Undefined Symbol when compiled with -O2 only on Mac - c

I'm working on a library that must compile on linux and mac os X. Until now, I had no problem, compiling with "-g" worked well under both OS.
I tried to compile with some optimization ("-O2") and it works well under linux but I get an Undefined Symbol when I try to link a program with my library under mac os X.
Does anyone have any clue what I should look for?
nm mylib.a | grep _the_symbol
This returns same thing for linux and mac (no leading underscore under linux) :
154:00000018 C _the_symbol
377: U _the_symbol
Here is the compile line under linux for the program using the library:
/usr/bin/gcc -std=c99 CMakeFiles/prod-cons.dir/prod-cons.c.o -o prod-cons -rdynamic -L/home/claferri/dev/build/src ../src/libckaapi.a -lpthread -Wl,-rpath,/home/claferri/dev/build/src
And under mac :
/usr/bin/gcc -std=c99 -Wl,-search_paths_first -headerpad_max_install_names -fPIC CMakeFiles/prod-cons.dir/prod-cons.c.o -o prod-cons -L/Volumes/Data/claferri/Work/build/src ../src/libckaapi.a /usr/lib/libpthread.dylib

Here's a guess at a workaround: try building the library with the -fno-common flag. If you have multiple definitions of this variable, you'll need to add "extern" to all but one.

Note that the following is a guess, and I can't say for certain unless/until you provide the exact compiler flags you're using -- but Xcode defaults to setting -fvisibility=hidden, which would hide pretty much any symbol in your library, unless it's declared as visible.
You can do the same on Linux, but GCC's default is not to hide symbols.
You'll find more information here: http://gcc.gnu.org/wiki/Visibility

Related

-lm doesnt work unless it's at the end of the command [duplicate]

This question already has answers here:
Why does the order in which libraries are linked sometimes cause errors in GCC?
(9 answers)
Closed 2 years ago.
Im currently writing a program for a uni asssessment and they have a set line to compile it, so if it doesn't work with that it won't be accepted.
They command they use is
gcc -Wall -ansi -lm program.c -o program.out
My program will not compile that way, and it'll give me a undefined referance error (Referring to my log10 using math.h library)
if i use:
gcc -Wall -ansi program.c -o program.out -lm
it works
What could be my issue?
Im using windows 10 64bit and have windows bash installed and gcc.
This would be explained if your instructors are using gold and you are using GNU ld. These are two linkers, both are part of the GNU project, and both are commonly used with GCC.
If you are using GNU ld, you get the "traditional" behavior:
The order of specifying the -L and -l options, and the order of specifying -l options with respect to pathname operands is significant.
This means that you have to put -lm after any object files and libraries that depend on it.
However, if you are using gold, the -l options may appear first.
If you have gold installed on your system, you can test it yourself.
Here is what I get:
$ gcc -lm program.c
/tmp/ccJmBjmd.o: In function `main':
program.c:(.text+0x15): undefined reference to `sin'
collect2: error: ld returned 1 exit status
But if I use gold, it works fine:
$ gcc -lm program.c -fuse-ld=gold
-lm needs to be at the end of the command, most likely in the first case with the literal the compiler is optimizing out the call to any function and therefore does not need to link against the library. This is called constant folding and for example we can see in the gcc docs on Other Built-in Functions Provided by GCC says:
GCC includes built-in versions of many of the functions in the
standard C library. The versions prefixed with __builtin_ are always
treated as having the same meaning as the C library function even if
you specify the -fno-builtin option. (see C Dialect Options) Many of
these functions are only optimized in certain cases; if they are not
optimized in a particular case, a call to the library function is
emitted.

LIBRARY_PATH environment variable not being used / read with gcc

My LIBRARY_PATH variable is exported, but I still have to pass the -L option to gcc in order to link to my library.
If I understand the GCC documentation correctly 3.20 Environment Variables Affecting GCC, the LIBRARY_PATH environment variable should be looked so that I only have to specify the -l option.
When I run
gcc -Wall cog.c -L$HOME/lib -lutil
the program is compiled, and I get an a.out as expected.
If I run
gcc -Wall cog.c -lutil
I get an undefined reference error.
As far as I can tell, I've properly exported the environment variable.
cassiopeia~> export LIBRARY_PATH=$HOME/lib
cassiopeia~> ls $LIBRARY_PATH
libutil.a
Any clues?
For what it's worth, I'm using Fedora 23 64bit and gcc version 5.3.1 20160406 (Red Hat 5.3.1-6).
Your distro probably is multilib-enabled. If this is the case, all path strings to libraries are expanded with the architecture for this machine (typically 32-bit or 64-bit). So, if you specify
$HOME/lib
as your search path, multilib might expand it to
$HOME/lib/x86_64-linux/4.6
or
$HOME/lib/x86_32-linux/4.6
You can check if this is the case by invoking gcc once using
gcc --print-search-dirs
This makes gcc respond with all search paths in use for config and libraries.

cc linker gives no error, but the resulting library is not linked to libgstvideo-1.0

I'm using cmake to generate a gstreamer library. In the end cmake uses the following command for linking:
/usr/bin/cc -fPIC -shared -Wl,-soname,libmacq-gstmelpi.so -o libmacq-gstmelpi.so <OBJECT_FILES> -lmacq-melpi -lmacq-icar-tools -lmacq-gstmecimeta -lgstreamer-1.0 -lgobject-2.0 -lglib-2.0 -lgstvideo-1.0
Note the -lgstvideo-1.0 flag at the end. The command runs fine, no error is produced, and the resulting library is created just fine. However when I use the library, I get undefined symbol error. So I use ldd to check; and amongst all the output of ldd; libgstvideo-1.0.so is not to be found.
This problem occurs on Ubuntu 14.04 on a armhf architecture. The problem does not occur on opensuse 13.1 (i586) nor on opensuse 13.1 (armv7hl), since in that case ldd libmacq-gstmelpi.so | grep gstvideo gives:
libgstvideo-1.0.so.0 => /usr/lib/libgstvideo-1.0.so.0 (0xb715f000)
EDIT :
I have another library, very similar where a very similar command works just fine; the resulting library is correctly linked to libgstvideo-1.0.so
/usr/bin/cc -fPIC -shared -Wl,-soname,libmacq-gstplugins.so -o libmacq-gstplugins.so <OBJECT_FILES> -lmacq-icar-tools -lmacq-gstmecimeta -lgstapp-1.0 -lgstbase-1.0 -lgstreamer-1.0 -lgobject-2.0 -lglib-2.0 -lgstvideo-1.0
Some remarks on what I have checked and tried:
/usr/lib/arm-linux-gnueabihf/libgstvideo-1.0.so exists. (the other libraries in /usr/lib/arm-linux-gnueabihf/ are found without problem, libmacq-gstmelpi.so is linked to /usr/lib/arm-linux-gnueabihf/libgstreamer-1.0.so.0 without problem)
I tried changing the order, and putting the -lgstvideo-1.0 flag before all other -l flags; no succes.
replacing cc with c++ or gcc; the commands work, but output is the same
removing -lgstvideo-1.0 on a system where the build worked. The resulting library builds (links) without error; yet upon execution I have the same undefined symbol error as on Ubuntu. This proves that the missing symbol is in libgstvideo-1.0.so, and that I need it.
Are you actually using symbols from that particular library or do you just want to link to it to avoid linking it in the application that uses the library later on?
It could be a default compiler behavior that it skips linking of libraries when no symbol from these are actually used.
Try -Wl,--no-as-needed to your flags. In this case the library should get linked - not checking whether its symbols are actually used or not.
EDIT:
After chatting it turned out that the actual desired symbols are in gstbase-1.0 and not gstvideo-1.0. Since gstvideo-1.0 pulls in gstbase-1.0 as a dependency this worked but would cause problem as the linker may remove this dependency since no symbols from this particular library are being used. Linking directly to gstbase-1.0 seemed to solve all issues.

GCC math.h in fedora v/s Ubuntu

I am using sin function in one of my c program.
The program runs prefectly on fedora machine but ginving reference issue in ubuntu machine
I have compile it using -lm like following:
gcc -lm kepler.c -o a.out --working on fedora, but not on ubuntu
gcc kepler.c -lm -o a.out --Working on both, fedora and ubuntu
/tmp/ccshH33a.o: In function `kepler':
/tmp/28/kepler.c:7: undefined reference to `sin'
collect2: error: ld returned 1 exit status
So can any one explains about the position of -lm, and why its working on fedora machine for both cases?
Thanks in Adavance.
gcc like all common compilers internally launches different programs : at least a compiler and a linker. -lm option is intended for linker, as is -o a.out. BTW,-o a.outis just a no-op sincea.outis the default name when-o` option is absent.
Why -lm before kernel.c may fail : as explained in the answer to the question referenced by Cornstalks in comment, linker processes its arguments left to right, so dependant libraries must come after modules calling them
Why could it succeed on fedora : object modules will always be included in the executable program, so it can makes sense to always put them first in the argument list passed to the linker, of for the linker to process them first
Why did it work on Fedora and not on Ubuntu : I assume it is not same version of gcc, and the one on Fedora is more tolerant.
Anyway, the correct way is to always put dependant libraries after the modules (or other libraries) calling them.

Undefined Symbols when linking against a mixed C and Fortran in OS X 10.6.4

I'm trying to compile a code (not mine) that consists of mixed Fortran and C source files, which are compiled into a library. This library can either be linked against directly, or (more usefully) driven from a python class. I have previously successfully built the code as 32-bit with g77 and gcc, but I've encountered a situation in which the code uses big chunks of memory, and needs to be 64-bit.
I've attempted to build as both 64-bit only, or as a universal binary, with gfortran 4.2.3 (binary dist from the AT&T R project) and the system gcc (4.2). The source files build correctly, but when I attempt to link against the library, I get many "Undefined Symbols" errors for a number of the Fortran functions. An nm on the library shows that the symbols appear to exist, but obviously the linker isn't finding them.
Here are two (of many) of the compile commands (which produce no errors):
/usr/local/bin/gfortran -arch ppc -arch i386 -arch x86_64 -fPIC -fno-strength-reduce -fno-common -ff2c -Wall -c lsame.f
gcc -c -I/Users/keriksen/Research/atomic_data/fac -I/Users/keriksen/Research/atomic_data/fac/faclib -O2 -fPIC -fno-strength-reduce -fno-common pmalloc.c
And the link step, which bombs:
gcc -o sfac sfac.c stoken.c -I/Users/keriksen/Research/atomic_data/fac -I/Users/keriksen/Research/atomic_data/fac/faclib -O2 -fPIC -fno-strength-reduce -fno-common -L/Users/keriksen/Research/atomic_data/fac -lfac -lm -lgfortran -lgcc
A sample Undefined Symbol:
"_acofz1", referenced from:
_HydrogenicDipole in libfac.a(coulomb.o)
_HydrogenicDipole in libfac.a(coulomb.o)
and the corresponding nm that shows that symbol exists:
niobe:atomic_data/fac[14] nm libfac.a | grep acof
0000000000000000 T _acofz1_
0000000000002548 S _acofz1_.eh
U _acofz1
Am I doing something stupid, like not including a necessary switch to the linker, or is something more subtle going on here?
Per Yuji's suggestion:
The cfortran.h header file (available on the Web, and apparently fairly widely used) that handles the C/Fortran compatibility issues does not handle gfortran out of the box. I hacked it (incorrectly) to ignore this fact, and paid for my insolence.
The specific issue is that gfortran emits object code containing symbols with a trailing underscore, gcc does not.
The correct thing to do was to set the environment variable CPPFLAGS to -Df2cFortran. Once this macro is set, cfortran.h adds the necessary underscore to the symbols in the calling C functions' object code, and the linker is happy.
The other surprising thing in this particular instance was the configure script looks at the F77 environment variable for the name of the fortran compiler. Once I set F77 to "/usr/local/bin/gfortran -arch x86_64" the generated Makefile was correct, and it produced 64-bit only object code. I'm not sure if this is standard configure behavior, or if it a peculiarity of this particular script.
The upshot is that I now have a 64-bit shared library which plays nicely with a 64-bit python I downloaded. I'm half-considering going back and trying to produce a universal library that will work with the system python, but I'm not sure I want to tempt fate.

Resources