Fix assertion failure on aliases of dynamic weak symbols

H.J. Lu hjl@lucon.org
Wed Jul 18 15:58:00 GMT 2007


On Tue, Jul 17, 2007 at 06:55:59AM -0700, Ian Lance Taylor wrote:
> "H.J. Lu" <hjl@lucon.org> writes:
> 
> > It sounds familar. I looked at the binutils archive. Unfortunately,
> > it starts around May 10, 1999 :-(. I will add a testcase for this.
> 
> By the way, a couple of months ago I asked Nick to ask Red Hat whether
> it would be possible for them to release the internal archive from
> before the inclusion into sourceware.  I could hook the old changes
> into the existing CVS repository as part of a conversion to
> subversion, just as I did for gcc (merging the FSF revision history
> into the egcs history).  So far I think the request has vanished into
> Red Hat's legal department.

That is too bad.

I think the proper way to fix this is

1. Select a different section for linker created symbol to avoid false
symbol alias in a shared library.
2. Match symbol types when creating symbol aliases in a shared library.

H.J.



More information about the Binutils mailing list