C# 4.0
-
Christian Graus wrote:
I can use const to tell a user when they can trust my code not to change their stuff.
erm, I don't see how that would work. I thought that const was more of a promise to yourself, or to implementors, not to users. Nothing prevents you from calling methods and setting fields, right? Ok, let me start it again. What is the const keyword supposed to do in parameters that has an effect outside the method?
"Computer Science is no more about computers than astronomy is about telescopes." - Edsger Dijkstra
Here's an example of what I think he's saying:
public void DoSomething(const Employee emp)
{
emp.Age = 32; // This would not compile.
}An object is passed by reference, so its properties are settable. You don't want to make the Age property read only, because normally, you want the consumer to be able to set it. But what if DoSomething was meant to be a final validation, and you didn't want anything in the object to change, because it could mess up state elsewhere? The const keyword would allow the developer to leave a property or field writable, but still be able to restrict when it could be written to.
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
-
You of course would write it, to handle that case.
MrPlankton
MrPlankton wrote:
You of course would write it, to handle that case.
I thought so. :) Then it's the same as with the current lack of optional method arguments - you have to write overloads and people complain. Now you would have to write void methods...
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
Free beer!
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
really interesting!
-
here here!
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
I strongly Agree! C# cut-offs from c++ were really too deep. cons keyword (both on parameters and members signatures) shoud be a MUST in any oo language. And what about the annoying lack of default in parameters? I'd like to see c# designers and gurus dealing with Office PIA... I'm sure they would have a private build of csc.exe with default parameters implementeed .... InvokeExcelStuff(theOnlyNeededParameter, null, null, null, null, null, null, null, null, null, null, null, null, null, null....)
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
-
MrPlankton wrote:
You of course would write it, to handle that case.
I thought so. :) Then it's the same as with the current lack of optional method arguments - you have to write overloads and people complain. Now you would have to write void methods...
Well you could cast it to the method you want even if there is no left side argument. (int)functA("stuff"); I'm just saying having the return type as part of the signiture would be nice to have from time to time.
MrPlankton
-
I'd love to see a const keyword on parameters to methods, and optional parameters. Both of which seem simple enough.
Christian Graus No longer a Microsoft MVP, but still happy to answer your questions.
What would you expect 'const' to do? Would it just be the programmers promise not to change the contents of the object (generating an error if an assignment was attempted)? Would it also be able to determine if a method call on the object changed the state of the object and generate an error? If const is accepted as a parameter keyword should it also be accepted as a method modifier? How would optional parameters differ from the current params keyword?
-
Well you could cast it to the method you want even if there is no left side argument. (int)functA("stuff"); I'm just saying having the return type as part of the signiture would be nice to have from time to time.
MrPlankton
MrPlankton wrote:
I'm just saying having the return type as part of the signiture would be nice to have from time to time
I know, I've been there myself. ;)
-
Here's an example of what I think he's saying:
public void DoSomething(const Employee emp)
{
emp.Age = 32; // This would not compile.
}An object is passed by reference, so its properties are settable. You don't want to make the Age property read only, because normally, you want the consumer to be able to set it. But what if DoSomething was meant to be a final validation, and you didn't want anything in the object to change, because it could mess up state elsewhere? The const keyword would allow the developer to leave a property or field writable, but still be able to restrict when it could be written to.
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
Ok, I think I get it. So, I would be able to call methods and get properties and fields, but forbidden from setting fields and properties. Well, there are still the methods to change the data, so I guess that optimization-wise, this doesn't help. So, it's mostly syntactic sugar, a contract constraint. I'd put it in the same category as varargs, in terms of usefulness. You say that VB.NET has this feature today? How does it work when you import a VB.NET library to another language, like C#?
"Computer Science is no more about computers than astronomy is about telescopes." - Edsger Dijkstra
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
I'd like to see operator overloading supported in Interfaces, but this obvious has problems for languages that don't support op overloading. Here's what I'm thinking interface INumeric { ... overloads for +, -, /, *, remainder, (others) } class stats<T> where T:INumeric { ... methods to calculate staticial values on type T } This doesn't have to be done with interfaces (that just seems easiest). If all the numerics shared a common base class that would also be ok. I realize that this can be done with out op overloading (using only method calls), but I'd like to see it build-in so the standard types implement a basic set of operators, and users could implement their own (possibly) wacky types. That way the possibly complex methods used to calculate need only be written once. Consider, calculating Navier–Stokes equations over a large region, then as you drill down into a small region you find that you need more precision than the ~15 of doubles, so you write a high precision class that is much slower but provide the precision required. If there were a natural way to create such a class then it could just be plugged in via generics rather than requiring a bunch of complex equations to be rewritten for the new type or for method (rather than operator) expansion. Ok, that's enough wackiness for one day
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
Default initialization on auto-implemented properties!!
I'm on a Fuzzy State: Between 0 an 1
-
Sunny Ahuwanya wrote:
I think they already degraded the language in C# 3 by adding extension methods
I think of them as of an improvement and use them. :)
Pawel Krakowiak wrote:
I think of them as of an improvement and use them. Smile
Can anyone explain to me how extension methods are an improvement? Besides helping to sell LINQ and encouraging programmers to write code in a non-portable, non object oriented manner, what is the point of extension methods?
Sunny Ahuwanya "The beauty of the desert is that it hides a well somewhere" -- Antoine de Saint Exupéry
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
I have 5 that I want Basic stuff: 1. I'd like compile-time calculation and compile-time execution support. I'd like to have a compile-time assert, as well as simple calculations. I don't want to do calculations at runtime on static data I have a compile time. 2. I'd like to see language support for mixins. 3. I'd like to be able to write "const-correct" code the way I can in C++. I'd like to be able to mark a method read-only so that I know it doesn't modify the data. I'd like to be able to have a read-only object and pass it to other functions knowing that they cannot modify it. This really helps when reading other people's code, too, 'cause you can easily determine their intent. The following I think are features that the language will eventually have -- mark my words. 1. Language support for mocking. For instance, I should be able to instantiate an interface and get some dummy default implementation. This will help with stubbing. And I should be able to configure the mock in a simple way to do some other "basic" functions. 2. Language support for unit testing. I realize that everyone's thinking why, 'cause NUnit works, etc. But if you allow your mind to wonder why it would be like to have language support for Unit Testing, you can imagine some interesting possiblities. For instance, how about methods won't compile unless they have tests. Or if every path doesn't isn't executed, then you get a compiler warning. Maybe those aren't the best examples, but smarter people can come up with something more interesting.
-
Oh, well then... Treat the
usingdirective as an error. While you're at it, require full attribute names.There are so many things wrong with extension methods that I *still* get baffled how the geniuses at M$ included that as a feature. I guess LINQ must have been a HUGE thing for every team to bend their libraries and languages to 'force' it to fit in. I just want the option not to use it or make me use it in the "good way". This is just like the way you could set option strict in classic VB. Luckily almost all Microsoft extension methods come with in their exclusive namespaces. So all I have to do now is not include those namespaces BUT that doesn't prevent some other genius at XYZ company to pack extension methods with regular methods in their libraries and forcing ME to write bad code. Chip on my shoulder :)
Sunny Ahuwanya "The beauty of the desert is that it hides a well somewhere" -- Antoine de Saint Exupéry
-
S. Senthil Kumar wrote:
The methods themselves would have to be declared const, just like in C++.
Yeah, I wrote that myself, without reading your post. :doh:
S. Senthil Kumar wrote:
if a method declared const in v1 of the library became non const in v2, the "constness" guarantee will get broken (unless there is a runtime check).
I'm not sure I understand, could you please explain?
Cheers, Vıkram.
"You idiot British surprise me that your generators which grew up after Mid 50s had no brain at all." - Adnan Siddiqi.
Let's say you are using a library with the following code
// Assembly SomeLib.dll
class C
{
void someMethod() const {...}
}and the code using it looks like this
// Assembly App.exe
void usingMethod(const C obj)
{
obj.someMethod(); // ok, because someMethod is const
}Now let's say a new version of SomeLib.dll comes out and in that version, someMethod() becomes non-const. With C++, App.exe would require recompiling with the modified header file, and the compiler would be able to flag the error in
usingMethod(non const method call on a const object). With .NET and binary compatibility, there's no recompiling necessary - you just drop in the new DLL and now the const guarantee is broken. It would be a silent breaking change, unless there is a runtime check.Regards Senthil [MVP - Visual C#] _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
There are so many things wrong with extension methods that I *still* get baffled how the geniuses at M$ included that as a feature. I guess LINQ must have been a HUGE thing for every team to bend their libraries and languages to 'force' it to fit in. I just want the option not to use it or make me use it in the "good way". This is just like the way you could set option strict in classic VB. Luckily almost all Microsoft extension methods come with in their exclusive namespaces. So all I have to do now is not include those namespaces BUT that doesn't prevent some other genius at XYZ company to pack extension methods with regular methods in their libraries and forcing ME to write bad code. Chip on my shoulder :)
Sunny Ahuwanya "The beauty of the desert is that it hides a well somewhere" -- Antoine de Saint Exupéry
Sunny Ahuwanya wrote:
There are so many things wrong with extension methods
Care to list some of them? I get that they can pollute the list of methods in a class and can cause calls to unintended methods, what else do you find wrong?
Regards Senthil [MVP - Visual C#] _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
Pawel Krakowiak wrote:
I think of them as of an improvement and use them. Smile
Can anyone explain to me how extension methods are an improvement? Besides helping to sell LINQ and encouraging programmers to write code in a non-portable, non object oriented manner, what is the point of extension methods?
Sunny Ahuwanya "The beauty of the desert is that it hides a well somewhere" -- Antoine de Saint Exupéry
Sunny Ahuwanya wrote:
non-portable, non object oriented manner
Hmm, that's interesting. Is it not part of the ECMA spec? Considering there's no runtime support required for extension methods, I can't see any other portability concerns. Or maybe you are considering them non-portable because the calling code won't compile without the presence of the source code containing the extension methods? And yeah, they are non object oriented if you consider static methods in a static class non-object oriented as well.
Regards Senthil [MVP - Visual C#] _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
So now that C# 4.0 is being talked about, I was wondering what people thought would be good additions to the language. Sorry if this is a repost, but I went through several pages, and didn't see anything, so... What I'd frankly love to see would be tuples. Rather than having to use multiple 'out' parameters, you'd just return multiple values:
public int,int MinMax(int[] numbers)
{
int min, max;
// Code to calculate min/maxreturn min, max;
}What do you think? What would be good for the next version?
Kyosa Jamie Nordmeyer - Taekwondo Yi (2nd) Dan Portland, Oregon, USA
I don't know what I'd put in, but I know what I'd 'take out'. I'd restrict the use of the 'var' keyword to the Linq domain, only. I see major abuse coming, and I don't relish maintaining code which uses the 'var' keyword in a profligate fashion. D. T. Doutel