Re: Re: GEC Compilation of Externals

Brian Heilig wrote:
> I remember now. I figured out how to make statements work with all
> three compilers by adding a semi-colon, just as you suggested. What I
> couldn't figure out were functions, for example:
> 
> sizeof: INTEGER is
>     -- Size to allocate (in bytes)
>   external
>     "C inline use <windows.h>"
>   alias
>     "sizeof (struct _COMMTIMEOUTS)"
>   end

I think ECMA says it should be:

  sizeof: INTEGER is
      -- Size to allocate (in bytes)
    external
      "C inline use <windows.h>"
    alias
      "return sizeof (struct _COMMTIMEOUTS);"
    end

> Eric, you mentioned implementing externals according to the ECMA
> specification. I agree that you have done that, but I don't think the
> inline mechanism is not specified well (from ECMA-367):
> 
> 8.31.14 Semantics: C Inline semantics
> In an external routine er of the inline form, an External_name of the
> form alias C_text denotes the algorithm defined, according to the
> semantics of the C language, by a C function that has:
> • As its signature, the signature specified by er.
> • As its body, C_text after replacement of every occurrence of $a,
> where the lower name of a is the lower name of one of the formal
> arguments of er, by a.
> 
> When implementing a function, simply replacing C_text will not do. A
> return statement is necessary.

I don't think you read the above semantics correctly. There is nothing
which says that C_text should be a C expression. It should be an
algorithm, in other word a piece of C code. The semantics rule of ECMA
defines what the generated result will be. Then it's up to developer
to make sure that the result will be correct C code. After all this
standard is about Eiffel, not C. So the alias clause is just interpreted
as a sequence of characters with special treatments for $ values, and
not as structured C code.

> In addition, consider the following
> example from ETL3 (http://se.inf.ethz.ch/~meyer/ongoing/etl/others.pdf
> para 31.11):
> 
> an_inline_function (x,y: INTEGER): INTEGER
>   external "[
>     C
>       inline
>       use <stdio.h>
>     ]"
>   alias "[
>     if ($x > 5) {
>       some_c_function ($y, cvar++);
>     }
>   ]"
> end
> 
> This can't work because C_text does not return a value as the
> signature of `an_inline_function' suggests it should.

But the ECMA standard and the Eiffel compiler expect that the developer
knows what it does when writing C code in inline external routines.
Perhaps the above should have been written:

an_inline_function (x,y: INTEGER): INTEGER
   external "[
     C
       inline
       use <stdio.h>
     ]"
   alias "[
     if ($x > 5) {
       return some_c_function ($y, cvar++);
     } else
       return 0;
   ]"
end


-- 
Eric Bezault
mailto:ericb-D6Qt/9opevxWk0Htik3J/[email protected]
http://www.gobosoft.com




------------------------ Yahoo! Groups Sponsor --------------------~--> 
Everything you need is one click away.  Make Yahoo! your home page now.
http://us.click.yahoo.com/AHchtC/4FxNAA/yQLSAA/saFolB/TM
--------------------------------------------------------------------~-> 

To Post a message, send it to:   [email protected]
To Unsubscribe, send a blank message to: [email protected] 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/gobo-eiffel/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/