C++0x [[annotations]]

[Sektor van Skijlen]

Is there any official proposal for annotations in C++0x?

So far annotations have been "implicitly" used in many proposals as some
(usually) free-form text enclosed in [[ ]] (for example, n2493, n2509, n1943 -
not used in later updates). Is there any consistent proposal for the general
mechanism of annotations?

 

[Pete Becker]

Is there any official proposal for annotations in C++0x?

So far annotations have been "implicitly" used in many proposals as some
(usually) free-form text enclosed in [[ ]] (for example, n2493, n2509, n1943 -
not used in later updates). Is there any consistent proposal for the general
mechanism of annotations?

N2418

 

[Sektor van Skijlen]

Dnia Wed, 5 Mar 2008 21:43:50 -0500, Pete Becker skrobie:

On 2008-03-05 18:54:55 -0500, Sektor van Skijlen
<[email protected]> said:

Is there any official proposal for annotations in C++0x?

So far annotations have been "implicitly" used in many proposals as some
(usually) free-form text enclosed in [[ ]] (for example, n2493, n2509, n1943 -
not used in later updates). Is there any consistent proposal for the general
mechanism of annotations?

N2418

Ah, yes, overlooked. Tx.

 

[Alf P. Steinbach]

* Sektor van Skijlen:

Dnia Wed, 5 Mar 2008 21:43:50 -0500, Pete Becker skrobie:

On 2008-03-05 18:54:55 -0500, Sektor van Skijlen
<[email protected]> said:

Is there any official proposal for annotations in C++0x?

So far annotations have been "implicitly" used in many proposals as some
(usually) free-form text enclosed in [[ ]] (for example, n2493, n2509, n1943 -
not used in later updates). Is there any consistent proposal for the general
mechanism of annotations?

N2418

Ah, yes, overlooked. Tx.

The N2418 paper[1] refers to some existing schemes but fails to mention the
broadly used existing scheme that is most like the one in the paper, namely
Microsoft's attribute programming[2] for Visual C++.

Which uses single brackets.

There is a very large overlap in both notation and terminology, so a direct
adoption of this proposal could be very confusing to Windows programmers...


Cheers,

- Alf (in the "let's not reinvent the wheel again" mood)


Notes:
[1] <url: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2418.pdf>
[2] General discussion in article <url:
http://msdn2.microsoft.com/en-us/magazine/cc301337.aspx>, technical at <url:
http://msdn2.microsoft.com/en-us/library/zkwy014e(VS.71).aspx>.

 

[Pete Becker]

There is a very large overlap in both notation and terminology, so a
direct adoption of this proposal could be very confusing to Windows
programmers...

Microsoft is well aware of the implications of this proposal and does
not oppose it.

 

[Sektor van Skijlen]

Dnia Thu, 6 Mar 2008 11:48:00 -0500, Pete Becker skrobie:

On 2008-03-06 11:19:37 -0500, "Alf P. Steinbach" <[email protected]> said:

Microsoft is well aware of the implications of this proposal and does
not oppose it.

Steps taken by Microsoft might be less interresting, as Microsoft had invented
C++/CLI, a language that has been evaluated by the ISO C++ standard committee
as a language that would draw people out of C++ standard.

Some time ago I read the document that the ISO C++ Standard Committee produced
after evaluating C++/CLI and the only thing that makes me wonder is that ISO
C++ standard committee did not sue Microsoft, as Sun did for J++ (for C++
there were even much stronger bases for violating the rules of the standard).
I state that there were some interresting reason why it did not happen.

Just btw.

 

[Bo Persson]

Dnia Thu, 6 Mar 2008 11:48:00 -0500, Pete Becker skrobie:


Steps taken by Microsoft might be less interresting, as Microsoft
had invented C++/CLI, a language that has been evaluated by the ISO
C++ standard committee as a language that would draw people out of
C++ standard.

Some time ago I read the document that the ISO C++ Standard
Committee produced after evaluating C++/CLI and the only thing that
makes me wonder is that ISO C++ standard committee did not sue
Microsoft, as Sun did for J++ (for C++ there were even much
stronger bases for violating the rules of the standard). I state
that there were some interresting reason why it did not happen.

ISO just produces and maintains standards, it does not enforce them.


Bo Persson

 

[Sektor van Skijlen]

Dnia Sat, 8 Mar 2008 10:38:52 +0100, Bo Persson skrobie:

ISO just produces and maintains standards, it does not enforce them.

Yes, but the problem is that their implementation of C++/CLI works against
standard C++, especially that they recall their implementation as "C++". If
they called it, say, C++# (and never try to state that it is C++), there would
be no problem.

Or maybe the C++ standard committee stated that the C++/CLI is "not fierce"
because it's unlikely that anybody want to use it?