Cast from (long double*) to (const double*)

[ferran]

How can I cast from (long double*) to (const double*)

I have tried:
const double* Value1 = (const double*)Value2;

The compiler does not complain but the actual results when I access
the const double* are incorrect.

Note that the (long double*) is a pointer to an array of long doubles.

[David Harmon]

On Sat, 10 Apr 2004 18:24:06 +0100 in comp.lang.c++, "John Harrison"
<jo************ *@hotmail.com> wrote,

Granted, but I was just trying to emphasise to the OP that there has to be
some work done to perform the conversion (he was trying to cast), whether
that's done with an explicit loop, or in the internals of std::vector
constructor or std::copy.

Good point; certainly the ultimate conversion from long double to double
must be done one at a time somewhere. Perhaps it could be done without
needing to store the whole set of results. We don't know how the
results are to be used, but perhaps some kind of adapter class might
avoid a copying step. I guess if the destination is hard coded to
require an array of double there is not much choice but to build one.

[ferran]

Thankyou guys, I have implement a 'FOR' loop in order to copy al the
values from the array of long doubles to one of const doubles, and it
works perfectly.
One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(LondD oubleArray, LondDoubleArray + N, ConstDoubleArra y);
or
std::vector<dou ble>ConstDouble Array (LondDoubleArra y, LondDoubleArray +
N);
Thanks a lot for your help, I was getting a bit crazy.

[ferran]

Thankyou guys, I have implement a 'FOR' loop in order to copy al the
values from the array of long doubles to one of const doubles, and it
works perfectly.
One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(LondD oubleArray, LondDoubleArray + N, ConstDoubleArra y);
or
std::vector<dou ble>ConstDouble Array (LondDoubleArra y, LondDoubleArray +
N);
Thanks a lot for your help, I was getting a bit crazy.

[Kevin Goodsell]

ferran wrote:

Thankyou guys, I have implement a 'FOR' loop in order to copy al the
values from the array of long doubles to one of const doubles, and it
works perfectly.
One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(LondD oubleArray, LondDoubleArray + N, ConstDoubleArra y);
For something like this I can think of no good reason to use a 'for'
loop instead of std::copy.
or
std::vector<dou ble>ConstDouble Array (LondDoubleArra y, LondDoubleArray +
N);

This is fine if a vector is what you want, and still preferable to the
'for' loop, in my opinion.

On the other hand, your std::copy version is not fine if
ConstDoubleArra y is actually a vector. If it is an empty vector, you'd
want to pass std::back_inser ter(ConstDouble Array) as the third argument.
If it's a vector that you've already resized to N elements, you want
ConstDoubleArra y.begin() as the third argument.

-Kevin
--
My email address is valid, but changes periodically.
To contact me please use the address from a recent posting.

[Kevin Goodsell]

ferran wrote:

Thankyou guys, I have implement a 'FOR' loop in order to copy al the
values from the array of long doubles to one of const doubles, and it
works perfectly.
One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(LondD oubleArray, LondDoubleArray + N, ConstDoubleArra y);
For something like this I can think of no good reason to use a 'for'
loop instead of std::copy.
or
std::vector<dou ble>ConstDouble Array (LondDoubleArra y, LondDoubleArray +
N);

This is fine if a vector is what you want, and still preferable to the
'for' loop, in my opinion.

On the other hand, your std::copy version is not fine if
ConstDoubleArra y is actually a vector. If it is an empty vector, you'd
want to pass std::back_inser ter(ConstDouble Array) as the third argument.
If it's a vector that you've already resized to N elements, you want
ConstDoubleArra y.begin() as the third argument.

-Kevin
--
My email address is valid, but changes periodically.
To contact me please use the address from a recent posting.

[Rolf Magnus]

Kevin Goodsell wrote:

One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(LondD oubleArray, LondDoubleArray + N, ConstDoubleArra y);

For something like this I can think of no good reason to use a 'for'
loop instead of std::copy.

The OP was asking for a reason to use std::copy instead of a 'for' loop,
not the other way round.

[Kevin Goodsell]

Rolf Magnus wrote:

Kevin Goodsell wrote:

One last question, besides the looking style, is there any reason why
should I use instead of a 'FOR' loop any of the next statements (that
were mentioned)?

std::copy(Lo ndDoubleArray, LondDoubleArray + N, ConstDoubleArra y);

For something like this I can think of no good reason to use a 'for'
loop instead of std::copy.

The OP was asking for a reason to use std::copy instead of a 'for' loop,
not the other way round.

Yeah, it looks like I misread it.

The reasons are that they are shorter, simpler, easier to understand,
and harder to get wrong than the 'for' loop.

-Kevin
--
My email address is valid, but changes periodically.
To contact me please use the address from a recent posting.