[dchest]

The problem is not in == vs =, it's in allowing assignment as an expression. Go fixes this, for example, by now allowing them: if a = 0 {} gives compile error.

[Grue3]

The problem is that == and = look similar. Banning assignment as expression only makes the language less expressive. Common Lisp fixes this, by having assignment and comparison look completely different: (setf variable value) vs (eql variable value)

[jimmaswell]

Bit of a bummer, I often find using assignment as an expression useful, like if (x = y as String) == null or such in c#. I get the feeling Go sacrifices a bit too much freedom.

[sigkill]

I'm not a professional programmer by any stretch, but wouldn't this problem completely go away if the C and C++ compilers detect a single = inside an if statement and outright refuse it. Either that or automatically substitute the assignment with checking for instances inside the if-condition.

The reason I'm asking is, because I don't know if there's ever a case where you'd want to assign right inside the if-condition. Is there?

[comex]

There can be:

    if (Foo *foo = getFoo()) { // getFoo returns null if there is no foo
        // There is a foo
    }

However, this problem is basically entirely addressed by compiler warnings, which generally ask you to add an extra set of parens for this case.

[benjamincburns]

Or you could also blame it on not having branch statements which enforce the use of a boolean type.

Edit: Actually, forget I said this. It's a dumb statement. 'if(a = true)' would always pass, regardless of the value of 'a'.