From kde-core-devel  Mon Jan 12 16:00:17 2004
From: Thiago Macieira <thiago.macieira () kdemail ! net>
Date: Mon, 12 Jan 2004 16:00:17 +0000
To: kde-core-devel
Subject: Re: Static functions
Message-Id: <200401121400.24291.thiago.macieira () kdemail ! net>
X-MARC-Message: https://marc.info/?l=kde-core-devel&m=107392331411607
MIME-Version: 1
Content-Type: multipart/mixed; boundary="--Boundary-02=_YSsAAmPtU9UaBJT"


--Boundary-02=_YSsAAmPtU9UaBJT
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Bo Thorsen wrote:
>I don't "feel" anything about it - if the symbols are gone from the
> symbol table, then that's all good. But I don't believe so, the file
> static functions are necessary to achieve this. Prove me wrong :-)

You are right. Simple testing reveals:

$ cat test.cpp
int v1;
static int v2;

namespace
{
  int v3;
  static int v4;
};

main()
{
}
$ nm -C test.o
00000008 B (anonymous namespace)::v3
0000000c b (anonymous namespace)::v4
00000000 T main
00000000 B v1
00000004 b v2
$ readelf -s test.o

Symbol table '.symtab' contains 11 entries:
   Num:    Value  Size Type    Bind   Vis      Ndx Name
     0: 00000000     0 NOTYPE  LOCAL  DEFAULT  UND
     1: 00000000     0 FILE    LOCAL  DEFAULT  ABS test.cpp
     2: 00000000     0 SECTION LOCAL  DEFAULT    1
     3: 00000000     0 SECTION LOCAL  DEFAULT    2
     4: 00000000     0 SECTION LOCAL  DEFAULT    3
     5: 00000004     4 OBJECT  LOCAL  DEFAULT    3 v2
     6: 0000000c     4 OBJECT  LOCAL  DEFAULT    3=20
_ZN13_GLOBAL__N_v12v4E
     7: 00000000     0 SECTION LOCAL  DEFAULT    4
     8: 00000000     4 OBJECT  GLOBAL DEFAULT    3 v1
     9: 00000008     4 OBJECT  GLOBAL DEFAULT    3=20
_ZN13_GLOBAL__N_v12v3E
    10: 00000000    23 FUNC    GLOBAL DEFAULT    1 main


Note how 'v2' and  'v4' have a type of 'b' (meaning local .bss), while=20
'v1' and 'v3' are 'B' (global .bss). The readelf output also reveals=20
that there are three globally-exported symbols: v1, main and one=20
C++-mangled symbol ending in 'v3'.

Using a neat trick of gcc extension in ELF targets, we get more:
$ cat test.cpp
int v5 __attribute__((visibility("hidden")));
int v6 __attribute__((visibility("protected")));
int v7 __attribute__((visibility("internal")));

main()
{ }
$ nm test | grep 'v[567]$'
080494e4 B v5
080494e8 B v6
080494ec B v7
$ readelf -s test | grep 'v[567]$'
    44: 080494e8     4 OBJECT  GLOBAL PROTECTED   22 v6
    49: 080494ec     4 OBJECT  GLOBAL INTERNAL   22 v7
    61: 080494e4     4 OBJECT  GLOBAL HIDDEN   22 v5

$ nm test.so | grep 'v[567]$'
00001724 b v5
00001728 B v6
0000172c b v7
$ readelf -s test.so | grep 'v[567]$'
    22: 00001728     4 OBJECT  GLOBAL PROTECTED   19 v6
    41: 0000172c     4 OBJECT  LOCAL  INTERNAL   19 v7
    42: 00001724     4 OBJECT  LOCAL  HIDDEN   19 v5
    44: 00001728     4 OBJECT  GLOBAL PROTECTED   19 v6

And here comes the fun part: there are macros for those defined in=20
kdemacros.h:
#define KDE_NO_EXPORT __attribute__ ((visibility("hidden")))
#define KDE_EXPORT __attribute__ ((visibility("visible")))
#define KDE_PACKED __attribute__((__packed__))
# define KDE_DEPRECATED __attribute__ ((deprecated))

So we just have to use what we already have.

=2D-=20
  Thiago Macieira  -  Registered Linux user #65028
   thiagom (AT) mail (dot) com
    ICQ UIN: 1967141   PGP/GPG: 0x6EF45358; fingerprint:
    E067 918B B660 DBD1 105C  966C 33F5 F005 6EF4 5358

--Boundary-02=_YSsAAmPtU9UaBJT
Content-Type: application/pgp-signature
Content-Description: signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQBAAsSYM/XwBW70U1gRApwcAKCTq4G8KWAwiyPj6U1VlZtm306JtgCdHfxh
br2gsOya45cQmPLKI64EU2E=
=t7UR
-----END PGP SIGNATURE-----

--Boundary-02=_YSsAAmPtU9UaBJT--