Finding a record representation used by gnat

Finding a record representation used by gnat

[Mark]

Is it possible to find the representation used for a record compiled using 
GNAT 3.11a2 (DEC ALPHA).
I was hoping that either a gcc switch or a gnat utility to tell me this.


For those who want to know why ...

I am also compiling the same source using GNAT 3.10p under NT, and then
sending messages between the two different machine. As expected they
allocate fields differently.

The PC version is for testing, and my intention is to put a representaion
clause only on this. 

Later multiple identical machines with the same compiler will send the same
messages to each other (without the overhead of defining the representation
and allowing the representation to optimised by the compiler).
 

Re: Finding a record representation used by gnat

[Mats Weber]

Mark wrote:
> 
> Is it possible to find the representation used for a record compiled using
> GNAT 3.11a2 (DEC ALPHA).
> I was hoping that either a gcc switch or a gnat utility to tell me this.
> 
> For those who want to know why ...
> 
> I am also compiling the same source using GNAT 3.10p under NT, and then
> sending messages between the two different machine. As expected they
> allocate fields differently.

Don't do that. Different versions of the same compiler on the same
platform may come up with different record layouts. And there are also
big endian/little endian and alignment issues.

You should either use GLADE (or parts thereof) or create your own
architecture-independent protocol for transmitting your data.

Re: Finding a record representation used by gnat

[Stephen Leake]

"Mark" <mark@lourow.demon.co.uk> writes:

> Is it possible to find the representation used for a record compiled using 
> GNAT 3.11a2 (DEC ALPHA).
> I was hoping that either a gcc switch or a gnat utility to tell me this.

See LRM 13.5.2, attributes 'position, 'first_bit, 'last_bit. Write a
short program that prints the results.

-- Stephe