Common areas and as

Nick Burrett nick at
Thu Oct 30 01:52:05 PST 2003

David Marston wrote:
> In message <3F9CE5C6.8080505 at>
>           Nick Burrett <nick at> wrote:

>>Thanks for the analysis.  I get the impression that we aren't making 
>>symbol 'x' an member of area 'x'.  We probably aren't generating a 
>>relocation at all.
> They were there, just wrong. I've found the problem now, and can't
> believe I didn't see it the first time given I'd changed the same piece
> of logic in quite a few places already.
> (It's the changes in reloc.c in particular that are for this problem,
> but I thought I'd make sure it was changed everywhere)

Looking at your patch, I was sure I had made exactly the same fixes. 
However on closer inspection, I forgot to change reloc.c. :-)

Running the testsuite now gives me one expected failure.  It appears 
that at least the C compiler is now up to scratch.

I'm now running the libstdc++ testsuite which will take a few hours.



Nick Burrett
Network Engineer, Designer Servers Ltd.

More information about the gcc mailing list