2. Old way (Win32 API and COM):
MyFunction()
{
error_1 = doSomething();
if (error_1)
display error
else
{
continue processing
if (error_2)
display error
else
continue processing
}
}
3. It uses Exceptions
Exception handling enables programmers to remove
error-handling code from the “main line” of the program’s
execution.
Examples:
◦ Without exceptions: DivideByZeroNoExceptionHandling.cs
◦ With exceptions: DivideByZeroTest.sln
4. try {
// code that requires common cleanup or
// exception-recovery operations
}
catch (InvalidOperationException) {
//code that recovers from an InvalidOperationException
// (or any exception type derived from it)
}
catch (SomeOtherException) {
// code that recovers from an SomeOtherException
// (or any exception type derived from it)
}
catch {
// code that recovers from any kind of exception
// when you catch any exception, you usually re-throw
throw;
}
finally {
// code that cleans up any operations started
// within the try block. This code ALWAYS executes.
}
5. A try block contains code that requires common cleanup or
exception-recovery operations.
The cleanup code should be put in a single finally block.
The exception recovery code should be put in one or more
catch blocks.
◦ Create one catch block for each kind of type you want to handle.
A try block must have at least one catch or finally
block.
6. A catch block contains code to execute in response to
an exception.
If the code in a try block doesn’t cause an exception to
be thrown, the CLR will never execute the code in any of
its catch blocks.
You may or may not specify a catch type in
parantheses after catch :
◦ The catch type must be of type System.Exception or a type
that derived from System.Exception
◦ If there is no catch type specified, that catch block handles any
exception. This is equivalent to having a catch block that
specifies System.Exception as a catch type.
◦ CLR searches for a matching catch type from top to bottom.
◦ If CLR cannot find any catch type that matches the exception, CLR
continues searching up the callstack to find a catch type.
7. Once the catch block that matches the exception is
found, you have 3 choices:
1. Re-throw the same exception, notifying the higher-up call stack
of the exception
2. Throw a different exception, giving richer exception information
to code higher-up in the call stack
3. Let the code continue from the bottom of the catch block
In choices 1-2, an exception is thrown and code starts
looking for a catch block whose type matches the
exception thrown
In choice 3, the finally block is executed
You can also specify a variable name like
catch(Exception e) to access information specific
to the exception.
8. The CLR does not completely eliminate memory leaks. Why?
Even though GC does automatic memory clean-up, it only
cleans up if there are no references kept on the object.
Even then there may be a delay until the memory is required.
Thus, memory leaks can occur if programmers inadvertently
keep references to unwanted objects.
C# provides the finally block, which is guaranteed to
execute regardless of whether an exception occurs.
If the try block executes without throwing, the finally block
executes.
If the try block throws an exception, the finally block still executes
regardless of whether the exception is caught.
This makes the finally block ideal to release resources from
the corresponding try block.
Example: UsingExceptions.cs
9. Local variables in a try block cannot be accessed in the
corresponding finally block, so variables that must be
accessed in both should be declared before the try block.
Placing the finally block before a catch block is a syntax
error.
A try block does not require a finally block, sometimes no
clean-up is needed.
A try block can have no more than one finally block.
Avoid putting code that might throw in a finally block.
◦ Exception handling will still work but the CLR will not keep the information
about the first exception just thrown in the corresponding try block.
10. The using statement simplifies writing code in which you
obtain a resource.
The general form of a using statement is:
using ( ExampleObject e = new ExampleObject() )
{
e.SomeMethod();
}
This using statement code is equivalent to:
{
ExampleObject e = new ExampleObject();
try
{
e.SomeMethod();
}
finally
{
if ( e != null )
( ( IDisposable ) e ).Dispose();
}
}
11. In .NET, only objects of class Exception and its derived
classes may be thrown and caught.
Exceptions thrown in other .NET languages can be caught with
the general catch clause.
Class Exception is the base class of .NET’s exception class
hierarchy.
A catch block can use a base-class type to catch a hierarchy
of related exceptions.
A catch block that specifies a parameter of type
Exception can catch all exceptions.
12. Class Exception’s properties are used to formulate error
messages indicating a caught exception.
Property Message stores the error message associated with an
Exception object.
Property StackTrace contains a string that represents the method-
call stack.
When an exception occurs, a programmer might use a
different error message or indicate a new exception type.
The original exception object is stored in the InnerException
property.
Other properties:
HelpLink specifies the location of a help file that describes the
problem.
Source specifies the name of the application or object that caused
the exception.
TargetSite specifies the method where the exception originated.
13. The CLR generates SystemExceptions, derived from class
Exception, which can occur at any point during program
execution.
If a program attempts to access an out-of-range array index,
the CLR throws an exception of type
IndexOutOfRangeException.
Attempting to use a null reference causes a
NullReferenceException.
15. Example: Convert.ToInt32
Search for “Convert.ToInt32” in the Index of the
Visual Studio online documentation.
Select the document entitled Convert.ToInt32 Method.
In the document that describes the method, click the link
ToInt32(String).
The Exceptions section indicates that method
Convert.ToInt32 throws two exception types.
16. They are not an “exceptional event”, a rare event that occurs.
They are not just errors.
They are specific results returned when a method could not
complete its task.
Example: when should Transfer throw?
public class Account {
public void Transfer(Account from, Account to,
decimal amount) {
…
}
}
When the Transfer method detects any of such possible
conditions and cannot transfer the money, then it should
notify the caller that it failed by throwing an exception.
17. When implementing your own methods, you should throw an
exception when the method cannot complete its task.
Associating each type of malfunction with an appropriately
named exception class improves program clarity.
1. What Exception-derived type are you going to throw?
You must select a meaningful type
You can select a type defined in FCL that matches your semantics
If not, you may need to define your own type
Your exception type hierarchy should be shallow and wide: that is, create
as few base classes as possible
1. What string message are you going to pass to the exception
type’s constructor?
If the exception is handled, no one will see this exception message.
If the exception is unhandled, then the code will probably log this
message and a developer would want to understand what went wrong
using this message. So the message should give enough detail as
possible.
Message does not need to be localized.
18. User-defined exception classes should derive directly or
indirectly from class Exception of namespace System.
Exceptions should be documented so that other
developers will know how to handle them.
User-defined exceptions should define three constructors:
◦ a parameterless constructor
◦ a constructor that receives a string argument
(the error message)
◦ a constructor that receives a string argument and an Exception
argument (the error message and the inner exception object)
20. try {
// code that might fail…
}
catch (Exception) {
…
}
How can you write code that can recover from all
situations???
A class library should never ever swallow all exceptions. The
application should get a chance to handle the exception.
You can catch all exceptions only if you are going to process it
and re-throw it again.
21. The ability to keep cleanup code in a dedicated location and making sure
this cleanup code will execute
The ability to keep code that deals with exceptional situations in a central
place
The ability to locate and fix bugs in the code
Unified error handling: all .NET Framework classes throw exceptions to
handle error cases
Old Win32 APIs and COM returns a 32-bit error code. Exceptions include
a string description of the problem.
Exceptions also include a stack trace that tells you the path application
took until the error occurred.
You can also put any information you want in a user-defined exception of
your own.
The caller could ignore the error returned by a Win32 API, now the caller
cannot continue with exceptions. If the application cannot handle the
exception, CLR can terminate the application.
from or to could be null
from or to object might not refer to an open account
from account might have insufficient funds
to account might cause an overflow
the amount might be 0 or negative