bug in tupleof ?

Derek Parnell derek at nomail.afraid.org
Wed Mar 7 22:32:30 PST 2007


On Thu, 08 Mar 2007 00:51:05 -0500, Guillaume Chereau wrote:

> Is that normal that this very simple code :
 ...

> Is it a bug in the compiler, or I am missing something about properties ?

I'm not sure if its a bug or not, but I have an explanation <G>

Have a look at this modified code ...

 import std.stdio;

 class Test {int i; float x;}

 int main()
 {
    auto t = new Test();
    t.i = 10;
    t.x = 4.2;

    writefln("(%s, %s)", t.tupleof);

    // define the tuple and show it contains .init values.
    auto f = t.tupleof;
    writefln("(%s, %s)", f);

    // update the tuple's fields.
    f[0] = 9;
    f[1] = 2.4;
    writefln("(%s, %s)", f);

    // show that the object hasn't changed.
    writefln("(%s, %s)", t.tupleof);

    // Now assign values back into the object.
    t.tupleof[0] = f[0];
    writefln("(%s, %s)", t.tupleof);
    t.tupleof[1] = f[1];
    writefln("(%s, %s)", t.tupleof);

    return 0;
 }

When you code "writefln(t.tupleof)" the compiler seems to convert that into
"writefln(t.i, t.x)", but when you code "auto f = t.tupleof" the compiler
seems to convert that into a tuple instance based on the datatypes in the
Test class, and does not use the data values that the 't' instance
contains.

The variable 'f' can be thought of a fixed length array but in which each
element has a different datatype.

-- 
Derek
(skype: derek.j.parnell)
Melbourne, Australia
"Justice for David Hicks!"
8/03/2007 5:23:29 PM


More information about the Digitalmars-d-learn mailing list