Re: [PATCH] Add attribute "artificial"

[Devang Patel <dpatel at apple dot com>]

On Aug 10, 2005, at 5:07 PM, Daniel Berlin wrote:

> On Wed, 2005-08-10 at 16:51 -0700, Devang Patel wrote:
>
> > On Aug 10, 2005, at 4:45 PM, Daniel Berlin wrote:
> >
> >
> > > If you actually promise to do the work, i have no problem with the
> > > current patch (in theory) as a starting point.
> >
> > No, I do not. If that is the requirement then I am withdrawing my
> > patch. In my opinion having user visible attribute is just fine. Mark
> > himself admitted that it is useful in some cases.
>
> Also, why would you say "we change in in 4.2" when we can't really
> remove user attributes very easily.
> I get the feeling you think we should just this patch in now and
> "improve it" later, when it's actually important that we get it right
> now since it's involving decisions we can't easily undo later.

Because, I think it is useful to have such user visible attribute.  
Again, Mark himself admitted that it is useful in some cases.  
However, if you or someone wants make it table based, then they are  
free to do so.

> Except, again, if it's user visible, it shouldn't be marked compiler
> internal.
> That's the point of being "user visible"

It is not "marked" as compiler internal. Docs say that it is  
"suitable" for compiler provided intrinsics headers.

In any case, I have answered all questions regarding how to handle it  
in stabs and how to handle it in DWARF. Stuart has answered all  
questions/doubts regarding with ICC does in this situation. I  
convinced Daniel Jacobowitz that it is useful to have such feature.  
Geoff OKed this in principle before I actually produced patch. But we  
are still discussing this since last two weeks. I do not think we are  
making any forward progress on this, so this is my last email on this  
thread.

Thanks!
-
Devang