https://sourceware.org/bugzilla/show_bug.cgi?id=28719
Nick Clifton <nickc at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |nickc at redhat dot com
--- Comment #1 from Nick Clifton <nickc at redhat dot com> ---
Hi Achim,
I am very concerned about how adding a DWARF debug section to the linker
script can cause this effect, and I think that we need to know more about the
underlying cause. Otherwise it is possible that this problem will happen again
in the future when more debugging sections are added.
Are you able to compare working and non-working(*) binaries and check a few
things ? For example if you compile the test code to an object file and
examine the relocations in it, are there any relocations that reference the
fputs symbol for the .debug_loclists section ?
Also - when you perform the link where fputs becomes non-weak, are you able
to create a linker map (-Wl,-Map,foo.map) and then look inside it to find out
what has happened to the fputs symbol. Similarly, what kind of output do you
get if you add the -Wl,--trace-symbol=fputs ?
Cheers
Nick
(*) working = does not run and non-working = runs because there are no weak
symbols.
--
You are receiving this mail because:
You are on the CC list for the bug.
