Difference between revisions of "Linux development"
(→CMake: note about clean cmake run; see Trac #1055)
m (moved Navit SVN installation on Linux to Linux: Make autocomplete easier, keep links short and specific, unify OS articles)
Revision as of 09:42, 3 December 2012
Taking care of dependencies
First, let's make sure we are in our home directory: this is only for the sake of making this tutorial simple to follow. You can save that directory anywhere you want, but you will have to adapt the rest of the instructions of this guide to your particular case.
Now, let's grab the code from SVN. This assumes that you have subversion installed.
There are currently two methods of compiling Navit - using GNU autotools or using CMake. GNU autotools was the old method, but will become deprecated in favour of CMake.
CMake builds Navit in a separate directory of your choice - this means that the directory in which the SVN source was checked out remains untouched.
mkdir navit-build cd navit-build
Once inside the build directory just call the following commands:
cmake ~/navit make
Note: CMake will autodetect your system configuration on the first run, and cache this information. Therefore installing or removing libraries after the first CMake run may confuse it and cause weird compilation errors. If you install or remove libraries/packages (for example because the first CMake run complained about them missing), always do a clean CMake run:
rm -r ~/navit-build/* cmake ~/navit
Running the compiled binary
The binary is called navit and can be run without arguments:
cd ~/navit-build/navit/ ./navit
It is advised to just run this binary locally at the moment (i.e. not to install system-wide). Note that for this to work, Navit must be run from the directory where it resides (that is, you must first change your working directory, as described above). If Navit is run from another directory, it will not find its plugins and image files, and will not start.
Go to the directory where you checked out the svn repository:
Start the make dance:
./autogen.sh && ./configure && make
It is better to use autogen and configure before make, due to the heavy flux of changes in the svn version. Also, putting them all in one command ensures that they go through the next step only if the previous step succeeded. In the end, hopefully, you get a fully compiled navit binary in the navit/ subfolder (so basically, it looks like ~/navit/navit/navit. Yeah.)
Running the compiled binary
Here, I am skipping the usual "make install" (in my case, "sudo checkinstall make install") step, because I don't care about installing navit systemwide. I run navit directly from its svn folder, and it makes things much simpler (and safer).
To execute navit, you can simply click on the binary file (if you are sure it is compiled properly) and it should launch. If you prefer to launch it from a terminal, you need to go into the directory containing the binary, first, like so:
cd ~/navit/navit/ ./navit
You don't need to recompile everything to update navit to the latest SVN; with svn update only the edited files will be downloaded. Just go to the navit directory (e.g. /home/CHANGEME/navit) and run:
You then follow this by either the instructions in CMake or in make. Actually, with CMake you can miss out the step where you make the build directory - just use the old one. With make you can usually miss out the ./autogen.sh and ./configure steps.
Prebuilt binaries exist for many distributions.
Configuring the beast
This is an entirely different matter, young padawan. Good luck :)
You can also check a post describing a Navit configuration on Ubuntu Jaunty.