Walter Bright:
Why aren't you using Ada if this is critical to you? (I'm not
being sarcastic, this is a fair question.)
It's not critical...
Ada is not fun, too much new stuff to learn and to adapt to, and
I can't influence Ada evolution in any way. My hope for preferred
future system language is not in Ada. Still, Ada contain some
nice ideas.
I'm not saying "freeze the design".
But we are currently close to this... Lately the only ideas I've
seen seriously discussed are the ones about reference counting by
Andrei. Even the proposal about the tracking of memory ownership
was not much discussed.
I'm saying that if things are wrapped in enough bubble wrap,
few programmers will want to use the language. After all, I
don't wear a firesuit or helmet when I drive my car.
The new pre-condition is optional (you use it for Phobos structs,
etc), and for the programmer that later uses the
function/struct/class its usage its totally transparent, so the
wearing of firesuit/helmet metaphor is not good enough.
If I define a library type named Nibble that is represented with
an ubyte and accepts only values in [0, 15] with an enum
precondition, I can use it like (once Kenji patch to convert
arrays of structs is merged):
Nibble[] arr = [5, 18, 3, 1];
The usage is totally transparent for the user, no firesuits or
helmets are necessary or visible (yet that code will give a
compile-time error).
I understand. But poor C++ interop is preventing quite a few
people from using D.
I understand and I encourage this part of D design to keep going
on.
Validating printf format strings is not.
I don't care much of printf/writef. That was just an example, and
not even the most important.
2. If it exists, why aren't you using it?
If a writeln template with compile-time format string testing
goes in Phobos I'll surely use it (despite the template bloat it
will cause).
I understand that - yet another basic type in the system.
Perhaps a good enough FlagsEnum can be implemented with pure D
library code.
And formalizing a programmer idiom is often not a bad idea.
Sorry, this is just hand-waving.
I have seen it's often true ::-) New languages are essentially
invented for this purpose: to turn programmer idioms into
compiler-enforced features with a short nice syntax. The idiom of
passing a pointer + length to a C function is replaced by a much
better dynamic array of D. Even OOP is an idiom that used to be
implemented badly in C by lot of people. This list of examples
can become very long.
A reasonable optimum point is not equal to "nobody can find any
fault with it". A reasonable optimum point is where the faults
are more acceptable than the known alternatives.
I think the free mixing of signed and unsigned integral values is
not a good idea in D. I think that there are various ways to
refine immutable value range propagation.
Bye,
bearophile
