Bug 14728 – Problems in the switch statement with enumerated data types

Found by Denis M... https://vk.com/topic-471951_32042913 I think it can not write, but still it is a strange behavior and sometimes even compile. ````` import std.stdio; void main() { int x = 1; long y = 2; switch(x, y) { case x : writeln("1"); break; // Error: variable x cannot be read at compile time case y : writeln("2"); break; default: break; } } ````` import std.stdio; // compiles void main() { int x = 1; int y = 2; switch(x, y) { case x : writeln("1"); break; case y : writeln("2"); break; default: break; } } I think that should be banned in the switch to send the enumerated arguments T...

Here "," is the comma operator, which works as in C (i.e. the first expression is evaluated but its result is discarded). It is the same situation as e.g. if (x, y) { ... } The compiler should probably warn that the expression "x" has no side effects.

(In reply to Vladimir Panteleev from comment #1) > Here "," is the comma operator, which works as in C (i.e. the first > expression is evaluated but its result is discarded). It is the same > situation as e.g. > > if (x, y) { ... } > > The compiler should probably warn that the expression "x" has no side > effects. Yes, probably, it would be nice if the compiler warns about this. But I see the second problem. Why is it compiled? int x = 1; int y = 2; but it is not compiled into a switch statement? I think that this option should also be compiled: int x = 1; long y = 2;

As of dmd v2.077.1 both scenarios compile and you get a "Deprecation: Using the result of a comma expression is deprecated" Mark as FIXED?

> Mark as FIXED? Yes I think so. Otherwise please reopen. Btw it's since 2.074.1, see https://run.dlang.io/is/W3Ms3E (the All mode tests the code sample against all D compilers and shows a summary).