Re: ENVIRONMENT VARIABLES, Forest, POSIX

Re: ENVIRONMENT VARIABLES, Forest, POSIX

[Scott Moody]

The question of 'getenv' is obviously easy to solve
by making the direct call to the unix you happen to be compiling
on TODAY. What about those other non-standard unix calls
like 'getwd' -- all those that whenever I try to compile C code
from SunOS to SOlaris I hit.

That was why the POSIX interface seems to be the correct abstraction.

I am somewhat worried also about the Forest copywrite restrictions.. 
. but am more worried about the 'thick' binding that Forest provides.
I can't look at the POSIX manual, and write Ada code. I have to look at
say Forest, and see how (syntax) they let me access that code.

One answer would be to get Ada.POSIX to do it .. then place it into
the language standard, and push the burdon onto ada compilers to 
provide portability..
 The ada.command_line is a start at this..

-scott

Re: ENVIRONMENT VARIABLES, Forest, POSIX

[David Emery]

In article <DoBn2H.Ft1@plato.ds.boeing.com>, scott@plato.ds.boeing.com wrote:

> The question of 'getenv' is obviously easy to solve
> by making the direct call to the unix you happen to be compiling
> on TODAY. What about those other non-standard unix calls
> like 'getwd' -- all those that whenever I try to compile C code
> from SunOS to SOlaris I hit.
> 
> That was why the POSIX interface seems to be the correct abstraction.
> 
> I am somewhat worried also about the Forest copywrite restrictions.. 
> . but am more worried about the 'thick' binding that Forest provides.
> I can't look at the POSIX manual, and write Ada code. I have to look at
> say Forest, and see how (syntax) they let me access that code.
> 
> One answer would be to get Ada.POSIX to do it .. then place it into
> the language standard, and push the burdon onto ada compilers to 
> provide portability..
>  The ada.command_line is a start at this..
> 
> -scott

I'm very confused about what you need to write code that is not provided
by the POSIX standard.  Certainly the package specifications are an
incomplete specification of the interface, but that's no surprise, for 
POSIX or anything else...  

Do you not have a copy of IEEE Std 1003.5?

            dave