https://sourceware.org/bugzilla/show_bug.cgi?id=20159
--- Comment #9 from Paul Braman <paul.braman at nielsen dot com> ---
HJ, maybe you might be on to something ...
I took the test case and compiled it clean on both machines. If I use the
executable I created on Ubuntu 14.04, the test case runs as expected regardless
of which shared library it is running against and regardless of which VM I am
running it on. The reverse (executable built on 16.04) is also true (fails
regardless of library/VM).
[Someone might want to double check my work.]
[bramankp@ubuntu-1404:linkertest] gcc -fPIC -c -I. -o nothing.lo nothing.c
[bramankp@ubuntu-1404:linkertest] gcc -shared -rdynamic -o libnothing.so
nothing.lo
[bramankp@ubuntu-1404:linkertest] gcc -o something something.c faketime.c
-Wl,-rpath=. -L. -lnothing
[bramankp@ubuntu-1404:linkertest] ./something
42
42
[bramankp@ubuntu-1404:linkertest] md5sum something
1f2e26a68c8e7576fafda6aa0e947490 something
[bramankp@ubuntu-1404:linkertest] scp
bramankp@192.168.1.201:test/linkertest/something ./
bramankp@192.168.1.201's password:
something
100% 8704 8.5KB/s 00:00
[bramankp@ubuntu-1404:linkertest] md5sum something
3464ea630b6a075e6adf9b15aaf3412f something
[bramankp@ubuntu-1404:linkertest] ./something
42
1464701962
[bramankp@ubuntu-1404:linkertest]
This is just one of the example cases. I also moved the library built from
14.04 to 16.04 and then compiled the executable and ran it (and visa versa).
Dynamic linking, itself, does not seem to be the culprit. I'm not sure where
the problem lies now (GCC, glibc, or ld).
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils
