It's easy to let program figure out the dependency at compile time, (with gcc -MM). Nevertheless, link dependency (deciding which libraries should be linked to) seems to be difficult to figure out. This issue become emergent when multiple targets with individual libraries to link to are needed.
For instance, three dynamic library targets t1.so, t2.so and t3.so needs to be built. t1.so needs math library (-lm), while t2 and t3 don't. It would be tedious to write separate rules. A single rule requiring the three targets linked with math library saves the trouble. However, it causes inflation of target size since math library is unused for t2.so and t3.so.
Any ideas?
This is not as easy to figure out as finding needed headers. gcc -MM is just some fancy way to use the preprocessor, but it knows pretty much nothing about the way the code is used or works: you could include some headers full of #define's or introduce complex dependencies library dependencies.
I would stick with writing explicit linking dependencies for all targets (3 in your case). You can collect common dependencies in LDFLAGS.
It looks like ld's --trace option is a good start. The output needs formatting, but I think it contains all the right information.
My invocation looks something like this:
$ g++ -o foo a.o b.o -l sfml-graphics -l sfml-window -Wl,--trace
/usr/bin/ld: mode elf_i386
/usr/lib/gcc/i686-linux-gnu/4.6/../../../i386-linux-gnu/crt1.o
/usr/lib/gcc/i686-linux-gnu/4.6/../../../i386-linux-gnu/crti.o
/usr/lib/gcc/i686-linux-gnu/4.6/crtbegin.o
a.o
b.o
-lsfml-graphics (/usr/lib/gcc/i686-linux-gnu/4.6/../../../../lib/libsfml-graphics.so)
-lsfml-window (/usr/lib/gcc/i686-linux-gnu/4.6/../../../../lib/libsfml-window.so)
-lstdc++ (/usr/lib/gcc/i686-linux-gnu/4.6/libstdc++.so)
-lm (/usr/lib/gcc/i686-linux-gnu/4.6/../../../i386-linux-gnu/libm.so)
-lgcc_s (/usr/lib/gcc/i686-linux-gnu/4.6/libgcc_s.so)
/lib/i386-linux-gnu/libc.so.6
(/usr/lib/i386-linux-gnu/libc_nonshared.a)elf-init.oS
/lib/i386-linux-gnu/ld-linux.so.2
-lgcc_s (/usr/lib/gcc/i686-linux-gnu/4.6/libgcc_s.so)
/usr/lib/gcc/i686-linux-gnu/4.6/crtend.o
/usr/lib/gcc/i686-linux-gnu/4.6/../../../i386-linux-gnu/crtn.o
Have you tried using 'nm'? It gives you a list of defined and undefined symbols in object/library files (see documentation here.
There's an approach mentioned in this post by Bernd Strieder that I am considering using -
1. Use nm to generate a list of symbols in all object/library files involved.
2. This file is parsed and basically the (U)ndefined and (T)ext symbols
and the symbols of main functions are filtered out and mapped to their
object files. I found that U and T symbols suffice, which reduces the
overall problem considerably compared to the linker, which has to
consider all symbols.
3. The transitive hull of the dependency relation according to U and T
symbols between object files is being calculated.
4. A list of object files needed to resolve all dependencies can be
printed for any object file.
5. For any main object file, a make target to link it is arranged.
Related
I'm attempting to do a release of some software and am currently working through a script for the build process. I'm stuck on something I never thought I would be, statically linking LAPACK on x86_64 linux. During configuration AC_SEARCH_LIB([main],[lapack]) works, but compilation of the lapack units do not work, for example undefiend reference to 'dsyev_' --no lapack/blas routine goes unnoticed.
I've confirmed I have the libraries installed and even compiled them myself with the appropriate options to make them static with the same results.
Here is an example I had used in my first experience with LAPACK a few years ago that works dynamically, but not statically: http://pastebin.com/cMm3wcwF
The two methods I'm using to compile are the following,
gcc -llapack -o eigen eigen.c
gcc -static -llapack -o eigen eigen.c
Your linking order is wrong. Link libraries after the code that requires them, not before. Like this:
gcc -o eigen eigen.c -llapack
gcc -static -o eigen eigen.c -llapack
That should resolve the linkage problems.
To answer the subsequent question why this works, the GNU ld documentation say this:
It makes a difference where in the command you write this option; the
linker searches and processes libraries and object files in the order
they are specified. Thus, foo.o -lz bar.o' searches libraryz' after
file foo.o but before bar.o. If bar.o refers to functions in `z',
those functions may not be loaded.
........
Normally the files found this way are library files—archive files
whose members are object files. The linker handles an archive file by
scanning through it for members which define symbols that have so far
been referenced but not defined. But if the file that is found is an
ordinary object file, it is linked in the usual fashion.
ie. the linker is going to make one pass through a file looking for unresolved symbols, and it follows files in the order you provide them (ie. "left to right"). If you have not yet specified a dependency when a file is read, the linker will not be able to satisfy the dependency. Every object in the link list is parsed only once.
Note also that GNU ld can do reordering in cases where circular dependencies are detected when linking shared libraries or object files. But static libraries are only parsed for unknown symbols once.
I have a C project using several object files that needs to be linked in a specific order to find all needed symbols.
For example this command works fine (lib2.o depends on lib1.o etc.)
gcc -o my_app main.o lib1.o lib2.o lib3.o -lm
but
gcc -o my_app main.o lib3.o lib2.o lib1.o -lm
ends with undefined reference to `my_variable' errors.
This is a known behavior and can be solved for example by adding these objects to GROUP section in a linker script.
Now I'd like to share these object as a static library with my colleagues. So...
ar -rcs mylib.a lib1.o lib2.o lib3.o
gcc -o my_app main.o mylib.a -lm
Unfortunately this gives the same undefined reference errors like the specifying the objects in incorrect order.
I have not found any linker or archiver options to make it working and also no solution by googling even if I think that this problem should be relatively common.
Do please anybody know a solution?
regards
Jan
This might be a link order problem. When the GNU linker sees a library, it discards all symbols that it doesn't need. It also does that in the sequential order form left to right.
Recent versions of gcc/ld default to linking with --as-needed flag.
This means if you write -lmylib.a before the C file the library will automatically get excluded (the order matters when testing if things are "needed" like this)
You can fix this with either:
gcc -L. -o example example.c -lmylib.a
gcc -L. -Wl,--no-as-needed -o example example.c -lmylib.a
The latter of which passes --no-as-needed to the linker, which would cause the library to still be linked, even if you didn't call any function external from it.
Your error implies that the problem is in one of your lib?.o files [lib{later}.o depends on lib{earlier}.o]
How did you manage to compile them?
Were there any compilation warnings?
It has been a while since I used C, but I think that you will need to include dependent libraries within the library that has the dependency - this may be the reason why you can't find too many references to the problem, because it does not really exist.
I am trying to link a C++ module using GCC, essentially like this:
gcc -c hello.c
g++ -c world.cpp
gcc -ohello -lstdc++ hello.o world.o
Note that I use -lstdc++ to link the C++ module in, so that I can use gcc instead of g++. The problem is that I'm getting the error:
undefined reference to `operator new(unsigned long)'
(Assuming that world.cpp contains at least one call to new.)
This error is fixed if I put -lstdc++ at the end of the linker line, like this:
gcc -ohello hello.o world.o -lstdc++
I am aware that this question has been asked many times here, but I have a special requirement. I am not directly calling GCC. I am using a build system for a different programming language (Mercury) which is calling GCC on my behalf, and I can't easily modify the way it calls GCC (though I can specify additional libraries using the LDFLAGS environment variable). So I have two additional requirements:
I cannot use g++ to link (only gcc) -- that is why I am doing the -lstdc++ trick above rather than simply linking with g++).
I don't think that I can control the order of the linker commands -- Mercury will put the .o files on the command-line after any libraries.
I understand the basic reason why the order is important, but what is baffling me is why did this break now? I just updated to Ubuntu 11.10 / GCC 4.6.1. I have been successfully compiling this program for years using precisely the above technique (putting -lstdc++ first). Only now has this error come up. An unrelated program of mine links against OpenGL using -lgl and that too broke when I upgraded and I had to move -lgl to the end of the command-line. I'm probably going to discover that dozens of my programs no longer compile. Why did this change? Is there something wrong with my new system or is that the way it is now? Note that these are ordinary shared libraries, not statically linked.
Is there anything I can do to make GCC go back to the old way, where the order of libraries doesn't matter? Is there any other way I can convince GCC to link libstdc++ properly without moving it after the .o files on the command-line?
If Mercury puts object files after libraries, Mercury is broken. Libraries belong after object files - always. You may sometimes get away with the reverse order, but not reliably. (Static libraries must go after the object files that reference symbols in the static library. Sometimes, a linker will note the symbols defined by a shared library even when none of the symbols are used; sometimes, the linker will only note the shared library symbols if the shared library provides at least one symbol.)
I'm porting an existing system from Windows to Linux. The build is structured with multiple static libraries. I ran into a linking error where a symbol (defined in libA) could not be found in an object from libB. The linker line looked like
g++ test_obj.o -lA -lB -o test
The problem of course being that by the time the linker finds it needs the symbol from libA, it has already passed it by, and does not rescan, so it simply errors out even though the symbol is there for the taking.
My initial idea was of course to simply swap the link (to -lB -lA) so that libA is scanned afterwards, and any symbols missing from libB that are in libA are picked up. But then I find there is actually a recursive dependency between libA and libB! I'm assuming the Visual C++ linker handles this in some way (does it rescan by default?).
Ways of dealing with this I've considered:
Use shared objects. Unfortunately this is undesirable from the perspective of requiring PIC compliation (this is performance sensitive code and losing %ebx to hold the GOT would really hurt), and shared objects aren't needed.
Build one mega ar of all of the objects, avoiding the problem.
Restructure the code to avoid the recursive dependency (which is obviously the Right Thing to do, but I'm trying to do this port with minimal changes).
Do you have other ideas to deal with this? Is there some way I can convince the binutils linker to perform rescans of libraries it has already looked at when it is missing a symbol?
Just do this:
g++ test_obj.o -lA -lB -lA -o test
When the linker reads the first libA on the command line, it'll discard the object/symbols in it that noone has depended on yet, e.g. all the symbols libB needs but not test_obj.o. So you just make it read libA again, and it'll pick up the those symbols as well.
While #nos provides a simple solution, it doesn't scale when there are multiple libraries involved and the mutual dependencies are more complex. To sort out the problems ld provides --start-group archives --end-group.
In your particular case:
g++ test_obj.o --start-group -lA -lB --end-group -o test
I'm having a problem with my compiler telling me there is an 'undefined reference to' a function I want to use in a library. Let me share some info on the problem:
I'm cross compiling with gcc for C.
I am calling a library function which is accessed through an included header which includes another header, which contains the prototype.
I have included the headers directory using -I and i'm sure it's being found.
I'm first creating the .o files then linking them in a separate command.
So my thought is it might be the order in which I include the library files, but i'm not sure what is the correct way to order them. I tried with including the headers folder both before and after the .o file.
Some suggests would be great, and maybe and explanation of how the linker does its thing.
Thanks!
Response to answers
there is no .a library file, just .h and .c in the library, so -l isn't appropriate
my understanding of a library file is that it is just a collection of header and source files, but maybe it's a collection of .o files created from the source?!
there is no library object file being created, maybe there should be?? Yes seems I don't understand the difference between includes and libraries...i'll work on that :-)
Thanks for all the responses! I learned a lot about libraries. I'd like to put all the responses as the accepted answer :-)
Headers provide function declarations and function definitions. To allow the linker find the function's implementation (and get rid of the undefined reference) you need to ask the compiler driver (gcc) to link the specific library where the function resides using the -l flag. For instance, -lm will link the math library. A function's manual page typically specifies what library, if any, must be specified to find the function.
If the linker can't find a specified library you can add a library search path using the -L switch (for example, -L/usr/local/lib). You can also permanently affect the library path through the LIBRARY_PATH environment variable.
Here are some additional details to help you debug your problem. By convention the names of library files are prefixed with lib and (in their static form) have a .a extension. Thus, the statically linked version of the system's default math library (the one you link with -lm) typically resides in /usr/lib/libm.a. To see what symbols a given library defines you can run nm --defined-only on the library file. On my system, running the command on libm.a gives me output like the following.
e_atan2.o:
00000000 T atan2
e_asinf.o:
00000000 T asinf
e_asin.o:
00000000 T asin
To see the library path that your compiler uses and which libraries it loads by default you can invoke gcc with the -v option. Again on my system this gives the following output.
GNU assembler version 2.15 [FreeBSD] 2004-05-23 (i386-obrien-freebsd)
using BFD version 2.15 [FreeBSD] 2004-05-23
/usr/bin/ld -V -dynamic-linker /libexec/ld-elf.so.1 /usr/lib/crt1.o
/usr/lib/crti.o /usr/lib/crtbegin.o -L/usr/lib /var/tmp//ccIxJczl.o -lgcc -lc
-lgcc /usr/lib/crtend.o /usr/lib/crtn.o
It sounds like you are not compiling the .c file in the library to produce a .o file. The linker would look for the prototype's implementation in the .o file produced by compiling the library
Does your build process compile the library .c file?
Why do you call it a "library" if it's actually just source code?
I fear you mixed the library and header concepts.
Let's say you have a library libmylib.a that contains the function myfunc() and a corresponding header mylib.h that defines its prototype. In your source file myapp.c you include the header, either directly or including another header that includes it. For example:
/* myapp.h
** Here I will include and define my stuff
*/
...
#include "mylib.h"
...
your source file looks like:
/* myapp.c
** Here is my real code
*/
...
#include "myapp.h"
...
/* Here I can use the function */
myfunc(3,"XYZ");
Now you can compile it to obtain myapp.o:
gcc -c -I../mylib/includes myapp.c
Note that the -I just tells gcc where the headers files are, they have nothing to do with the library itself!
Now you can link your application with the real library:
gcc -o myapp -L../mylib/libs myapp.o -lmylib
Note that the -L switch tells gcc where the library is, and the -l tells it to link your code to the library.
If you don't do this last step, you may encounter the problem you described.
There might be other more complex cases but from your question, I hope this would be enough to solve your problem.
Post your makefile, and the library function you are trying to call. Even simple gcc makefiles usually have a line like this:
LIBFLAGS =-lc -lpthread -lrt -lstdc++ -lShared -L../shared
In this case, it means link the standard C library, among others
I guess you have to add the path where the linker can find the libraray. In gcc/ld you can do this with -L and libraray with -l.
-Ldir, --library-path=dir
Search directory dir before standard
search directories (this option must
precede the -l option that searches
that directory).
-larch, --library=archive
Include the archive file arch in the
list of files to link.
Response to answers - there is no .a library file, just .h and .c in the library, so -l isn't approriate
Then you may have to create the libraray first?
gcc -c mylib.c -o mylib.o
ar rcs libmylib.a mylib.o
I have encountered this problem when building a program with a new version of gcc. The problem was fixed by calling gcc with the -std=gnu89 option. Apparently this was due to inline function declarations. I have found this solution at https://gcc.gnu.org/gcc-5/porting_to.html