Bug 7637 – writeln doesn't take custom toString into account

import std.stdio; struct Foo { string toString(bool x) { return ""; } } void main() { Foo foo; writeln(foo.toString()); // nogo: ok writeln(foo); // compiles } The second writeln() call shouldn't compile. It seems writeln uses a default toString if it doesn't find a toString() function that takes exactly zero arguments. Note that it will call the custom toString if 'x' has a default argument.

The above code doesn't compile on 2.059 Win32. However this compiles: import std.stdio; struct Foo { string toString(bool x) { return ""; } } void main() { Foo foo; writeln(foo.toString(true)); // nogo: ok writeln(foo); // compiles } PS E:\DigitalMars\dmd2\samples> rdmd bug.d toto Foo()

(In reply to comment #1) > The above code doesn't compile on 2.059 Win32. nogo in a comment means that line doesn't compile. The second call is the interesting one.

This is current std.format design, not accepts-invalid bug. Today, std.format.formatValue (it is internally used by writeln and writefln) only supports following toString specializations: http://dlang.org/phobos/std_format.html ---- const void toString(scope void delegate(const(char)[]) sink, FormatSpec fmt); const void toString(scope void delegate(const(char)[]) sink, string fmt); const void toString(scope void delegate(const(char)[]) sink); const string toString(); If there is no toString() that matches required signatures, formatValue formats the object with 'TypeName(field-formattings-separeted-with-comma)', instead of reporting 'invalid toString signature detected' error. There is two design decides: - Ignoring invalid toString makes more user-defined types formattable. - Shouldn't prevent any kind of toString that user really wants.

Ok fair enough.