Where Is IntelliSense For "goto" statements?
-
Sorry.. I see the case you are trying to make, but in this specific case there is a simpler goto-less construct:
private static void AltFoo(int x)
{
Console.WriteLine(String.Concat("Option: ", x));if (x != 1) { Console.WriteLine("English"); } if (x == 0 || x == 1) Console.WriteLine("Spanish"); } Console.WriteLine();}
I expect your response will be something along the line of 'extend the example to more complex cases and the if-then approach becomes unwieldy'. Granted. Of the three examples (mine and your two), your second example is the best solution. It is considerably easier to understand and is more maintainable. Suppose the specification adds support for three other languages: French, German, and Italian. The goto version becomes impossible (without a lot of convoluted spaghetti code), but this code is very clear:
private static void AltFoo(int x)
{
Console.WriteLine(String.Concat("Option: ", x));switch (x) { case 0: DisplayEnglish(); DisplaySpanish(); DisplayGerman(); DisplayItalian(); break; case 1: DisplaySpanish(); break; case 2: DisplayGerman(); break; case 3: DisplayItalian(); break; default: DisplayEnglish(); break; } Console.WriteLine();}
private static void DisplayItalian()
{
Console.WriteLine("Italian");
}private static void DisplayGerman()
{
Console.WriteLine("German");
}private static void DisplaySpanish()
{
Console.WriteLine("Spanish");
}private static void DisplayEnglish()
{
Console.WriteLine("English");
}cgh1977 wrote:
I see the case you are trying to make
cgh1977 wrote:
I expect your response will be something along the line
Actually, I don't really use gotos, and was merely throwing in a generally pushed point of view, hence the use of "some people" in my first sentence. And in my example, I'd always choose the goto-less version or some similar variant.
Regards, Nish
Blog: blog.voidnish.com Most recent article: An MVVM friendly approach to adding system menu entries in a WPF application
-
Or conversely the non serious could just goto hell.
-
throw a hissy-fit, whydoncha?
-
throw a hissy-fit, whydoncha?
-
Or conversely the non serious could just goto hell.
-
Hi, Not a proper softy (I'm a hardware guy who programs on occasion) I have used a goto only in RTOS situations where if the program counter/accumlator goes beyond a certain point it will make a whole in the wall as it goes beyond reach. I was one of the spagetti basic guys who gave the goto a bad rep. It's just a way of moving around the memory so thats why the .NET compilers use it (or a jmp or jump instruction) it's only when people who didn't know what it was or how to use it got involved did it get a bad reputation. Glenn (let the abuse begin!)
I totally agree with you. Goto is used everywhere by everyone, right now, in C#, but it is abstracted to look like more complex statements. Goto can be a powerful tool in the right situation, especially in native c/c++. I agree that a lot of programmers could really screw up a pile of mud if they were given powerful tools, but I don't think that means you outlaw back-hoes. It means you smack the hand of the ones who run with scissors around others, or go ahead and poke them in the eye so they know the consequences. Allow me to protect me from myself. Don't force me to be protected from myself.
Opacity, the new Transparency.
-
Richard Blythe wrote:
I've been trying to dig up a good example of me using the "goto" statement.
I haven't used goto's, but my understanding is that they can be helpful in error handling, e.g. goto HandleError. It allows you to write code after such goto statements in such a way that it can assume that no error has occurred, thus simplifying things. One could argue that this is what exceptions are for. But the same reasoning that says that goto's are bad would seem to apply to exceptions, maybe even more so. Using goto to jump ahead within a function to execute some error handling code would seem less confusing than throwing the control flow completely out of the current context to who knows where.
That was a good point, and that's why we use pokémon error handling: to put a limit to that who knows where. The gotta catch'all approach is good when calling third party code in particular when it has the potential to be wrote letter by who knows who. By the way, I browsed my more relevant solution for an use of goto, and found only one, and I'm proud of it:
\[DebuggerNonUserCode()\] private static int Sqrt(int number) { //Newton's method aproximation for positive integers //if (n == 0) return 0; int x, \_x = number >> 1; back: x = (\_x + number / \_x) >> 1; if (x <= \_x) return \_x; \_x = x; goto back; }I did my best at that time to increase performance of that code without going to a non-portable solution. I wonder if anybody will find a better way to write it.
-
Richard Blythe wrote:
I don't use the: "goto" statement very often but I have started using it more frequently.
>SMACK!!< You what? :| I haven't found a need for a GOTO in the last, oh, 15 years...
A guide to posting questions on CodeProject[^]
Dave Kreskowiak Microsoft MVP Visual Developer - Visual Basic
2006, 2007, 2008
But no longer in 2009...Dave Kreskowiak wrote:
>SMACK!!< You what? I haven't found a need for a GOTO in the last, oh, 15 years...
You've obviously never coded in C for a low power micro with no exception handling... GOTO has it uses - mostly for bailing out to a common exit point when something goes pearshaped and you need to clean up after yourself. Only other option in these circumstances is horribly nested if statements - using a GOTO results in much cleaner/readable code. A TRY/CATCH block would do the trick nicely, but most low power (I'm not talking ARMs here, old crusty Renesas M16s and the like) micros used in industry don't have exception handling.
-
I don't use the: "goto" statement very often but I have started using it more frequently. Does anyone know why Visual Studio does not implement IntelliSense for goto statements? For example: // VS code editor does not show IntelliSense with //available "Label_" when entering this goto statement if (condition == true) goto Label_ExitCode; //... Label_ExitCode: //...
The mind is like a parachute. It doesn’t work unless it’s open.
Ever heard of try/catch - if then else ... Goto usage is the ultimate spaghetti type programming, so you don't want to go that way. It's not a crusade, it's a fact. Believe me - a former spaghetti analyser that has programmed for years in language like Data General/UBB (Universal Business BASIC). For a long time in those languages (interpreter based) the "do while/for loop" instruction didn't exist and every line in your program had a linenumber. With the goto you could force a kind of do while or even if then else. Also a goto made the program go faster (interpreter) skipping the lines that it didn't needed. The goto was of course abused making the readability very poor. Here's an example of a do while in old BASIC 100 if i<10 then 110 ... (instruction) 120 ... 130 ... 140 goto 100 150 end if Gradually the linenumber became labels and vb3,vb4,... took on this way of programming. @start 100 if i<10 then 110 ... 120 goto @start 130 end if Those (nostalgic) days are gone now... :-) I think the only reason why MS didn't completely erase the command is backwards compatibility with VB6 where you had to use it with (on) error handling. Now you have an (even better) initiative for this, the try/catch/finaly block. The finaly part is certainly an improvement. Conclusion : "Spaghetti is nice as long as it's on your plate accompagnied with some tomatoesauce, cheese and some tabasco".
modified on Tuesday, July 13, 2010 5:20 AM
-
I don't use the: "goto" statement very often but I have started using it more frequently. Does anyone know why Visual Studio does not implement IntelliSense for goto statements? For example: // VS code editor does not show IntelliSense with //available "Label_" when entering this goto statement if (condition == true) goto Label_ExitCode; //... Label_ExitCode: //...
The mind is like a parachute. It doesn’t work unless it’s open.
Richard Blythe wrote:
I don't use the: "goto" statement very often but I have started using it more frequently. Does anyone know why Visual Studio does not implement IntelliSense for goto statements?
Try it in VB, maybe C# does not have it
-
Ever heard of try/catch - if then else ... Goto usage is the ultimate spaghetti type programming, so you don't want to go that way. It's not a crusade, it's a fact. Believe me - a former spaghetti analyser that has programmed for years in language like Data General/UBB (Universal Business BASIC). For a long time in those languages (interpreter based) the "do while/for loop" instruction didn't exist and every line in your program had a linenumber. With the goto you could force a kind of do while or even if then else. Also a goto made the program go faster (interpreter) skipping the lines that it didn't needed. The goto was of course abused making the readability very poor. Here's an example of a do while in old BASIC 100 if i<10 then 110 ... (instruction) 120 ... 130 ... 140 goto 100 150 end if Gradually the linenumber became labels and vb3,vb4,... took on this way of programming. @start 100 if i<10 then 110 ... 120 goto @start 130 end if Those (nostalgic) days are gone now... :-) I think the only reason why MS didn't completely erase the command is backwards compatibility with VB6 where you had to use it with (on) error handling. Now you have an (even better) initiative for this, the try/catch/finaly block. The finaly part is certainly an improvement. Conclusion : "Spaghetti is nice as long as it's on your plate accompagnied with some tomatoesauce, cheese and some tabasco".
modified on Tuesday, July 13, 2010 5:20 AM
My primary reason is for complex validation that share the same exception. Here's an example:
public void validateSomething(someComplexType)
{
if (first_conditionFails)
goto Label_Error;
else if (complexAlgorithmRequired)
{
//init vars for algorithm...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
else //perform second phase of algorithm
{
//init vars for second phase...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
}
}
//skip the end-of-routine error code
return;Label_Error:
//set class vars...
//to an invalid state...
//...
throw new someException();
}How would you implement this with try/catch without a whole bunch of "throw someException()"?
The mind is like a parachute. It doesn’t work unless it’s open.
-
My primary reason is for complex validation that share the same exception. Here's an example:
public void validateSomething(someComplexType)
{
if (first_conditionFails)
goto Label_Error;
else if (complexAlgorithmRequired)
{
//init vars for algorithm...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
else //perform second phase of algorithm
{
//init vars for second phase...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
}
}
//skip the end-of-routine error code
return;Label_Error:
//set class vars...
//to an invalid state...
//...
throw new someException();
}How would you implement this with try/catch without a whole bunch of "throw someException()"?
The mind is like a parachute. It doesn’t work unless it’s open.
You can put the codes under
Label_Errorin a separate function, sayCatchError(). Then, replace every line of yourgoto label_errorstatement with a statement calling theCatchErrorfunction. This would be much more readable and at the same time avoid all the pitfalls ofgoto. -
My primary reason is for complex validation that share the same exception. Here's an example:
public void validateSomething(someComplexType)
{
if (first_conditionFails)
goto Label_Error;
else if (complexAlgorithmRequired)
{
//init vars for algorithm...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
else //perform second phase of algorithm
{
//init vars for second phase...
if (first_conditionFails)
goto Label_Error;
else if (second_conditionFails)
goto Label_Error;
}
}
//skip the end-of-routine error code
return;Label_Error:
//set class vars...
//to an invalid state...
//...
throw new someException();
}How would you implement this with try/catch without a whole bunch of "throw someException()"?
The mind is like a parachute. It doesn’t work unless it’s open.
Hi Richard, I didn't know for for what purpose you were trying to use the goto that's why I assumed it was for as try catch system (is the most common use). On your case it's not the try/catch that has to be used. How I would solve this problem, is by using validationfunction with a return value. This way you could even put your validation in a separate class or library. It's a standalone function. Best practice to seperate this because if your using winforms, you could easily use the same validation when you decide to use ASP.net or something else. Compile the lib as a dll, put the function public and you can use it in any language. The enum is practical in .net because the intellisense kicks in but you could easily test it on the values 1,2,3 In fact you could consider the return value as a kind of a label. I'll put it in VB because I know this best but you'll get the idea. Enum ValidateError firstcondition = 1 secondcondition = 2 thirdcondition = 3 End Enum Private Sub ValidateMessage(ByVal someComplexType) Select Case ValidateSomething(someComplexType) Case ValidateError.firstcondition MsgBox("first") Case ValidateError.secondcondition MsgBox("second") Case ValidateError.thirdcondition MsgBox("third") Case Else MsgBox("ok") End Select End Sub Private Function ValidateSomething(ByVal someComplexType) As ValidateError If someComplexType < 0 Then Return ValidateError.firstcondition End If If someComplexType > 100 Then Return ValidateError.secondcondition End If If someComplexType > 1000 Then If someComplexType > 2000 Then Return ValidateError.thirdcondition End If End If '... End Function What do you think of it ? Regards Serge
-
Dave Kreskowiak wrote:
>SMACK!!< You what? I haven't found a need for a GOTO in the last, oh, 15 years...
You've obviously never coded in C for a low power micro with no exception handling... GOTO has it uses - mostly for bailing out to a common exit point when something goes pearshaped and you need to clean up after yourself. Only other option in these circumstances is horribly nested if statements - using a GOTO results in much cleaner/readable code. A TRY/CATCH block would do the trick nicely, but most low power (I'm not talking ARMs here, old crusty Renesas M16s and the like) micros used in industry don't have exception handling.
Antzzz wrote:
You've obviously never coded in C for a low power micro with no exception handling...
Actually, I have, a very long time ago. Since you mentioned Visual Studio, you're not writing code for a micro, but for some flavor of Windows. I'm assuming you're using a managed language and not C++, which, frankly, Intellisense always has, and still does to this day, suck at.
A guide to posting questions on CodeProject[^]
Dave Kreskowiak Microsoft MVP Visual Developer - Visual Basic
2006, 2007, 2008
But no longer in 2009... -
Antzzz wrote:
You've obviously never coded in C for a low power micro with no exception handling...
Actually, I have, a very long time ago. Since you mentioned Visual Studio, you're not writing code for a micro, but for some flavor of Windows. I'm assuming you're using a managed language and not C++, which, frankly, Intellisense always has, and still does to this day, suck at.
A guide to posting questions on CodeProject[^]
Dave Kreskowiak Microsoft MVP Visual Developer - Visual Basic
2006, 2007, 2008
But no longer in 2009... -
I don't use the: "goto" statement very often but I have started using it more frequently. Does anyone know why Visual Studio does not implement IntelliSense for goto statements? For example: // VS code editor does not show IntelliSense with //available "Label_" when entering this goto statement if (condition == true) goto Label_ExitCode; //... Label_ExitCode: //...
The mind is like a parachute. It doesn’t work unless it’s open.
Richard, Yo've gotten a lot of negative replies (that I can see) because the GOTO statement is considered to be in the top 10 of BAD CODING PRACTICES in OOP. This i because GOT equates to 'Spaghetti Code' and thats something none of us ever want to hae to deal with. If you are coding properly you don;t need to use GOTO in an OOP langauge; assuming you are using an update or current version. If you were using say VB 3 or something that may be a differnet story.