There are several ways to obtain a working installation of Hamlib. In the following sections discuss installing from a package manager, building from source, and installing Hamlib project supplied binaries on Microsoft Windows@registeredsymbol{}. @menu * Unix binary packages:: * Source options:: * Building from source:: * MS Windows 32 binaries:: @end menu @node Unix binary packages @section Installing binary packages on Linux and BSD @cindex Binary packages, Linux, BSD @cindex Linux binary packages @cindex BSD binary packages The easiest way to install a released version of Hamlib on a Linux based distribution or a BSD variant is through the provided @dfn{package manager}. While package managers vary according to the distribution (it's easy to lump BSD variants in this group too) their end goal is to provide ready to use software packages. Since such a wide variety of package managers exist, it is best to recommend that the documentation for your chosen distribution be your guide. @node Source options @section A variety of Hamlib sources @cindex Source options Distribution packages are most often official Hamlib releases and in some cases could be quite old and lacking support for newer radios or rotors. In some cases support is improved in existing radio or rotor back ends and bugs are fixed in newer releases. Often times to get the improved support/bug fixes, building from source will be required. Relax, it's not hard. :-) Source code is available as official releases, testing snapshots, daily development snapshots, and the bleeding edge of development directly from the @url{http://git-scm.com/, Git} repository. As a rule, even the bleeding edge tarballs should configure and compile without error even though certain implementation work may be in progress and may be incomplete or have errors. @menu * Source releases:: * Source snapshots:: * Git clone:: @end menu @node Source releases @subsection Getting released source @cindex Getting released source @cindex Source, getting released @cindex Source, obtaining releases Official Hamlib source releases, commonly called @dfn{tarballs} can be found on the @url{http://sourceforge.net/projects/hamlib/files/hamlib/, SourceForge.net Hamlib files} Web page. The most recent release is listed first. @node Source snapshots @subsection Getting source snapshots @cindex Getting source snapshots @cindex Source, getting snapshots @cindex Source, obtaining snapshots @cindex Source, daily snapshots Testing release candidates and daily snapshots of the development repository are available via the World Wide Web from @url{http://n0nb.users.sourceforge.net/, Hamlib Git daily snapshots}. These are not official releases but are provided for testing new features and bug fixes. Testing release snapshots are only posted for a few weeks before a planned release. The daily development snapshot is made and posted each day by around 1030 UTC. Daily snapshots @i{should} compile but sometimes a bug creeps in that prevents compilation. If that should happen, please report it to the @email{hamlib-developer@@lists.sourceforge.net, hamlib-developer mailing list}. @node Git clone @subsection Git repository @cindex Git repository @cindex Git clone The source repository can be @dfn{cloned} which copies the repository to your computer including its entire history, branches, and release tag information. In other words, once the @command{git} @option{clone} command is finished a complete copy of the Hamlib development will be on your computer. You can do quite a lot with this as nothing is hidden from view since the entire history of Hamlib is right there all the way from the very first commit to the present. None of the meta-data is hidden away on some central server. To clone the repository use the following command: @example git clone git://git.code.sf.net/p/hamlib/code hamlib @end example Odds are that you will want to run the above command in a sub directory of your home directory. The @file{hamlib} directory will be created by Git and the @dfn{master} branch will be checked out for you as the @dfn{working copy}. The master branch is one of several branches used in Hamlib development. It is the main branch of new features and bug fixes. The working copy will be the latest revision of every file at the time of the clone. Later updates from the developers will require using another Git command to update your local repository. @xref{Working with Git}. @node Building from source @section Building from source @cindex Building from source @cindex Source, building from Building from source will be required for various reasons. Perhaps only an older release is provided by your distribution, or you'd like to test recent changes to Hamlib---either a specific back end or API changes---and offer a report to the developers, or you'd like to take part in development and offer your contribution to the project, or you'd just like to learn how to build a relatively comprehensive package from source. Any is a good reason to build from the source code. Before going further, this manual assumes familiarity with working from the command prompt in a Linux/BSD/Unix like system's @dfn{shell} environment, either in a @dfn{virtual console} (a text only screen with no graphics) or in a @dfn{terminal} in a desktop environment (@command{xterm}, @command{rxvt}, @command{konsole}, @command{gnome-terminal}, @command{xfce4-terminal}, @command{terminal}, etc.). If this is new to you, take some time and read up on using the shell. A good tutorial can be found at @url{http://linuxcommand.org/, LinuxCommand.org} which also offers an in-depth book that can be purchased or downloaded for no cost (the Hamlib project is not associated with nor has any interest in the sale of this book, it just looks like a very good effort on the part of its author). Let's get started. @menu * Compiling source tarballs:: * Bootstrapping from a Git clone:: * Other make targets:: * Parallel build trees:: * Adding debugging symbols:: * Compiling MS Windows 32:: @end menu @node Compiling source tarballs @subsection Compiling source tarballs @cindex Compiling source tarballs @cindex Source tarballs, compiling Before proceeding, it is essential to read the information in the files, @file{README}, @file{INSTALL}, and @file{README.betatester} supplied in the Hamlib @dfn{top-level} directory which will be named something like @file{hamlib-3.0~git} where the latter part is the release version. In this case the @samp{3.0~git} indicates this is a development snapshot of the Git master branch. These files provide detailed information for compiling Hamlib and will vary some from release to release. Compiling from a source tarball whether it is an official release or a testing or daily development snapshot follows the same set of commands, known as the @dfn{three step} which are each run from the top-level directory: @example ./configure make sudo make install @end example @menu * configure:: * make:: * make install:: * ldconfig:: @end menu @node configure @subsubsection @command{configure} @cindex configure The @command{./configure} command examines your system and checks it for any packages that are required or good to have options for compiling Hamlib. The leading @file{./} tells the shell to only run the @command{configure} command found in the current directory. It's always possible that a @command{configure} command could be lurking elsewhere and we don't want to run that! Run: @example @command{./configure} @end example from the top-level directory. @quotation Note Some distributions are configured so commands can only be run from directories listed in the @env{PATH} environment variable. The @file{./} is necessary or the @command{configure} command will not be run as the @dfn{current directory} (defined as @file{.}) is not in the @env{PATH}. This is considered a default security feature so that only programs provided by the distribution are run. @env{PATH} can be modified for your own session, but that is a topic for the LinuxCommand.org reference above. @end quotation Of course, things are usually complicated a bit by options and Hamlib is no exception. The good news is that the defaults, i.e. no options, work well in most situations. Options are needed to enable the compilation of certain portions of Hamlib such as the language bindings. Optional features usually require that more development tools are installed. The @file{INSTALL}, and @file{README.betatester} files in the Hamlib top-level directory will have details on the options available for that release. A useful option is @samp{--prefix} which tells @command{configure} where in the file system hierarchy Hamlib should be installed. If it is not given, Hamlib will be installed in the @file{/usr/local} file system hierarchy. Perhaps you want to install to your home directory instead: @example @command{./configure --prefix=~/local} @end example All of the files will be installed in the @file{local} directory of your home directory. @file{local} will be created if it does not exist during installation as will several other directories in it. Installing in your home directory means that @dfn{root}, or superuser (administrator) privileges are not required when running @command{make install}. On the other hand, some extra work will need to be done so other programs can use the library. (TODO: describe this in an appendix). Another useful option is @samp{--help} which will give a few screens full of options for @command{configure}. If in a desktop environment the scroll bar can be used to scroll back up through the output. In either a terminal or a virtual console Linux supports the @key{Shift-PageUp} key combination to scroll back up. As luck would have it, @key{Shift-PageDown} can be used to scroll down toward the end of the output and the shell prompt. After a fair amount of time, depending on your computer, and a lot of screen output, @command{configure} will finish its job. So long as the few lines previous to the shell prompt don't say ``error'' or some such failure message Hamlib is ready to be compiled. If there is an error and all of the required packages listed in @file{README.betatester} have been installed, please ask for help on the @email{hamlib-developer@@lists.sourceforge.net, hamlib-developer mailing list}. @node make @subsubsection @command{make} @cindex make The @command{make} command is responsible for running the @dfn{compiler} which reads the source files and from the instructions it finds in them writes @dfn{object} files which are the binary instructions the @acronym{CPU} of a computer can execute. @command{make} then calls the @dfn{linker} which puts the object files together in the correct order to create the Hamlib library files and its executable programs. Run: @example @command{make} @end example from the top-level directory. Any error that causes @command{make} to stop early is cause for a question to the @email{hamlib-developer@@lists.sourceforge.net, hamlib-developer mailing list}. In general @command{make} will take longer than @command{configure} to complete its run. As it is a system command and therefore found in the @env{PATH}, prefixing @command{make} with @file{./} will cause a @samp{command not found} error from the shell. @node make install @subsubsection @command{make install} @cindex make install Assuming that you have not set the installation prefix to your home directory, root (administrator) privileges will be required to install Hamlib to the system directories. Two popular methods exist for gaining root privileges, @command{su} and @command{sudo}. @command{sudo} is probably the most popular these days, particularly when using the @url{http://www.ubuntu.com, Ubuntu} family of distributions. Run: @example @command{sudo make install} @end example as root from the top-level directory. Running @command{make install} will call the installer to put all of the newly compiled files and other files (such as this document) in predetermined places set by the @samp{--prefix} option to @command{configure} in the directory hierarchy (yes, this is by design and @command{make} is not just flinging files any old place!). A lot of screen output will be generated. Any errors will probably be rather early in the process and will likely be related to your @var{username} not having write permissions in the system directory structure. @node ldconfig @subsubsection @command{ldconfig} @cindex ldconfig Once the installation is complete one more step is required if Hamlib has never been installed from a local build before. The @command{ldconfig} command tells the system library loader where to find the newly installed Hamlib libraries. It too will need to be run with root privileges: Run: @example @command{sudo ldconfig} @end example as root from any directory. @quotation Note Subsequent installations of Hamlib will not need to have @command{ldconfig} run after each installation so long as no new backend libraries nor a newer major version of Hamlib were installed. A developer adding a new backend (see @file{README.developer}) will need to run @command{ldconfig} for the new back end library to be found by the library loader (this refers a new back end, i.e. a sibling to the existing Icom, Kenwood, Yaesu, etc.@: back ends, not a new model in an existing back end). Likewise, any time a new back end appears in a source tarball, @command{ldconfig} will need to be run for the library loader to find it. There is no harm running @command{ldconfig} after each run of @command{make install}. @end quotation On some distributions a bit of configuration will be needed before @command{ldconfig} will add locally compiled software to its database. Please consult your distribution's documentation. @node Bootstrapping from a Git clone @subsection Bootstrapping from a @command{git clone} @cindex Bootstrapping from a Git clone @cindex Git clone, bootsrapping Choosing to build from from a @command{git clone} requires a few more development tools (notice a theme here?) as detailed in @file{README.developer}. The most critical will be the GNU Autotools (@command{autoconf}, @command{automake}, @command{libtool}, and more) from which the build system consisting of @file{configure}, the various @file{Makefile.in}s throughout the directory structure, and the final @file{Makefile}s are generated. In the top-level directory is the @command{autogen.sh} script from which the build system is @dfn{bootsrapped}---the process of building the Hamlib build system. At its completion @command{autogen.sh} will run @command{configure} and will pass any command line arguments it receives to @command{configure}. Environment variables intended for the compiler may also be set on the @command{autogen.sh} command line. After the configuration is complete, the build may proceed with the @command{make} step as for the source tarballs above. Or @command{configure --help} may be run, and @command{configure} run again with specific options in which case the @file{Makefile}s will be regenerated and the build will proceed. @node Other make targets @subsection Other @command{make} targets @cindex Other @command{make} targets @cindex @command{make}, other targets Besides @command{make install}, other @dfn{targets} exist when running @command{make}. Running @command{make clean} from the top-level directory removes all of the generated object and executable files generated by running @command{make} freeing up considerable disk space. @quotation Note During development of individual source files, it is not necessary to run @command{make clean} each time before @command{make}. Simply run @command{make} and only the modified file(s) and any objects that depend on them will be recompiled. This speeds up development time considerably. @end quotation To remove even the generated @file{Makefile}s, run @command{make distclean} from the top-level directory. After this target is run, @command{configure} will need to be run again to regenerate the @file{Makefile}s. This command may not be as useful as the @file{Makefile}s do not take up much space. @node Parallel build trees @subsection Parallel build trees @cindex Parallel build trees @cindex Build, parallel trees One feature of the GNU build system used by Hamlib is that the object files can be kept in a directory structure separate from the source files. While this has no effect on the @command{make} targets described above, it does help the developer find files in the source tree! One such way of using parallel builds is described in @file{README.developer}. Parallel builds can be very useful as one build directory can be configured for a release and another build directory can be configured for debugging with different options passed to @command{configure} from each directory. The generated @file{Makefile}s are unique to each build directory and will not interfere with each other. @node Adding debugging symbols @subsection Adding debugging symbols @cindex Adding debugging symbols When additional debugging symbols are needed with, for example, the GNU Debugger, @command{gdb}, the needed compiler and linker options are passed as environment variables. Run: @example @command{../hamlib/configure CFLAGS="-ggdb3 -O0" CXXFLAGS="-ggdb3 -O0"} @end example from a build directory intended for a debugging build. The @samp{-ggdb3} option tells the C compiler, this case the GNU C Compiler, @command{gcc}, to add special symbols useful for @command{gdb}. The @samp{-O0} option tells @command{gcc} to turn off all optimizations which will make it easier to follow some variables that might otherwise be optimized away. @samp{CFLAGS} and @samp{CXXFLAGS} may be set independently for each compiler (sadly, a bug in the build system source file in Hamlib 1.2.15.3 and earlier did not preserve @samp{CXXFLAGS}). @quotation Note There are a number compiler options available for controlling debugging symbols and setting optimization levels. Please consult the compiler's manual for all the details. @end quotation @node Compiling MS Windows 32 @subsection Compiling for MS Windows 32 bit architectures @cindex Compiling for MS Windows 32 @cindex MS Windows 32, compiling Currently compiling is done on an Xubuntu 10.10 virtual machine using @url{http://www.mingw.org/, MinGW32}. @file{README.build-win32} in the @file{scripts} directory has details on how this is accomplished. Work is ongoing to correct build issues in the @url{http://www.cygwin.com/, Cygwin} environment running on MS Windows. @node MS Windows 32 binaries @section Pre-compiled binaries for MS Windows 32 bit architectures @cindex Pre-compiled binaries for MS Windows 32 @cindex MS Windows 32, compiled binaries @cindex MS Windows 32, pre-compiled binaries Pre-compiled binaries for MS Windows 32 bit architectures (Windows NT and newer) are available for both official releases and daily development snapshots. Official releases are available through the @url{http://sourceforge.net/projects/hamlib/files/hamlib/, SourceForge.net file download service}. Daily development snapshots are available from @url{http://n0nb.users.sourceforge.net/}. Beginning with the Hamlib 1.2.15.3 release a self-extracting installer is available. Among its features are selecting which portions of Hamlib are installed. The @env{PATH} environment variable will need to be set manually per the included @file{README.win32-bin} file. Daily development snapshots feature both a .ZIP archive and the self extracting installer. Bug reports and questions about these archives should be sent to the @email{hamlib-developer@@lists.sourceforge.net, hamlib-developer mailing list}.