Why do we use finally blocks?

2019-01-07 05:52发布

问题:

As far as I can tell, both of the following code snippets will serve the same purpose. Why have finally blocks at all?

Code A:

try { /* Some code */ }
catch { /* Exception handling code */ }
finally { /* Cleanup code */ }

Code B:

try { /* Some code */ }
catch { /* Exception handling code */ }
// Cleanup code

回答1:

  • What happens if an exception you're not handling gets thrown? (I hope you're not catching Throwable...)
  • What happens if you return from inside the try block?
  • What happens if the catch block throws an exception?

A finally block makes sure that however you exit that block (modulo a few ways of aborting the whole process explicitly), it will get executed. That's important for deterministic cleanup of resources.



回答2:

Note that (in Java at least, probably also in C#) it's also possible to have a try block without a catch, but with a finally. When an exception happens in the try block, the code in the finally block is run before the exception is thrown higher up:

InputStream in = new FileInputStream("somefile.xyz");
try {
    somethingThatMightThrowAnException();
}
finally {
    // cleanup here
    in.close();
}


回答3:

You may want to put the code that you want to anyway get executed irrespective of what happens in your try or catch block.

Also if you are using multiple catch and if you want to put some code which is common for all the catch blocks this would be a place to put- but you cannot be sure that the entire code in try has been executed.

For example:

conn c1 = new connection();
try {
    c1.dosomething();
} catch (ExceptionA exa) {
    handleexA();
    //c1.close();
} catch (ExceptionB exb) {
    handleexB();
    //c1.close();
} finally {
    c1.close();
}


回答4:

Finally always gets executed, where as your code after the catch may not.



回答5:

Even though our application is closed forcefully there will be some tasks, which we must execute (like memory release, closing database, release lock, etc), if you write these lines of code in the finally block it will execute whether an exception is thrown or not...

Your application may be a collection of threads, Exception terminates the thread but not the whole application, in this case finally is more useful.

In some cases finally won't execute such as JVM Fail, Thread terminate, etc.



回答6:

Because you need that code to execute regardless of any exceptions that may be thrown. For example, you may need to clean up some unmanaged resource (the 'using' construct compiles to a try/finally block).



回答7:

There may be times when you want to execute a piece of code no matter what. Whether an exception is thrown or not. Then one uses finally.



回答8:

finally ALWAYS executes, unless the JVM was shut down, finally just provides a method to put the cleanup code in one place.

It would be too tedious if you had to put the clean up code in each of the catch blocks.



回答9:

If catch block throws any exception then remaining code will not executed hence we have to write finaly block.