Re: Extra version stuff in glibc?

[Jeff Bailey]

On Fri, Mar 22, 2002 at 01:36:17PM -0500, Roland McGrath wrote:
> Check the .map files in your libc build.  libc.map seems to be wrong.
> Try removing it and doing a rebuild.  If it comes out different,
> then try to reproduce the original failure with a clean build.
> This is exactly what the 2002-03-14 libc changes should have fixed.

I can now reproduce the failure - I'm trying to grok the Makefiles
enough to guess the correct fix.  I've found the place where the maps
are supposedly generated, but as far as I can tell, the rule just sets
a few variables.

Whenever --enable-oldest-abi=foo is specified (Seemingly regardless of
the value that I choose for foo), it seems to go back as far as 2.0.
If I leave it off completely, then it 2.2.5 shows for all of the
functions up top, except for:

GLIBC_2.2.6 {
  global:
    __errno_location;
}GLIBC_2.2.5;
HURD_CTHREADS_0.3 {
  global:
    __libc_getspecific;
    __mutex_init; __mutex_lock; __mutex_lock_solid; __mutex_trylock;
    __mutex_unlock; __mutex_unlock_solid;
    __spin_lock; __spin_lock_init; __spin_lock_solid; __spin_try_lock;
    __spin_unlock;
    _cthread_exit_routine; _cthread_init_routine;
    cthread_fork; cthread_detach;
    cthread_keycreate; cthread_getspecific; cthread_setspecific;
}GLIBC_2.2.6;

I'll get more info over the course of this evening.  I've been doing
this while sitting on hold for tech suport. =)

-- 
Some people play the lottery, some people use Ebay. Personally, I've
had much better luck with Ebay than the lottery
 - hendridm