Re: Versions.def (was Re: cvs commit: src/lib/libc/gen Symbol.map)

From: Alexander Kabaev <kabaev_at_gmail.com>
Date: Tue, 27 Nov 2007 19:56:23 -0500
On Tue, 27 Nov 2007 13:15:31 -0800
Jason Evans <jasone_at_freebsd.org> wrote:

> Alexander Kabaev wrote:
> > On Tue, 27 Nov 2007 16:22:21 +0000 (UTC)
> > Jason Evans <jasone_at_FreeBSD.org> wrote:
> >>   Modified files:
> >>     lib/libc/gen         Symbol.map 
> >>   Log:
> >>   Add _pthread_mutex_init_calloc_cb to libc's map, for which malloc
> >> defines a stub.
> > 
> > Is it a new function in FreeBSD 8.0 or are you planning to commit
> > it to FreeBSD 7.0-current. If it is for 8.0 only, it does not
> > belong where you put it.
> 
> My intent is to MFC this change after FreeBSD 7.0 is released.  This 
> brings up some questions and problems.  This change adds 
> _pthread_mutex_init_calloc_cb as a private interface that is
> implemented by lib{c,thr,kse}, so from the user's perspective there
> is no API change, but of course the libraries have to be in sync with
> each other.
> 
> 1) Do we need to introduce a new namespace to
> src/lib/libc/Versions.def for such a change?
> 
> 2) If we do introduce a new namespace in RELENG_7, does that mean
> that every time we MFC an interface, we have to add a new namespace?
> 
> 3) Suppose that we do add the FBSDprivate_1.1 namespace.  It it a 
> successor to FBSD_1.1 or FBSDprivate_1.0?  It appears that we cannot 
> have it both ways, which leads me to suspect that FBSDprivate_1.0
> should not be a successor to FBSD_1.0 (i.e. it should have no
> dependencies), and that FBSDprivate_1.1 should be a successor to
> FBSDprivate_1.0.
> 
> FWIW, it looks like once we get the Versions.def issues figured out,
> I'm guessing that _pthread_mutex_init_calloc_cb should be in 
> src/lib/libc/stdlib/Symbol.map since its definition is in the stdlib 
> directory rather than the gen directory.
> 
> Thanks,
> Jason
Hi Jason,

sorry for the fuss. I didn't realize that you have added new symbol to
FBSDPrivate section. No rules apply to it, so you are in the clear.

--
Alexander Kabaev

-- 
Alexander Kabaev

Received on Wed Nov 28 2007 - 00:25:37 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:23 UTC