Improvements to support packaging on Fedora

[ekohl]

My goal is to create a Fedora package. My intent for packaging is to:

• Use cmake to install
• Modify the code so it can install the binary to /usr/bin and data to /usr/share/railcontrol
• Add a systemd service file

To get there I'd like to use CPack to build packages and bring parity to make dist, but enhanced to support installing to the GNU hierarchy using GNUInstallDirs. Fedora encourages running a test suite so I'd like to implement CTest to bring parity to make test.

Fedora packaging policy also disallows any vendored dependencies so that should also become optional.

There are also related changes I'd like make for better development, like adding CI because I don't like manually verifying I didn't break anything.

This is based on what I posted in #15 (comment)

[ekohl]

To get there I'd like to use CPack to build packages and bring parity to make dist

#18

Fedora packaging policy also disallows any vendored dependencies so that should also become optional.

#15

There are also related changes I'd like make for better development, like adding CI because I don't like manually verifying I didn't break anything.

#17

[teddych]

On Cygwin I get the following errors with cmake 4.1.3:

CMake Deprecation Warning at CMakeLists.txt:1 (cmake_minimum_required):
Compatibility with CMake < 3.10 will be removed from a future version of
CMake.

Update the VERSION argument value. Or, use the ... syntax
to tell CMake that the project requires at least but has been updated
to work with policies introduced by or earlier.

Could you please have a look at this?

[ekohl]

It's a warning, not an error. It comes from:

railcontrol/CMakeLists.txt

Line 1 in b6e12dd

cmake_minimum_required(VERSION 3.5)

It's a heads up that CMake will drop compatibility for version 3.5. It's probably safe to raise it to a higher version which makes the warning go away.

[teddych]

For the moment I can live with the cmake version.
But the second VERSION-error is more critical, because this refers to the RailControl-version that is not reported anymore at RailControl's startup.

[ekohl]

But the second VERSION-error is more critical, because this refers to the RailControl-version that is not reported anymore at RailControl's startup.

I don't know what you mean: there is only 1 warning, which is from CMake. It refers to the VERSION argument to cmake_minimum_required in the line I linked.

[teddych]

Sorry, I got somehow an empty RAILCONTROL_VERSION variable, but can't reproduce it anymore. Will have an eye on it. But for the moment this is OK for me as is.

[teddych]

I'm not very familiar with all the cmake and build tools stuff...

[ekohl]

#26

[Fuddl1]

Modify the code so it can install the binary to /usr/bin and data to /usr/share/railcontrol

@ekohl I'm about to create a pull request, can you please review and test on fedora this change: Fuddl1@1aa075d

[ekohl]

I'm afraid my break is over and daily life is eating more of my time again so I'll be a bit slower in responding.

@ekohl I'm about to create a pull request, can you please review and test on fedora this change: Fuddl1@1aa075d

That looks like a smaller implementation of #24. There I chose to make the inclusion of GNUInstallDirs optional because I wanted to maintain the option to install everything in the same directory because I think for Windows that may be an easier model. Not sure actually, but my abstractions in #24 can also allow for Windows to use the right paths in a user's home directory.