handling large and fast incoming streams of data. Predictive maintenance, supply chain optimization, and Internet-of-Things analysis are examples where Hadoop provides the scalable storage, processing, and analytics platform to gain meaningful insights from granular data that is typically only valuable from a large-scale, aggregate view. One architecture useful for capturing and analyzing streaming data is the Lambda Architecture, represent...
SYS-CON Events announced today that Ocean9will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Ocean9 provides cloud services for Backup, Disaster Recovery (DRaaS) and instant Innovation, and redefines enterprise infrastructure with its cloud native subscription offerings for mission critical SAP workloads.
With billions of sensors deployed worldwide, the amount of machine-generated data will soon exceed what our networks can handle. But consumers and businesses will expect seamless experiences and real-time responsiveness. What does this mean for IoT devices and the infrastructure that supports them? More of the data will need to be handled at - or closer to - the devices themselves.
SYS-CON Events announced today that SoftLayer, an IBM Company, has been named “Gold Sponsor” of SYS-CON's 18th Cloud Expo, which will take place on June 7-9, 2016, at the Javits Center in New York, New York. SoftLayer, an IBM Company, provides cloud infrastructure as a service from a growing number of data centers and network points of presence around the world. SoftLayer’s customers range from Web startups to global enterprises.
SYS-CON Events announced today that Linux Academy, the foremost online Linux and cloud training platform and community, will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Linux Academy was founded on the belief that providing high-quality, in-depth training should be available at an affordable price. Industry leaders in quality training, provided services, and student certification passes, its goal is to c...
SYS-CON Events announced today that CA Technologies has been named “Platinum Sponsor” of SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY, and the 21st International Cloud Expo®, which will take place October 31-November 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA. CA Technologies helps customers succeed in a future where every business – from apparel to energy – is being rewritten by software. From ...
In his session at @ThingsExpo, Eric Lachapelle, CEO of the Professional Evaluation and Certification Board (PECB), will provide an overview of various initiatives to certifiy the security of connected devices and future trends in ensuring public trust of IoT. Eric Lachapelle is the Chief Executive Officer of the Professional Evaluation and Certification Board (PECB), an international certification body. His role is to help companies and individuals to achieve professional, accredited and worldw...
SYS-CON Events announced today that Loom Systems will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Founded in 2015, Loom Systems delivers an advanced AI solution to predict and prevent problems in the digital business. Loom stands alone in the industry as an AI analysis platform requiring no prior math knowledge from operators, leveraging the existing staff to succeed in the digital era. With offices in S...
SYS-CON Events announced today that Interoute, owner-operator of one of Europe's largest networks and a global cloud services platform, has been named “Bronze Sponsor” of SYS-CON's 20th Cloud Expo, which will take place on June 6-8, 2017 at the Javits Center in New York, New York. Interoute is the owner-operator of one of Europe's largest networks and a global cloud services platform which encompasses 12 data centers, 14 virtual data centers and 31 colocation centers, with connections to 195 add...
SYS-CON Events announced today that T-Mobile will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. As America's Un-carrier, T-Mobile US, Inc., is redefining the way consumers and businesses buy wireless services through leading product and service innovation. The Company's advanced nationwide 4G LTE network delivers outstanding wireless experiences to 67.4 million customers who are unwilling to compromise on ...
SYS-CON Events announced today that HTBase will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. HTBase (Gartner 2016 Cool Vendor) delivers a Composable IT infrastructure solution architected for agility and increased efficiency. It turns compute, storage, and fabric into fluid pools of resources that are easily composed and re-composed to meet each application’s needs. With HTBase, companies can quickly prov...
SYS-CON Events announced today that Infranics will exhibit at SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Since 2000, Infranics has developed SysMaster Suite, which is required for the stable and efficient management of ICT infrastructure. The ICT management solution developed and provided by Infranics continues to add intelligence to the ICT infrastructure through the IMC (Infra Management Cycle) based on mathemat...
SYS-CON Events announced today that Cloudistics, an on-premises cloud computing company, has been named “Bronze Sponsor” of SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Cloudistics delivers a complete public cloud experience with composable on-premises infrastructures to medium and large enterprises. Its software-defined technology natively converges network, storage, compute, virtualization, and management into a ...
There are 66 million network cameras capturing terabytes of data. How did factories in Japan improve physical security at the facilities and improve employee productivity? Edge Computing reduces possible kilobytes of data collected per second to only a few kilobytes of data transmitted to the public cloud every day. Data is aggregated and analyzed close to sensors so only intelligent results need to be transmitted to the cloud. Non-essential data is recycled to optimize storage.
"I think that everyone recognizes that for IoT to really realize its full potential and value that it is about creating ecosystems and marketplaces and that no single vendor is able to support what is required," explained Esmeralda Swartz, VP, Marketing Enterprise and Cloud at Ericsson, in this SYS-CON.tv interview at @ThingsExpo, held June 7-9, 2016, at the Javits Center in New York City, NY.
SYS-CON Events announced today that Outlyer, a monitoring service for DevOps and operations teams, has been named “Bronze Sponsor” of SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Outlyer is a monitoring service for DevOps and Operations teams running Cloud, SaaS, Microservices and IoT deployments. Designed for today's dynamic environments that need beyond cloud-scale monitoring, we make monitoring effortless so you ...
My team embarked on building a data lake for our sales and marketing data to better understand customer journeys. This required building a hybrid data pipeline to connect our cloud CRM with the new Hadoop Data Lake. One challenge is that IT was not in a position to provide support until we proved value and marketing did not have the experience, so we embarked on the journey ourselves within the product marketing team for our line of business within Progress. In his session at @BigDataExpo, Sum...
Keeping pace with advancements in software delivery processes and tooling is taxing even for the most proficient organizations. Point tools, platforms, open source and the increasing adoption of private and public cloud services requires strong engineering rigor - all in the face of developer demands to use the tools of choice. As Agile has settled in as a mainstream practice, now DevOps has emerged as the next wave to improve software delivery speed and output. To make DevOps work, organization...
DevOps is often described as a combination of technology and culture. Without both, DevOps isn't complete. However, applying the culture to outdated technology is a recipe for disaster; as response times grow and connections between teams are delayed by technology, the culture will die. A Nutanix Enterprise Cloud has many benefits that provide the needed base for a true DevOps paradigm.