Hm, strange. The non-stripped version doesn't show anything useful it
gdb's backtrace as well:
(gdb) bt
#0 0x0000000000451c61 in strcmp ()
#1 0x000000000040f161 in strcmp ()
#2 0x000000000043848e in strcmp ()
#3 0x000000000040b22c in strcmp ()
#4 0x000000000040801c in strcmp ()
#5 0x00000036c781d974 in __libc_start_main () from /lib64/libc.so.6
#6 0x0000000000407609 in strcmp ()
#7 0x00007fffacffdbb8 in ?? ()
#8 0x0000000000000000 in ?? ()
That process is using a binary from "~/src". I wonder if you have time
today to join me in our Campfire chat to discuss the problem? We are
eager to fix this.
https://wildbit.campfirenow.com/9fe4f
Thanks,
Ilya Sabanin.
On 13.10.2009, at 21:08, Alexander V. Lukyanov wrote:
On Tue, Oct 13, 2009 at 08:01:33PM +0800, Ilya Sabanin wrote:
I just checked both binaries with "nm -D". "/usr/local/bin/lftp"
reported a long list of symbols. "~/src/lftp-4.0.2/src/lftp" reported
no symbols. So I guess you are right about non-stripped and stripped.
Can usage of a non-stripped binary somehow affect our setup? In a bad
way I mean.
No, the debug info is not loaded into memory when the program runs.
It is
only loaded when a debugger is used.
--
Alexander..
