|
|
Acer E1-571
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
Hi CP community.
We hired a guy that did this a while ago. It annoyed me then, and now I have a project that I'm refactoring that lo and behold has it as well.
Maybe I'm missing some recommended practice (I googled it), and I don't mean to start a war or anything. I just don't see an obvious use for things like this. The message is the same when a socket error occurs...
try
{
DoSomething();
}
catch (SocketException e)
{
Console.WriteLine(e.Message);
conn.connecting = false;
}
catch (Exception e)
{
Console.WriteLine(e.Message);
conn.connecting = false;
}
What purpose in life, universe, code, etc... does a practice like this serve?!
(Clarification for all of those who've been giving concrete reasons for catching different exception types. I get that. I do that as well.
I'm saying that the guy we hired previously would handle several exception types only to do the same thing as the catch-all block. Literally the same thing. Sometimes he would catch a type only to throw it to the main block doing nothing with the specific type. It bugged the hell out of me.
Now I'm refactoring another project that is completely unrelated and I see a similar practice which made my mind wander to here...)
modified 20-Jun-15 14:01pm.
|
|
|
|
|
jeeves77 wrote: What purpose in life, universe, code, etc... does a practice like this serve?!
Zero 
If the brain were so simple we could understand it, we would be so simple we couldn't. — Lyall Watson
|
|
|
|
|
Yes it has purpose. You should see this msdn article. [^]
But it wasn't utilized as suppose to.
modified 20-Jun-15 11:09am.
|
|
|
|
|
I looked, I saw: but I didn't see a purpose to having two exception catchers that do the same thing.
Yes, if the SocketException catcher logged it differently, or re-tried or similar, and the Exception catcher did a generic log-and-exit instead then it's would have a purpose.
But two exceptions (one of which catches everything) that do the same thing? I see no point, except to duplicate code and potentially cause a maintenance problem if the logging mechanism changes.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
We can't predict looking the above code since we don't know what Dosomething() method exactly do. What we can infer is that there is a socket processing behind since there is a catch exception on it.
|
|
|
|
|
It's irrelevant what DoSomething() does in regard to the catch-blocks. In case of an exception, either the first or the second catch block gets executed and they do exactly the same, so the more specific one is superfluous.
If the brain were so simple we could understand it, we would be so simple we couldn't. — Lyall Watson
|
|
|
|
|
They don't occur exactly with the same reason. Off course both catches exception but how each one will occur determine by the actual method logic. Suppose Dosomething() has DB process, then where you will think possible to be catched if error happens. On catch (Exception ex) block and where do you think the socket error will be catched on (SocketException sex). That is why the method logic determine the exception occurence.
|
|
|
|
|
It doesn't matter.
Regardless of what exception is thrown, the code executed will be identical. Which means that the separate catch for the "lesser" exception need not be there at all.
And if it needed be there, why put it in? All that does is leave a "hole" where duplicated code can become different over time and cause other problems - that's one of the advantages of inheritance: it removes the need to "copy and paste" code by reusing a single method in all derived classes. So if the method needs to be changed, it's in one single place instead of scattered all over the file and prone to being missed when the updates are done.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
We all know that managed .Net exception are derived from Exception class. If you are asking why the rest of derived Exception such as socketexception, Dbexception are created then that is another question. And it won't be necessary to allow multiple catch statement block from the compiler. Off course if you leave/comment out the above socketexception then the error definitely go to exception block, b/c of the reason that I stated above.
Onething sure to know here, you can't determine to refactor the code by looking that code only, even we don't know either a SocketException will occur or not.
G. Day
|
|
|
|
|
It doesn't matter if a SocketException or an AliensLandedOnWhiteHouseLawnException occurs with that code: the code that is executed is identical regardless.
That is the point. Not that Socket Exceptions are derived from Exception - we all know that - but that having a separate catch block is silly if the code executed is the same anyway!
Look at the code. Assume a SocketException occurs in the method. What lines of code are executed?
Now assume that a AliensLandedOnWhiteHouseLawnException occurs instead. What lines of code are executed?
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
It does matter, a self ref [^]
How do you know the code executed the same ? YOU DON'T KNOW WHAT THE METHOD(Dosomething()) DO. What we know both exception was not utilized as such.
|
|
|
|
|
Member 11394652 wrote: YOU DON'T KNOW WHAT THE METHOD(Dosomething()) DO
Seriously, it doesn't matter.
Look at the code we do have:
catch (SocketException e)
{
Console.WriteLine(e.Message);
conn.connecting = false;
}
catch (Exception e)
{
Console.WriteLine(e.Message);
conn.connecting = false;
}The code in each catch block is identical.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
Are you saying the message will always be the same ?
This is a trick, If you do get it. 
|
|
|
|
|
Yes. The message comes from the Exception object, and will the same regardless of which catch block catches it: for a SocketException it will print a socket based message, for an AliensLandedOnWhiteHouseLawnException, it will print a message in Klingon.
But it is irrelevant which catch block prints it, because they both use the same code to do that.
There is only any point in having separate catches if they do different things:
catch (SocketException e)
{
Console.WriteLine("Problem with socket: {0}", e.Message);
conn.connecting = false;
}
catch (Exception e)
{
Console.WriteLine("An unknown error occured: {0}", e.Message);
conn.connecting = false;
}Or even:
catch (SocketException e)
{
Console.WriteLine("It's life Jim: {0}", e.Message);
conn.connecting = false;
}
catch (AliensLandedOnWhiteHouseLawnException e)
{
Console.WriteLine("Klingons on the starboard bow, starboard bow, starboard bow\nKlingons on the starboard bow, starboard bow Jim!: {0}", e.Message);
conn.connecting = false;
}
catch (Exception e)
{
Console.WriteLine("But not as we know it: {0}", e.Message);
conn.connecting = false;
}See what we mean?
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
The fact that you still ignore the Dosomething() method do, you generalize that both catches block get the same exception. Even I gave you example that the method might throw DBException and you expect the SocketException get exception based on your assumption.
|
|
|
|
|
I don't need to know what DoSomething() does to reason this:
original code:
SocketException > catch block 1 > { Console.WriteLine(e.Message); conn.connecting = false; }
other Exception > catch block 2 > { Console.WriteLine(e.Message); conn.connecting = false; }
after removing catch block 1:
SocketException > catch block > { Console.WriteLine(e.Message); conn.connecting = false; }
other Exception > catch block > { Console.WriteLine(e.Message); conn.connecting = false; }
ergo: catch block 1 is redundant
If the brain were so simple we could understand it, we would be so simple we couldn't. — Lyall Watson
|
|
|
|
|
Yes you do need to have the catch block. In fact you may need to add more based on what Dosomething()does. What op need to do is to implement the appropriate logic in the catch block which I already said it in my first response.
|
|
|
|
|
Take the code from the original post and put it into Main() of a new console project. Replace conn.connecting with a bool connecting or throw it out. Instead of DoSomething() throw a SocketException or create a method DoSomething() and throw the SocketException there. Run it and take note of the console output. Remove the first catch block and run again. You'll notice that it's the same output. Ergo the first catch block is redundant in this particular code sample.
If the brain were so simple we could understand it, we would be so simple we couldn't. — Lyall Watson
|
|
|
|
|
No that is not the intension. Op misses to implement the appropriate logic behind once the error occurred in SocketException which he copied similar catch block in the Exception block.
|
|
|
|
|
True. But you missed that we were explicitly not talking about how it should be done right but about whether the two catch blocks make sense the way they are.
If the brain were so simple we could understand it, we would be so simple we couldn't. — Lyall Watson
|
|
|
|
|
They make sense since the Op knows that there is socket related exception which Op wrote to it's own catch block. To me the the code was only getting SocketException for a while and then he got other excepion which it couldn't caught by SocketException and then he added the same catch code block with Exception without taking the account of future coders since he got relief at that moment.
|
|
|
|
|
No, I don't.
But both catch blocks have the same code.
So they both do the same thing regardless of which exception gets fired.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
