I am building a shared lib on Ubuntu 18.04.1 that has to be linked explicitly with libdl.so.
All is good when using a non-libtool makefile - "-ldl" is supplied to the linker and the produced binary imports libdl.so.
But when trying to build with a libtooled automake, the produced shared library does not import libdl.so despite fact that Makefile.am has la_LDFLAGS set to "-Wl,--no-as-needed -ldl"
I suspect that this is due to the fact that libtool comes with its own implementation of dl (libltdl) and it is linked in statically instead of libdl.so
Is there any way, a command line option perhaps, that will enforce libtool to use the native implementation instead of libltdl? - I have spent a lot of time finding it in vain.
The libtool version is 2.4.6
Thanks
Related
I'm trying to build a framework with Autotools on MacOS. This frameworks contains a library, let's call it libmytools.dylib, and an executable that uses the library.
What I want is to link this library dynamically at runtime.
The executable is usually installed to /Library/Frameworks/mytools.Framework/Versions/Current/Commands/mytools
The library is usually installed to
/Library/Frameworks/mytools.Framework/Versions/Current/Libraries/libmytools.dylib
In case the framework is installed in that location, everything works fine, but when the framework is not installed on the machine but instead it's only embedded in another project, the library can't be found.
What I need is a so called "Runpath Dependent Library" as described here:
https://developer.apple.com/library/content/documentation/DeveloperTools/Conceptual/DynamicLibraries/100-Articles/RunpathDependentLibraries.html
So what I did was adding mytools_LDFLAGS = -rpath #executable_path/../Libraries to the makefile.
But unfortunately libtool refuses to link my executable with the dynamic -rpath
and shows the following error: libtool: link: only absolute run-paths are allowed.
I also tried mytools_LDFLAGS = -dynamic -rpath #executable_path/../Libraries but the result is the same...
What am I doing wrong??
This is the current libtool design. This thread from the libtool patches list may have a suggestion which will work for you. The suggestion is:
build as normal and then postprocess using install_name_tool
If you don't like this suggestion, it may be possible to adapt the hacky patch that started off that thread to work for you, and apply the patch to the generated libtool as part of bootstrap.sh configuration.
I downloaded tar 1.28 and am attempting to compile a static binary to use on system recovery media on Fedora Linux (x86_64). It doesn't have to include common libraries, such as libc or pthreads, but I'd like to eliminate dependencies on selinux, acl, attr, pcre and lzma, etc.
The README indicates I can use
./configure CC=gcc LDFLAGS=-static
but this creates a fully static binary (including libc), but is missing support for anything normally found in the shared libraries (SELinux, ACLs, etc). Reading the config.log shows failed attempts to find a static library for acl
So, I installed libselinux-devel-static RPM, and this time got the selinux support in there, but I can't find static libraries for anything else. Search results say that libacl.a is in libacl-devel, but not mine.
A fully static binary is ok, but really just want to get the odd-balls in there so I don't have to put the shared libraries on the media.
I was looking at this as well, and i would get the following error message:
CCLD tar
../gnu/libgnu.a(quotearg.o): In function `quote':
/home/anarcat/dist/tar-1.27.1/gnu/quotearg.c:968: multiple definition of `quote'
/usr/lib/gcc/x86_64-linux-gnu/4.9/../../../../lib/libacl.a(quote.o):(.text+0x0): first defined here
names.o: In function `gid_to_gname':
/home/anarcat/dist/tar-1.27.1/src/names.c:95: warning: Using 'getgrgid' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
names.o: In function `gname_to_gid':
/home/anarcat/dist/tar-1.27.1/src/names.c:155: warning: Using 'getgrnam' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
names.o: In function `uname_to_uid':
/home/anarcat/dist/tar-1.27.1/src/names.c:125: warning: Using 'getpwnam' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
names.o: In function `uid_to_uname':
/home/anarcat/dist/tar-1.27.1/src/names.c:65: warning: Using 'getpwuid' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
../lib/libtar.a(rtapelib.o): In function `rmt_open__':
/home/anarcat/dist/tar-1.27.1/lib/rtapelib.c:428: warning: Using 'gethostbyname' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
collect2: error: ld returned 1 exit status
Makefile:1304: recipe for target 'tar' failed
I don't quite get it, but apparently, "You cannot getpwnam and friends with the whole flexibility that glibc has without shared libraries".
So game over? unclear.
For Fedora you can enable Repository CERT Forensics Tools
and complete build dependencies with some static libs (non-exhaustive list)
zlib-static
libiconv-static
libffi-devel
gettext-devel
pcre-static
ncurses-static
nss-devel
glibc-static
libselinux-static
libsepol-static
libattr-devel
libacl-devel
if libacl.a is missing you have to download source first from:
git clone https://git.savannah.nongnu.org/git/attr.git
git clone https://git.savannah.nongnu.org/git/acl.git
and compile with (first libattr, then libacl, then tar):
./autogen.sh
./configure --prefix=/usr --libexecdir=/usr/lib --enable-static --disable-shared --disable-rpath
make CFLAGS='-static'
make install
I'm having trouble building the hmatrix library on OS X Lion. Looking at the .cabal file, it requires the gsl library, so I installed it with macports. The .a files are in /opt/local/lib and the .h files are in /opt/local/include/gsl
As suggested here I changed the built-type from Custom to Simple. (without that change I get a similar error).
When I use cabal configure I get the following output:
* Missing C library: gsl
This problem can usually be solved by installing the system package that
provides this library (you may need the "-dev" version). If the library is
already installed but in a non-standard location then you can use the flags
--extra-include-dirs= and --extra-lib-dirs= to specify where it is.
So I tried cabal --extra-include-dirs=/opt/local/include --extra-lib-dirs=/opt/local/lib configure, but I still get the same error. I can compile and link a c program that includes gsl. What files is cabal looking for? If I have the right files, how do I tell it how to find them?
libgsl.a is a universal binary:
$ file /opt/local/lib/libgsl.a
/opt/local/lib/libgsl.a: Mach-O universal binary with 2 architectures
/opt/local/lib/libgsl.a (for architecture x86_64): current ar archive random library
/opt/local/lib/libgsl.a (for architecture i386): current ar archive random library
ghc looks like it's 64-bit:
$ ghc --info
[("Project name","The Glorious Glasgow Haskell Compilation System")
,("GCC extra via C opts"," -fwrapv")
,("C compiler command","/usr/bin/llvm-gcc")
,("C compiler flags"," -m64 -fno-stack-protector -m64")
,("ar command","/usr/bin/ar")
,("ar flags","clqs")
,("ar supports at file","NO")
,("touch command","touch")
,("dllwrap command","/bin/false")
,("windres command","/bin/false")
,("perl command","/usr/bin/perl")
,("target os","OSDarwin")
,("target arch","ArchX86_64")
,("target word size","8")
,("target has GNU nonexec stack","False")
,("target has subsections via symbols","True")
,("Project version","7.4.2")
,("Booter version","7.4.2")
,("Stage","2")
,("Build platform","x86_64-apple-darwin")
,("Host platform","x86_64-apple-darwin")
,("Target platform","x86_64-apple-darwin")
,("Have interpreter","YES")
,("Object splitting supported","NO")
,("Have native code generator","YES")
,("Support SMP","YES")
,("Unregisterised","NO")
,("Tables next to code","YES")
,("RTS ways","l debug thr thr_debug thr_l thr_p dyn debug_dyn thr_dyn thr_debug_dyn")
,("Leading underscore","YES")
,("Debug on","False")
,("LibDir","/usr/local/Cellar/ghc/7.4.2/lib/ghc-7.4.2")
,("Global Package DB","/usr/local/Cellar/ghc/7.4.2/lib/ghc-7.4.2/package.conf.d")
,("Gcc Linker flags","[\"-m64\"]")
,("Ld Linker flags","[\"-arch\",\"x86_64\"]")
]
As an alternative to mac-ports you can use the nix package manager for mac. It does a pretty good job of taking care of the c dependancies for for the libraries available through it. In general I have been more happy with it then any other package manager on mac.
Unfortunately mac(darwin) unlike for linux does not have as many binaries available through nix so installing ghc often means waiting for it to compile.
The commands to install ghc and hmatrix after installation of nix are:
nix-env -iA nixpkgs-unstable.haskellPackages.ghc
nix-env -iA nixpkgs-unstable.haskellPackages.hmatrix
All of the needed dependencies will be taken care of for you.
I just tried it on my macbook pro and hmatrix seems to be working correctly in ghci after trying commands from the first few pages of the tutorial.
I'm not a mac person, but it really sounds like you haven't installed the "-dev" version. For a mac, I suspect you need to install gsl-devel in addition to gsl. If the problem persists, verify that you have libgsl0-dev on your library path.
I have a C++ shared library being called from my Python program. The C++ lib is compiled with HDF5 which I installed using homebrew on OS X, so it resides in /usr/local/lib. Now the problem is that I also have PyTables installed, which includes an older version of libhdf5, installed somewhere in /Library/Frameworks/EPD64.framework/Versions/Current/...
I can compile and link my library just fine, and otool -L indeed points to /usr/local/lib/libhdf5.dylib. However, when I try to run it from Python, there is a version mismatch error. Somehow the runtime linker is loading the older hdf5 library instead.
One workaround is to build hdf5 as a static library instead, but I'd like to find a neater solution - how can I persuade the runtime linker to use the newer library? I already tried setting DYLD_LIBRARY_PATH but that just broke everything else (Python and MacVim wouldn't start).
Can you provide more info? I have a few alternatives (and questions).
If I understood correctly, you have a library that uses lib_ver_a.dyld and pyTables that uses lib_ver_b.dyld. I.e. the same library but different version.
You could either:
Compile your library to use the same version as pyTables (DYLD_LIB... while you compile) and run your program with pyTables library: DYLD_LIBRARY_PATH=/correct/path/lib python myprog.py
Upgrade your xcode (which version you have? also which osx version?). This upgrades your libraries and pyTables (be careful, newest is not necessary the most stable)
You should also check which python you are calling (a custom one or apple one).
There is also a possibility to define a specific dyld while linking (Do you use gcc or ld for linking? I might miss some options here):
gcc /path/lib1.dyld myLib.c -o myLib.o
This way your library is linked to certain version of the library (caution! I don't remember the exact details, but I can dig them out).
I hope this helps.
br,
Juha
I am trying to compile a package on ubuntu 8.1
when executing this command: ./configure I get the follwoing error:
checking for Boost headers version >= 103700... no
configure: error: cannot find Boost headers version >= 103700
knowing that I installed needed boost packages using these command:
$ apt-get install libboost-dev libboost-graph-dev libboost-iostreams-dev
Can anybody help please?
thank you. Now it works but i get another error when running ./configure: checking boost/iostreams/device/file_descriptor.hpp usability... yes checking boost/iostreams/device/file_descriptor.hpp presence... yes checking for boost/iostreams/device/file_descriptor.hpp... yes checking for the Boost iostreams library... no configure: error: cannot not find the flags to link with Boost iostreams any ideas please?
It could be that the version of boost that you're getting from the Ubuntu repository is too old (it's suggested here that the highest version for 8.10 is 1.35; it looks like your configure script is asking for 1.37). You might need to build from source; there's some more info in the answers to the question I linked to which will hopefully help.
UPDATE:
From your new error, it sounds like configure now can't find the boost_iostreams library. On my system it's /usr/lib/libboost_iostreams-mt.[a|so] - do you have those files (possibly in a different directory depending on where you installed boost)?
You can also try running ldconfig in case there's a missing symlink (from, say,
libboost_iostreams-mt.so.1.37.0 to libboost_iostreams-mt.so).
Is this configure one generated by GNU autoconf? If it is, there should be a file called config.log in the same directory which contains a list of all the commands configure tried to run when looking for things. If there's anything in there about boost_iostreams could you post it?
One totally random guess: some examples I've found on the web link to boost_iostreams without the multi-threading suffix -mt - but I don't have those on my machine at all. Maybe your configure script is running into the same problem?
UPDATE 2
The configure script seems to be looking for a single-threaded debug build of the boost iostreams library, which won't be produced by default when building from source on linux. Also, the default on linux is not to name the libraries based on the build configuration (so the libs you found in /usr/lib might not be the ones you installed from source unless you overrode this). This stuff isn't really explained on the boost website, I only found out by looking in the Jamroot file (bjam --help works too)! Anyway, to get a library with the right build configuration, and named correctly, I need to go into the root of the boost source tree and run:
sudo bjam --with-iostreams --layout=tagged variant=debug threading=single install
For me this puts the libraries (libboost_iostreams-d.a and the shared versions) into /usr/local/lib where ld will find them by default, so this should be fine. If you need them to go somewhere else you can use the --prefix=... option to bjam eg. if you want them in /usr/lib you can do --prefix=/usr. If the package you're building needs more boost libraries you can remove the --with-iostreams and then they'll all be built (or replace iostream with the name of each other library you need).
A side note: I had to install the libbz2-dev package to get boost iostreams to build - it's easy to miss the error here if you build all of boost as there's so much output!