Re: A patch for default version and archive

[Ulrich Drepper <drepper at redhat dot com>]

"H . J . Lu" <hjl@valinux.com> writes:

> > I've already explained this on
> > occasion, the version information is a tuple.
> 
> That is not entirely true. I have given examples between ld.so and
> libc.so, libpthread.so and libc.so, where a symbol definition from
> one DSO overrides the other.

This doesn't change the situation.  There is still a defined symbol
which is known to exist.

> It sounds like you have a testcase. But I don't think my linker will
> successfully link. If you can show me such a testcase, I will withdraw
> my patch.

You want to make a versioned reference match if there is no symbol
definition with a version available.  I.e., a reference to symbol@ver1
matches symbol or symbol@ver1.  This is what you explained.


Anyway, you are introducing problems where none were before to fix an
issue [1] which does not commonly occur.  You are rendering the whole
versioning mechanism which safely detects mismatches before an
applications dies without warning useless since versions now can would
be modifiable at will.  There are other possibilities to achieve this
which are not sacrifizing the stability of the system.


Footnotes:
[1] Note: this is an issue only of closed source software.  And not
even many of those, only those programs which are not following the
rules set up from day one.  Therefore I do not even call it a problem.

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------