RE: [GD-General] The joy of type aliasing and C

[Crosbie F. <cr...@cy...>]

> From: Alen Ladavac
>
> Didn't see the whole thing, so perhaps this is a stupid
> question, but... is
> there any particular reason that they mention an lvalue? Because, as I
> understand, we are accessing the stored value of a float
> through an rvalue of
> type int, not lvalue of type int. (I mean, it could have just
> as well been
> (const int*), wouldn't make a difference.)


int x = ( int & ) f;
int x = * ( int * ) &f;

In both cases 'f' is an lvalue (in this example).

Perhaps one might avoid aliasing if one declared f as register or as
volatile, but it's pretty tricky to get the binary representation of a float
without going via an lvalue of some sort.

Which then prompts the questions:

1) Is it possible to convert a float to its native binary representation
without going via an lvalue (not even indirectly)?

2) What is the most portable way (that avoids risk of alias optimisation) of
getting a float into binary representation, in native form and/or a
particular standard form (IEEE, etc.)?