«

A few days ago, tom_usenet <> wrote
in a message in this group, concerning ways of constructing
objects:

> > int main(void)
> > {
> > MyStruct Blat; // Zeros members; calls only one constructor
> > }
>
> Yup, that's the best plan, but in generic code where you want default
> initialization but don't know the type, you should probably do:
>
> T t = T();

So, Tom (or anyone who knows something about this issue):
You're saying I should construct objects like this:

MyType Blat = MyType();

Instead of:

MyType Blat;

What's the difference? Why would the former provide
"default initialization" but the latter not?

As for "in generic code", I tried using former form
inside a template class and got the error message
"ISO C++ forbids initialization of member Blat".
But the latter form worked OK.

So I'm just wondering what this "T t = T()" business
is all about.

--
Cheers,
Robbie Hatley
Tustin, CA, USA
email: lonewolfintj at pacbell dot net
web: home dot pacbell dot net slant earnur slant

* "Robbie Hatley" <lonewolfintj at pacbell dot net>:
> A few days ago, tom_usenet <> wrote
> in a message in this group, concerning ways of constructing
> objects:
>
> > > int main(void)
> > > {
> > > MyStruct Blat; // Zeros members; calls only one constructor
> > > }
> >
> > Yup, that's the best plan, but in generic code where you want default
> > initialization but don't know the type, you should probably do:
> >
> > T t = T();
>
> So, Tom (or anyone who knows something about this issue):
> You're saying I should construct objects like this:
>
> MyType Blat = MyType();
>
> Instead of:
>
> MyType Blat;
>
> What's the difference? Why would the former provide
> "default initialization" but the latter not?
>
> As for "in generic code", I tried using former form
> inside a template class and got the error message
> "ISO C++ forbids initialization of member Blat".
> But the latter form worked OK.
>
> So I'm just wondering what this "T t = T()" business
> is all about.

§8.5/9, effectively, the form "T t = T" does not initialize t when
T is a non-static POD (in the 1997 standard non-static was implied,
in the 2003 standard, Technical Corrigendum 1, it is explicit).

There is a difference between 1997 and 2003 standard regarding
"T t = T()". 1997 $8.5/7: default-initialization. 2003: "value-
initialization".

Andrew Koenig was the man, and he has explained this subtle change
very clearly a number of times in this group; unfortunately my brain
does not seem able to retain the clear understanding I always have
after reading Andrew's explanations, and the definition in the HS
involves phrasing such as "the constructor ... is called" (which
invariably makes a swarm of buzzing flies descend on me when I use it
in this newsgroup), and I'm too lazy to Google, so, I'll pass...

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

"Robbie Hatley" <lonewolfintj at pacbell dot net> wrote in message
news:40e78441$0$5859$ om...
> A few days ago, tom_usenet <> wrote
> in a message in this group, concerning ways of constructing
> objects:
>
> > > int main(void)
> > > {
> > > MyStruct Blat; // Zeros members; calls only one constructor
> > > }
> >
> > Yup, that's the best plan, but in generic code where you want
default
> > initialization but don't know the type, you should probably do:
> >
> > T t = T();
>

You can find a useful discussion and some references here:
http://www.boost.org/libs/utility/value_init.htm.

Jonathan