Comment #0 by bearophile_hugs — 2010-11-08T04:03:26Z
This D code gives no compilation errors with DMD 2.050:
public class Foo {
public int x = 10;
}
public class Test : Foo {
public int x = 20;
}
void main() {}
This is equivalent C# code:
public class Foo {
public int x = 10;
}
public class Test : Foo {
public int x = 20; // warning
public static void Main() {}
}
The C# compiler shows this warning:
warning CS0108: `Test.x' hides inherited member `Foo.x'. Use the new keyword if hiding was intended
The following C# code gives no warnings:
public class Foo {
public int x = 10;
}
public class Test : Foo {
new public int x = 20; // OK
public static void Main() {}
}
A similar error (or warning), with a similar "new" keyword solution, may be added to D, so unwanted hiding of attributes may be avoided.
This is related to using the obligatory "override" keyword:
bug 3836
And this warning that I think is better to become an error ASAP:
bug 4216
Comment #1 by andrej.mitrovich — 2014-04-28T11:58:32Z
This should probably be in some DIP, or you could ask about it in the forums. It's worth discussing IMO.
Comment #2 by bearophile_hugs — 2014-04-28T19:57:59Z
(In reply to Andrej Mitrovic from comment #1)
> This should probably be in some DIP, or you could ask about it in the
> forums. It's worth discussing IMO.
In the D design there are thousands of corner cases that weren't kept into account during the D design, because of reasons.
Here I am asking for a warning/error, and for a new usage of "new" (or another semantically equivalent syntax), this is both a breaking change, the introduction of a warning (and a syntax change): things currently seriously frowned upon.
Even improvements that lot of people seem to like, like the removal of implicit concatenation of adjacent string literals, or a dangerous and bad corner case: https://github.com/D-Programming-Language/dmd/pull/2887 were refused with insufficient reasons), so are such wars still worth fighting for?
Comment #3 by razvan.nitu1305 — 2022-11-29T10:29:04Z