This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: glibc 2.2 and versions in libc and linuxthreads


On Thu, Nov 09, 2000 at 05:58:39PM -0800, Ulrich Drepper wrote:
> "H . J . Lu" <hjl@valinux.com> writes:
> 
> > I'd like to resolve this:
> > 
> > http://sources.redhat.com/ml/libc-hacker/2000-11/msg00057.html
> > 
> > before it is too late. I think this bug may be unique to Linux.
> 
> All tries of you to handle this I've seen so far are wrong.  There

Do you have any suggestions for my testcase?

> must not be any assumption being made on the version names in
> different DSOs.  Initially we thought about using different names

I disagree. If a definition from one library is designed to override
the one from another library, I believe both name and version should
match. Otherwise, we will see problems like my testcase.

To take it further, assume one day we want GLIBC_2.2 for pread in libc
and the one in linuxthreads has to be changed also, what do we do?

> altogether but dropped the idea.
> 

If we had done that, we might have seen thread applications failures
long time ago. Why do we have both GLIBC_2_0 and GLIBC_2_1 for
pthread_attr_init in libc? They are the same in libc. If the name
doesn't matter, why bother with GLIBC_2_0 and GLIBC_2_1?

-- 
H.J. Lu (hjl@valinux.com)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]