gcc 3.0.x/glibc 2.2.4 policy

[H . J . Lu]

On Tue, Aug 07, 2001 at 11:26:17AM +0200, Andreas Jaeger wrote:
> 
> I've tested Jakub's patch and it worked fine for me on i686 with glibc
> compiled by GCC 3.1 CVS and by GCC 2.95.3.
> 
> So how should we continue?  I see the following alternatives if
> everybody agrees that Jakub's patch is the right solution:
> 
> - release 2.2.4 this week, add Jakub's patch to CVS and release a
>   2.2.5pre1 next week.  This would imply a release of 2.2.5 next month
>   or so.
> - Add Jakub's patch now

The latest Jakub's patch is a big step towards the right direction.
But I don't think it is enough. We haven't resolved the run-time vs.
link-time version check of libgcc_s.so.1 in gcc. After it is resolved,
Jakub's patch may have to be modified. Personally, I don't believe
dlopening libgcc_s.so.1 is needed. In practice, we gain next to nothing
by dlopening libgcc_s.so.1.


H.J.