On Sat Aug 26, 2000 at 07:07:29AM +0200, Martin Schulze wrote:
> >
> > -> of which the second gives me an:
> > error in loading shared library /lib/libext2fs.so.2
> > symbol __divdi3
> > version GLIBC_2.0 not defined in file libc.so.6 with link time reference
>
> I've spent four days on fixing this, no luck. You're invited to
> develop the real fix. We went that far that symbols are in libraries
> where we didn't expect them... Doesn't help since the symbol is not
> on the rootfs.
You could add in a special case for specific symbols that are getting
stripped out for wierd reasons... Since in this case, you know that
__divdi3 is being stripped, you could just add that as a special
case so it is always included in the final /lib/libc.so.6.
In this case -- I notice that objump seems to print different sizes for
this symbol when we query libext2fs and libc:
[andersen@traveller book]$ objdump -T /lib/libc.so.6 | grep __divdi3
000c7230 g DF .text 0000019f GLIBC_2.0 __divdi3
[andersen@traveller book]$ objdump -T /lib/libext2fs.so.2 | grep __divdi3
00000000 DF *UND* 0000019e GLIBC_2.0 __divdi3
As opposed to the rest of the symbols, which are the same size...
[andersen@traveller book]$ objdump -T /lib/libc.so.6 | grep "\<malloc\>"
0004f0ec w DF .text 000001cd GLIBC_2.0 malloc
[andersen@traveller book]$ objdump -T /lib/libext2fs.so.2 | grep "\<malloc\>"
00000000 DF *UND* 000001cd GLIBC_2.0 malloc
[andersen@traveller book]$ objdump -T /lib/libc.so.6 | grep "\<printf\>"
00043e54 g DF .text 0000002f GLIBC_2.0 printf
[andersen@traveller book]$ objdump -T /lib/libext2fs.so.2 | grep "\<printf\>"
00000000 DF *UND* 0000002f GLIBC_2.0 printf
Pretty wierd...
-Erik
--
Erik B. Andersen email: [EMAIL PROTECTED]
--This message was written using 73% post-consumer electrons--
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
