[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[geoffk at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591



------- Additional Comments From geoffk at gcc dot gnu dot org  2003-06-20 21:24 -------
What we want to do with this is to stop giving anonymous namespace members
globally-visible names, and then we don't need to make their names unique.

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[bangerth at ices dot utexas dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591



------- Additional Comments From bangerth at ices dot utexas dot edu  2003-06-20 21:38 -------
Subject: Re:  Members of anonymous namespaces get name mangled with header name

> What we want to do with this is to stop giving anonymous
> namespace members globally-visible names, and then we don't need to make
> their names unique.

Yes, except for the open question of how this would interfere with "export". 
Not that I care about this at the moment, but should be considered before a 
fix.

On the other hand, I care very much about this particular PR, since one cannot 
use PCH with BOOST at present :-( I am mostly neutral whether this is fixed 
by munging names of giving them internal linkage, as long as it's being 
fixed...

W.

-------------------------------------------------------------------------
Wolfgang Bangerth              email:            bangerth@ices.utexas.edu
                               www: http://www.ices.utexas.edu/~bangerth/

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[geoffk at geoffk dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591



------- Additional Comments From geoffk at geoffk dot org  2003-06-20 23:02 -------
Subject: Re:  Members of anonymous namespaces get name mangled with header name

> From: Wolfgang Bangerth <bangerth@ices.utexas.edu>
> Date: Fri, 20 Jun 2003 16:37:39 -0500

> > What we want to do with this is to stop giving anonymous
> > namespace members globally-visible names, and then we don't need to make
> > their names unique.
> 
> Yes, except for the open question of how this would interfere with "export". 
> Not that I care about this at the moment, but should be considered before a 
> fix.

I don't see how it could possibly interfere with the current export
non-implementation.  If export gets implemented in future, it'll
require some way to communicate information between translation units,
and while it's doing that it can also ensure that names are unique
(which will surely be one of the smallest problems that need to be
solved).

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[bangerth at ices dot utexas dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591



------- Additional Comments From bangerth at ices dot utexas dot edu  2003-06-20 23:08 -------
Subject: Re:  Members of anonymous namespaces get name mangled with header name


> I don't see how it could possibly interfere with the current export
> non-implementation.  If export gets implemented in future, it'll
> require some way to communicate information between translation units,
> and while it's doing that it can also ensure that names are unique
> (which will surely be one of the smallest problems that need to be
> solved).

I totally agree. I don't see a problem there either. That was just one of the 
reasons I heard a while back when this topic was brought up before. (No, 
don't ask me when and where this was.) I'd favor going the way with internal 
linkage now, and think about the interaction with export when that is 
imminent. AFAIK, nobody ever even uttered he'd be interested in implementing 
export...

W.

-------------------------------------------------------------------------
Wolfgang Bangerth              email:            bangerth@ices.utexas.edu
                               www: http://www.ices.utexas.edu/~bangerth/

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |3.4

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4.0                       |---
            Version|unknown                     |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[kevina at gnu dot org]

------- Additional Comments From kevina at gnu dot org  2004-02-03 19:14 -------
A similar issue was brought up in the thread "Anonymous Namespaces" found
here http://gcc.gnu.org/ml/gcc/2004-01/msg02323.html and
http://gcc.gnu.org/ml/gcc/2004-02/msg00001.html .

The new unit-at-a-time optimizer is unable to optimize functions inside
anonymous namespaces namespace becuase they technically have external linkage.

The solution seams to be "to make the C++ front end not use (TREE_PUBLIC(x) ==
false) to mean a symbol with internal linkage.  Instead, a language-specific
decl flag ought to be allocated to mean exactly 'internal linkage',
and all local symbols ought to get TREE_PUBLIC false." according to Zack Weinberg.

I hope this will get fixed in time for Gcc 3.4.  I know nothing of Gcc internals
so I can't do it myself so I am hoping someone else will.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kevina at gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[larsbj at gullik dot net]

------- Additional Comments From larsbj at gullik dot net  2004-08-18 16:05 -------
Are there any plans to fix this for 3.5. f.ex.?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |larsbj at gullik dot net


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces get name mangled with header name

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-23 15:31 -------
*** Bug 21179 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hans at jukesan dot de


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces should be 'static'

[cpence at gmail dot com]

------- Additional Comments From cpence at gmail dot com  2005-08-01 02:24 -------
Three months later, any movement on this bug?  How is the PCH code in the 4.0
branch?  PCH support would be a godsend to people supporting large C++ projects
on GCC.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cpence at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces should be 'static'

[geoffk at geoffk dot org]

------- Additional Comments From geoffk at geoffk dot org  2005-08-01 04:31 -------
Subject: Re:  Members of anonymous namespaces should be 'static'


On 31/07/2005, at 7:24 PM, cpence at gmail dot com wrote:

> Three months later, any movement on this bug?  How is the PCH code  
> in the 4.0
> branch?  PCH support would be a godsend to people supporting large C 
> ++ projects
> on GCC.

There aren't any particular plans I know of to work on this bug,  
although I'd be happy to hear of some.  As far as I know PCH works  
fine in the 4.0 branch (but this bug isn't fixed).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces should be 'static'

[geoffk at gcc dot gnu dot org]

------- Additional Comments From geoffk at gcc dot gnu dot org  2005-09-04 18:34 -------
Although members of anonymous namespaces can be made 'static' on every target, they can all be given 
the same name on targets with weak symbol support, since on other targets the typeinfo is compared by 
name so different types must be given different names.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591

[Bug pch/10591] Members of anonymous namespaces should not be TREE_PUBLIC

[bkoz at gcc dot gnu dot org]

------- Additional Comments From bkoz at gcc dot gnu dot org  2005-09-13 02:46 -------


See:

http://gcc.gnu.org/ml/gcc-patches/2005-09/msg00772.html

This is a great idea. It would allow anonymous namespaces to be useful, and more
designs (at least g++ designs) could use them.

-benjamin

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10591