We have a application with various dynamic libs, which all use Boost 1.48 (static). Due to a third-party dependency on Boost Filesystem v2, we can't switch to a more recent Boost version for the application.
For a new functionality we develop a new dynamic lib for the application, which should also be used in some other projects.
Can we use a recent (static?) Boost lib for this new lib, without interfering the Boost lib already used in the application?
Any traps I should avoid?
Boost libraries generally do not support mixing different versions of libraries. Whether doing so would cause problems or not depends on many factors, among which are the libraries in question, your application design and the target platform. One source of problems can be symbol relocation which is part of the linking process on Linux and other UNIX-like systems. Even if you link with static libs of Boost, the linked symbols can still be exported from your binaries and may clash when you load your application. This can cause all sorts of undefined behavior and is often very difficult to debug.
In general, I would highly discourage from mixing different releases of Boost in the same application (i.e. a runtime process).
If you are using a static boost library to build your dynamic library, the static boost library will not create side effect with another boost static or dynamic library.
Related
I'm currently developing a shared library on Windows (dll) with c++.
The library depends on another external library.
What is the best way to link them together?
Link the external library as a static library into my shared lib?
Link it as a shared library and provide the dll to the application who is using my shared lib?
For the second case what happens if i create an application which uses my own created shared library and also the external library as a shared library?
If for example my shared library is build with the external library version 1.1 and the application uses the newer Version for example 1.3 ? Now the dlls should be different but how could i provide them to the main application?
Are there some best practices or recommendations on how to handle such a situation?
This really depends on what you want to do and how you are deploying your library.
shared/dynamic libraries (dll's on windows) have several advantages over static libraries
They can be distributed externally from the application, allowing your application binary to be smaller
They can be updated externally to the application
They are slightly more efficient as code is only executed if its needed rather than being bundled with the executable
Of course they have several weaknesses too
They can be distributed externally from your application and updated externally - allowing dll injection attacks
Trying to release dynamic libraries with the executable is one of the most painful and horrible things to do (especially on windows where there is no RPATH)
You may have to use a dll, dependent on your external libraries licensing, Qt for example requires shared library linking in many cases (not all).
Standard convention is usually to offer shared and static versions of your library with the shared version being completely linked to other shared libraries and the static version being an ar static library (includes all dependencies). The shared library variant then offers instructions on linking (i.e a .pc (pkgconfig) file) which specifies the versions of the other libraries to link to (i.e v1.1 of x.dll) in order to successfully compile/link.
I'm developing a .so shared library in C++ that uses standard library facilities (e.g. std::map or std::vector).
I had problems in the past though since any application pulling my shared library caused standard libraries symbols collisions and debugging quickly became a nightmare.
I'm a novice on linux, how can I avoid this problem and/or mitigate it if I have no control on the application which loads my .so?
In short, link your shared library statically against the C++ standard library. Find more details in libstdc++ static linking in dynamic library.
I have a very basic client/server project that uses boost::asio. It generates two executables, a client and a server.
When I run the client, I get the following:
./client: error while loading shared libraries:
libboost_system.so.1.55.0: cannot open shared object
file: No such file or directory
This means that the program requires the boost_system binary to be loaded dynamically at run-time. This makes sense, as one dependency of boost_asio is boost_system.
What does this mean for the ease of distributing my application to end-users?
1) Do I simply pop my development version of the boost_system binary on my system, which in this case is libboost_system.so.1.55.0? How do I ensure that when the user runs the client, it will find the dynamic archive? Obviously, on my system, even with my boost install it still didn't find the archive.
2) I am building on Linux and thus I have .so binaries. How will #1 change if I try to cross-compile my app for Windows with mingw-w64?
I am brand-spanking new to distributing C++ programs and working with dynamic/shared libraries.
When I compile statically, I get the following warning:
Using 'getaddrinfo' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
Suggestion:
1) If you use shared libraries, you'll definitely need to include those libraries your program actually uses alone with your executable.
2) Here is a list of the Boost libraries. Your program will require just a subset:
http://www.boost.org/doc/libs/1_49_0/more/getting_started/unix-variants.html
The only Boost libraries that must be built separately are:
Boost.Filesystem
Boost.GraphParallel
Boost.IOStreams
Boost.MPI
Boost.ProgramOptions
Boost.Python (see the Boost.Python build documentation before building and installing it)
Boost.Regex
Boost.Serialization
Boost.Signals
Boost.System
Boost.Thread
Boost.Wave
A few libraries have optional separately-compiled binaries:
Boost.DateTime has a binary component that is only needed if you're using its to_string/from_string or serialization features, or if you're targeting Visual C++ 6.x or Borland.
Boost.Graph also has a binary component that is only needed if you intend to parse GraphViz files. * Boost.Math has binary components for the TR1 and C99 cmath functions.
Boost.Random has a binary component which is only needed if you're using random_device.
Boost.Test can be used in “header-only” or “separately compiled” mode, although separate compilation is recommended for serious use.
Alternatively, you can link your program with static (.a) Boost libraries instead of shared (.so), in which case there will be NO runtime dependencies.
Or you can mix/match shared/statis as you wish.
The choice is yours.
Look at the Boost documentation: b2 Static and Shared libraries
As said, you need to compile boost with the static option, for example
bjam install --toolset=msvc variant=release link=static threading=multi runtime-link=static
You can have more information in this Thread
Do i have static or dynamic boost libraries?
Something to notice, if you do an ldd on your executable, you'll probably notice some runtime dependencies on gcc/libc libraries, even if you compile it in static mode.
That means your client platform has to have those libraries installed. 90% of the time they're there, but it might be more complicated when you compile with the latest version of the compiler and the client has an older one.
I am building "boost" libraries from boost source code and I have two options: to build it "static" or to build it "shared" (e.g. dynamic). Which is better idea?
I prefer dynamic (shared) linking but when I tried to build boost shared libraries (on Ubuntu Linux), I got lots of errors or warnings (why there are always errors, warning, notes and other stuff when compiling, grrrrrrrr), so I don't know if it was compiled alright?
Thanks.
Better is subjective. Shared cuts down on size, at the risk of dependencies. Static solves dependency issues but increases the size.
For your purposes, I'd say building it in which ever way gets you to code faster is the better solution.
You will almost always want to use shared libraries over static libraries. A key advantage to using shared libraries is that if the library is updated, you can replace the shared libraries with the newer version (assuming binary compatibility) and reap the benefits of the improved implementation without recompiling your application. Additionally, using shared libraries saves space, if multiple programs are using them.
As for the dependencies issue, it is possible to link against a specific version of a shared library, or to place your shared libraries in a special location that is specific to your program -- which doesn't save you space, but which does give you the flexibility associated with shared libraries -- so that should not be a reason to choose static libraries over shared libraries. I am actually hard pressed to come up with a single instance, on a typical desktop, laptop, or server machine where using static libraries is better than using shared libraries.
P.S. If you are trying to install Boost on Ubuntu Linux, just run "sudo apt-get install libboost1.37-dev". You were probably getting errors because you did not install all of Boost's dependencies. These are automatically downloaded and installed when you use Ubuntu's apt-get package manager to install it. Also, it is generally better to use an OSs package manager for installing software packages, than to build from source. For example, using the package system's version of Boost will make it more likely that your software will run smoothly on other Ubuntu Linux deployments which use the package manager's version of Boost.
P.P.S. Boost uses some very advanced features of C++. It kind of pushes C++ to the limit. It is not uncommon to see warnings when compiling Boost. In fact, I have built Boost quite a number of times on various operating systems, and I don't recall a time when there were no warnings.
Static libraries are used when you don't need to dynamically load a componenet into the program. It is compiled into the exe.
A shared library is loaded on runtime, and is usually used for plugins or extentions.
Imo a static library is better here since you will probably load the boost shared library anyway on the program's startup.
Why do you prefer a shared library?
The recommended way to use the Boost C++ libraries on Linux is via shared linking. On an Ubuntu Linux box already configured for development you should not get any errors at all. Compilation warnings are expected -- for various mindset, technical, and time-constraint issues there are a few produced. Since regular release testing covers Ubuntu, I would not worry about functionality of created libraries -- if there's .so, it should work.
How do you construct a library (static lib or a dll/so) so that it isn't sensitive to future updates to the system's C runtime librarires?
At the end of July, Microsoft updated a bunch of libraries, including the C runtime libraries. Our app is written with a mix of MFC/C++/VB and some third party libraries, including some that are closed source.
I've been busy recompiling all of the libraries we have source to, but I am wondering if it is really necessary? What will happen if we link in or load a library built against an earlier version of the C runtime?
When recompiling this stuff, what compiler and linker settings must be the same between the main application and the supporting libraries? I've discovered that the runtime library setting needs to be the same (we use the multi-threaded version /MD and /MDd) but I'm worried about other settings. I've actually pulled all the settings out into Visual Studio property sheets and I'm using the same sheets for all our different projects, but this doesn't work for 3rd party libraries and I'm thinking it is overkill.
I have noticed that the linker will spit out a warning about conflicting libraries, but it suggests to just ignore the default libraries. Is it safe to do so? It seems like a very ugly solution to the problem.
If you are loading the 3rd party libraries as DLLs, they may depend on different runtime versions than your executable as long as
you are not handing over parameters of types, that depend on the runtime libs (like STL types)
the 3rd party lib is able to load the version of the runtime, that it has been built with or is statically linked to the runtime
So, you don't have to recompile the DLLs.
If you are statically
linking to the libs or if you are handing over types defined in the runtime DLLs, you may get some problems with symbols, that are already imported in your lib, so most likely you will have to recompile it.
If you or your third-party components are statically linking against out-of-date C libraries, you are fine; upgrades to those libraries will not affect your program. Of course, you won't benefit from any bug fixes or performance upgrades or what-have-you either. If you do recompile your code to take advantage of your new settings, all runtime switches must be the same. This is always the case, regardless of when or why your libraries are compiled. It is probably safe to ignore the default libraries (I've been doing so for years without difficulty).
Dynamically-linked libraries are another story. If you rely on the target system having a particular version of a given dll, and it has some other incompatible version instead, then you are screwed. The traditional solution to this problem is to bundle all the dlls you need with your executable. Microsoft's new side-by-side assembly thing might also be able to help, but it's always been a little too hard to set up for me to bother with it. You might have better luck.