We Keep Coding

Try Catch Should be used, except it never should?

So I have been doing some research into how I should be doing try-catch-finally blocks and there is some conflicting information in every post I read. Can someone clarify?
One common idea is to not catch exceptions that you do not know what to do with at that point in the code. The exception will bubble up until it presumably gets to a global exception handler if nothing else catches it. So at that point you display a message to the user that an unknown type of exception occurred, log it, etc.
Now after reading it sounds like this is the only exception handler that you will need? You should not be using it for flow control, so you should be checking if something is returned as null or is invalid causing the exception and correcting it in code. ie. testing for null and doing something about it before it can cause the exception.
But then there are other things such as running out of memory that you would not be able to test for, it would just occur during an operation and the exception would be thrown. But I would not be able to do anything about this either so this would be bubbled up to the global handler.
I think there are some that I am missing, like when dealing with files there can be some thrown from the external code. File not found exception seems like one that may come up often, so I would catch it and in the finally block gracefully close down anything I opened related to other code/processing and then notify the user and log it right there?
The only reason why you would want to catch an exception is for the finally part of the block to make sure that whatever you started before the exception is closed/finalized in a known state? But even then you would want to throw this exception after performing these tasks so the user is notified by the global exception handler, there is no point duplicating this code at this point?
So other than a global exception handler you would have try-catch-finally blocks for these scenarios.
So assuming that I am missing something here, there may be the possibility that you want to try and catch a specific type of exception and then do something with it. I cannot think of anything that you would want to do in the catch block though since the global one would log/notify the user and if you have an exception that usually means that there is no deal for the code to continue on.
Is there any easy way to know which exceptions will be thrown from which modules? I think the only way I have read is to read the MSDN or component suppliers documentation, other than that there is no way to know what exception you would be trying to catch if you were looking for a specific one (not sure why you would)
This question came up since in my application I had a section of code in a try-catch block, and it ended up that when an exception occurred it was either because an object was null, or a string was invalid. Once I wrote code to handle those scenarios the try-catch block is no longer needed, if an exception is encountered now there is nothing the code can do to recover so it should be logged and let the user know so it can be fixed.
But this goes against what I have been reading and what has been preached to me, bad code is code with no try-catch blocks. So how does this all tie together and what piece am I missing here?

The first part of your question is all correct: you should only catch exceptions that you know how to handle. Otherwise, just let them bubble up until they reach code that can handle them.
(Note that "handle" doesn't mean "log" or "display an error". It means to correct the problem that caused the exception, or work around it in some way.)
If they never encounter code that can handle them, or if they are unhandlable exceptions (like OutOfMemory), then they will eventually reach the global unhandled exception handler. This is where you will log the exception (if appropriate), display a generic error to the user (if appropriate), and more often than not, terminate the application. You cannot simply continue as if nothing happened—the exception indicates that the application is in an unexpected state. If you try and continue, you're just going to crash, or worse.
I think there are some that I am missing, like when dealing with files there can be some thrown from the external code. File not found exception seems like one that may come up often, so I would catch it and in the finally block gracefully close down anything I opened related to other code/processing and then notify the user and log it right there?
FileNotFound is a good example of an exception that you will want to handle locally. In the same method (or perhaps one level up, in your UI code) that attempts to load the file, you'll have a catch block for FileNotFound exceptions. If appropriate, display a friendly error message to the user and ask them to choose another file. If it's internal code, give up and try something else. Whatever you need to do. There are few good reasons for FileNotFound to bubble up outside of your code.
This is sort of like using exceptions for flow control, but unavoidable. There is no way to avoid using exceptions (or error codes) for I/O, so you just need to handle the failure case. You could try and verify that the file exists first, before trying to open it, but that would not solve the race issue wherein the file gets deleted or becomes otherwise inaccessible between the time your verification code runs and when you actually try and open it. So now all you've done is duplicated your error-handling code in two places, which serves little purpose.
You have to handle exceptions like FileNotFound locally. The further away from the code that throws, the less likely you can do anything reasonable about it.
Another good example of this, aside from I/O-related exceptions, is a NotSupportedException. For example, if you try to call a method that isn't supported, you might get this exception. You will likely want to handle it and have code in the catch block that falls back to a safe alternative.
The only reason why you would want to catch an exception is for the finally part of the block to make sure that whatever you started before the exception is closed/finalized in a known state? But even then you would want to throw this exception after performing these tasks so the user is notified by the global exception handler, there is no point duplicating this code at this point?
This does not require catching the exception. You can have a try block with only a finally block. A catch block is not required. In fact, this is precisely what using statement implements. If you have state that needs to be cleaned up in the event of an exception being thrown, you should implement the IDisposable pattern and wrap usage of that object in a using block.
Is there any easy way to know which exceptions will be thrown from which modules? I think the only way I have read is to read the MSDN or component suppliers documentation, other than that there is no way to know what exception you would be trying to catch if you were looking for a specific one (not sure why you would)
Precisely. This is not really a problem, though, since you are only catching the exceptions that you can do something about. If you don't know that a module can throw a particular exception, you obviously can't have written code that can handle that exception.
The documentation will tell you all of the important exceptions that you might need to handle, like FileNotFound, SecurityException, or what have you.
This question came up since in my application I had a section of code in a try-catch block, and it ended up that when an exception occurred it was either because an object was null, or a string was invalid. Once I wrote code to handle those scenarios the try-catch block is no longer needed, if an exception is encountered now there is nothing the code can do to recover so it should be logged and let the user know so it can be fixed.
Avoiding exceptions in the first place is always the best option. For example, if you can design your application so that a null object or invalid string is impossible, great. That is what we call robust code. In that case, you don't need to catch these exceptions because there's no way that you can handle it. You thought you already handled the problem, so if an exception is getting thrown anyway, it is a sign of a bug. Don't gloss over it with a catch block.
But sometimes, catch blocks are still necessary, and you write code inside of the catch block to handle the problem. In that case, there's probably no reason to re-throw the exception or log it, so you don't have any code duplication.
But this goes against what I have been reading and what has been preached to me, bad code is code with no try-catch blocks. So how does this all tie together and what piece am I missing here?
Completely wrong. I don't know where you've been reading that, but it is nonsense. Exceptions are exceptional conditions. If your code has catch blocks strewn all over it, that is a sign that you are doing it wrong. Either you're using exceptions for flow control, you're swallowing exceptions in a misguided attempt to "improve reliability", or you don't know about the global unhandled exception handler.
Doesn't sound like you're missing anything to me.
The only thing I feel compelled to mention that doesn't fit strictly into any of your questions is that sometimes you might want to catch an exception and rethrow it as a different exception. The most common situation where you would do this is if you were designing a library of re-usable code. Inside of the library, you might catch internal exceptions and, if you cannot handle them, rethrow them as general exceptions. The whole point of a library is encapsulation, so you shouldn't let exceptions bubble up that the caller cannot possibly do anything about.

There is no true guide for exceptions management (raising and handling). Every app has to decide what level of flow control should be used and how exception has to be raised/handled.
General rules are:
exceptions are raised in exceptional situations
handle exception you can handle and do meaningful things for your app so
exception raising can be used in flow control, actually it's the only way you can reliably handle flow control when you are dealing with devices, so hardware interrupts. (printers, bill validators, file transfer...)
The rest is up to you. The meaning of exception management is made by you.

Imagine you need to download some files from an FTP server, one you don't control. Of course you can't trust other people so you need to prepare for temporary outages and such. So you wrap your downloading code in a try-catch-block and look for WebException if you catch one you might check for FtpStatusCode.ActionNotTakenFileUnavailableOrBusy if that was the error you simply retry. Similarly if you call a web service a 404 might be trouble, but a 429 means that you wait a little and retry, because you had been rate-limited.
Usually you can know which exceptions can be thrown by experience or documentation but C# lacks checked exceptions. Things like NullPointerException or ArgumentNullException can be properly handled with guards in your code. But other things, like errors external dependencies can sometimes be caught and handled by you without crashing the application.

can we swallow unhandled exception in windows service

I am using windows service in c# for file transfer using sFTP (ssh.net).
I have handled exception in each and every block of code still i am getting unhandled exception and my my windows service gets crashed.
i figured out about unhandled exception by using
AppDomain.CurrentDomain.UnhandledException
Is is possible to avoid this unhandled exception and avoid service from being crashed.
Thanks in advance.
Vijay

No, because UnhandledException does not change the fact that the process is going down. It just gives you the chance to do something (e.g. log the failure) before that happens.
Even if things could work like that, the fact is that your service has bugs. You should be looking more to fix them and less to hide them.

Is is possible to avoid this unhandled exception and avoid service from being crashed.
Yes, it is. It is also a bad design decision.
AppDomain.CurrentDomain.UnhandledException
This is a good place to put in a last resort handler that writes a crash report, then restarts the service.
Swallowing exceptions is not smart. EVERY exception you expect has to be handled.
I have handled exception in each and every block of code still i am getting unhandled
exception
Sounds like you do not know what you are doing. It is not necessary to handle exceptions in every code block - it totally overloads your code with exception handlers. It is necessary, though, to have proper exception handling where it makes sense and to catch EVERY SENSIBLE (expectable) exception.
Avoiding the service from crashing is not smart - because if you do not know what exception you have, then you may end up with a corrupt service instead of a crashed one. Fail fast and hard is still the ONLY way to write reliable server systems.

The only way to avoid an unhandled exception is not to have one! Assuming your service doesn't make use of any multi-threading capabilities which could throw exceptions on separate threads, you can get away with putting an exception handler around your ServiceBase.Run call
This way any exceptions that slip through can be caught at the very last minute and handled. Usually though you want to try and grab these lower down and handle accordingly
With threaded applications it becomes a little more difficult as each new thread that is spawned won't throw on the main thread and won't be handled by this 'catch-all' handler.
Edit:
I might add that I don't condone doing this - ideally you should be handling exceptions in the right places - if you are getting unhandled exceptions you aren't expecting, then you need to examine the stack and see what went wrong. Adding a handler to AppDomain.UnhandledException and writing the exception tree and stack to the log will help (you do have some logging mechanism, right?). Either that or debugging

Catch every exception of program?

I've a WPF application which runs globally fine. But sometimes, the client tell me that he gots some crashes. But I've no information about what happened except what he does.
Is there a way to put a global try{}catch(Exception){LOGEXCEPTION; throw;} somewhere to receive all exceptions generated by a GUI action, or any other option.
Something like a "Last remedy" to log exception that we didn't manage well?
Thank you!

By handling the following events, you should be able to catch the vast majority (if not all) unhandled exceptions in your application:
Application.DispatcherUnhandledException
AppDomain.CurrentDomain.UnhandledException

C# try/catch nightmare

I have a application with similar code (not written by me)
try
{
EnumerateSomeCoolHardwareDevice();
}
catch (Exception ex)
{
}
UPDATE - This is .NET C# & EnumerateSomeCoolHardwareDevice() is using SerialPort?
I know how bad this code is but it works like this for a reason!
My question thou: I can see that it crashes somewhere in the EnumerateSomeCoolHardwareDevice(); but it doesn't get caught by the Catch (...) - It just crashes with the send report dialog! This also currently only happen in the release build... Is their ANY reason why my exception will NOT be caught by the catch (...)?

My guess is that you're not getting an Exception in your language/framework but rather EnumerateSomeCoolHardwareDevice() does weird things that simply cause the OS to kill your process. Remember that hardware details are abstracted by frameworks like Java and .NET, so whenever you do something with hardware directly, you're probably relying on unmanaged resources ... and whatever goes wrong there can kill you, catch or not.

One possible reason would be if the EnumerateSomeCoolHardwareDevice() function uses threading. If an exception is thrown in a thread and isn't handled within it's thread then it can crash an application. This simple app can demonstrate what I mean:
public static void testThread()
{
throw new Exception("oh god it's broken");
}
static void Main(string[] args)
{
try
{
Thread thread = new Thread(testThread);
thread.Start();
Console.ReadKey(); //Just to make sure we don't get out of the try-catch too soon
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
}
If you run that the app will crash and burn rather than catching the exception as you might expect.

In .NET catch (Exception ex) will only catch .NET exceptions, not native exceptions.
This question (catching native exceptions in C#) may help.

Assuming .NET, if EnumerateSomeCoolHardwareDevice uses Win32 methods via PInvoke (to access the hardware) and an error occurs, most Native methods return an error code. If that error code is not handled, and another native method is called anyway (perhaps with empty out parameters from the failed call), a serious native error (such as bad memory access or something similar) can cause a straight program crash, no exception thrown.

Have you tried the attribute
assembly:RuntimeCompatibility(WrapNonExceptionThrows = true)
That should wrap any non-.Net exceptions into System.Exception so it will be catched in your code.

If it is only happening on the production machine and not the dev machines then it could be down to DLL mismatches. Double check ALL the referenced DLLs and frameworks are the same version.
Secondly if the error is not being thrown by the EnumerateSomeCoolHardwareDevice() then it will crash the app as there is no way for the exception to get back up the stack (or thats my understanding of try/catches) in my experience this has happened to me before.
Lastly, the microsoft error report usually allows you to inspect what will be sent to MS, this should let you see where the error happened and why (assuming it has readable information within it).
Check the Event Viewer as the error should also be logged there, and normally provides an invaluable source of detail regarding the error and with a bit of digging through the error listed there you should be able to trace the fault.

If you are in .Net version 1.1 use a no parameters catch block like
catch{
...
}
Prior to .Net 2.0 there could be native exceptions that do not derive from System.Exception.
Also hook to appdomain unhandled exception event and see what happens.

There may be a try..catch inside EnumerateSomeCoolHardwareDevice().
If the exception is caught and handled there, the outer exception won't be hit unless the Exception is thrown again.

(Assuming Java) Both Error and Exception are subclasses of Throwable. If there is an assertion failing in EnumerateSomeCoolHardwareDevice() for example you will get an Error.

My guess is that there's a stack overflow happening. The .NET VM simply shuts down Release build processes that encounter a stack overflow, no CLR exceptions thrown. There's probably an internal try / catch inside that function that catches StackOverflowException one way or other, that's why it's not propagating to your code in Debug builds either.
The easiest way to figure out what's going on is by making a debug build, attaching a debugger and instructing the debugger to break before any exceptions are thrown (in Visual Studio, Debug/Exceptions and tick "Thrown" for "Common Language Runtime Exceptions" and possibly other ones as well, in cordbg.exe "catch exception")

If it is crashing madly and is using a SerialPort object then it is probably because at some point it skips onto a background thread and an exception happens here. IIRC the .DataReceived event or however you get information back from a serial port returns data on a background thread. Should an exception be thrown in this routine then the entire application will bail.
Find the background thread and put some exception handling round it.

What types of exception have you seen behaving in this way? Do you have a list of them?
Some exceptions will keep propgating up the call stack, even if they've been caught in a catch block, such as ThreadAbortException.
Others are essentially unhandleable, such as StackOverflowException or ExecutionEngineException.
If it is one of these (or some others I have probably missed), then the behaviour is probably as expected. If it is some others, then a deeper look with more information will be necessary.

What is the best way to collect crash data?

So I am sold on the concept of attempting to collect data automatically from a program - i.e., popping up a dialog box that asks the user to send the report when something goes wrong.
I'm working in MS Visual Studio C#.
From an implementation point of view, does it make sense to put a try/catch loop in my main program.cs file, around where the application is run? Like this:
try
{
Application.Run(new myMainForm());
}
catch (Exception ex)
{
//the code to build the report I want to send and to
//pop up the Problem Report form and ask the user to send
}
or does it make sense to put try/catch loops throughout pieces of the code to catch more specific exception types? (I'm thinking not because this is a new application, and putting in more specific exception catches means I have an idea of what's going to go wrong... I don't, which is why the above seems to make sense to me.)
-Adeena

I think you are right, you would not know what's going to go wrong, which is the point.
However, you might as well consider adding a handler to the ThreadException event instead.
The code above will work but there will be scenarios where multi threading might be an issue with such code, since not all code inside your windows forms program will be running in the main Application.Run loop thread.
Here's a sample code from the linked article:
[STAThread]
static void Main()
{
System.Windows.Forms.Application.ThreadException += new ThreadExceptionEventHandler(ReportError);
System.Windows.Forms.Application.Run(new MainForm());
}
private static void ReportError(object sender, ThreadExceptionEventArgs e)
{
using (ReportErrorDialog errorDlg = new ReportErrorDialog(e.Exception))
{
errorDlg.ShowDialog();
}
}
More documentation on MSDN.
On a minor point, using the ThreadException event also allow your main message loop to continue running in case the exception isn't fatal (i.e. fault-tolerance scenarios) whilst the try/catch approach may requires that you restart the main message loop which could cause side effects.

From an implementation point of view, does it make sense to put a try/catch loop in my main program.cs file, around where the application is run?
Sure and always.
You should use Try/Catch-Blocks wherever you do something critical, that might raise an exception.
Therefore you can not really stick to a pattern for that, because you should now, when what exception will be raised. Otherwise these are unhandled exceptions, which let your program crash.
But there are many exceptions, that need not to stop the application entirely, exceptions, that just can be swallowed over, as they are expected and do not critically need the application to stop. Example for that are UnauthorizedAccessExceptions when moving or accessing data with your programm.
You should try to keep you Try/Catch-Blocks as small as needed, as well as to use not too much of them, because of performance.
Some out there use Try/Catch for steering the program's execution. That should entirely be avoided wherever possible, cause raising an Exception is performance killer number 1.

Wrapping a try catch around the whole application will mean the application will exit upon error.
While using a try and catch around each method is hard to maintain.
Best practice is to use specific try catches around units of code that will throw specific exception types such as FormatException and leave the general exception handling to the application level event handlers.
try
{
//Code that could error here
}
catch (FormatException ex)
{
//Code to tell user of their error
//all other errors will be handled
//by the global error handler
}
Experience will tell you the type of things that could go wrong. Over time you will notice your app often throwing say IO exceptions around file access so you may then later catch these and give the user more information.
The global handlers for errors will catch everything else. You use these by hooking up event handlers to the two events System.Windows.Forms.Application.ThreadException (see MSDN) and AppDomain.UnhandledException (see MSDN)
Be aware that Out of Memory exceptions and StackOverflowException may not be caught by any error catching.

If you do want to automatically get stack traces, Microsoft do allow you to submit them via the Error Reporting Services. All you need to do is sign up for a Digital Certificate from VeriSign and register it (for free) with Microsoft.
Microsoft then gives you a login for you to download the mini-dumps from a website, which are submitted when a user clicks "Send Error Report".
Although people can hit "Don't Send", at least it is a Microsoft dialog box and possibly not one that you have to code yourself. It would be running 24/7, you wouldn't have to worry about uptime of your web server AND you can submit workaround details for users AND you can deliver updates via Windows Update.
Information about this service is located in this "Windows Error Reporting: Getting Started" article.

The best approach is to sing up for
AppDomain.UnhandledException and Application.ThreadException In your application's main function. This will allow you to record any unhandled exceptions in your application. Wraping run in a try catch block does not catch every thing.

if you just want to capture crashes, then ignore all errors and let DrWatson generate a minidump for you. Then you can look at that ina debugger (windbg is preferred for minidumps) and it'll show you the line your code went wrong on, and also all the parameters, stack trace, and registers. You can set Drwatson to generate a full dump where you'll get an entire memory core dump to investigate.
I wouldn't recommend putting a try/catch around the entire app unless you want your app to never "crash" in front of the user - it'll always be handled, and probably ignored as there's nothing you can do with the exception at that point.
Sending the minidump to you is another matter though, here's an article, you'll have to do some work to get it sent via email/http/ftp/etc.