This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


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

Re: [PATCH] Handle weak symbols


Mark Mitchell <mark@codesourcery.com> writes:

> >>>>> "Andreas" == Andreas Jaeger <aj@suse.de> writes:
> 
>     Andreas> AFAIK this even does not work with 2.95.2 on PowerPC -
>     Andreas> but this bug is critical for glibc and we need to have it
>     Andreas> fixed. 
> 
> How can that be?  There are lots of people using glibc with old
> releases of GCC.  If this bug always existed, then it cannot be
> "critical" to have fixed -- somehow people got by until now.

It works fine on ia32 - by accident I guess - but not on other
platforms.

Btw. AFAIK we introduced in glibc 2.2.3 a change for GCC 3.0's
libstdc++ handling (the __cxa_atexit) - and this same patch uses weak
symbols and gets miscompiled by GCC.  

The other weak symbol usages in glibc worked by chance.

> Please explain in more detail why things worked OK with GCC 2.95.2 --
> which had the same bug -- but do not work with GCC 3.0.

It doesn't work with 2.95.2 on PowerPC and still does not work with 3.0.

I guess, it's better if Franz explains all the deatils since he looked
closer into these issue,

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj


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