Skip to content

Latest commit

 

History

History
59 lines (34 loc) · 2.18 KB

README.md

File metadata and controls

59 lines (34 loc) · 2.18 KB

Custom set of cerbero recipes

Install requisites

Install python >=3.7 since cerbero 1.20.0 requires it.

Install required python modules

pip install six pyparsing

Install aditional software required by flexvdi-client recipe when building for windows:

dnf install mingw32-nsis-3.06.1-2.fc34.x86_64

Download

Clone this repo and update the cerbero submodule:

$ git clone https://github.com/flexVDI/cerbero-recipes
$ cd cerbero-recipes
$ git submodule init && git submodule update

Use

  1. Configure ~/.cerbero/cerbero.cbc. Add home_dir, at least. E.g. home_dir = '/home/you/cerbero'
  2. Use cerbero.sh script in the root dir of the project. It includes the flexvdi.cbc file and runs from the cerbero config directory so that a) custom recipe paths and toolchains are set, and b) other configuration files are referenced without path.

NOTE: When cross-compiling on Linux to Windows, glib recipe seems to need more than 1024 open files. Cerbero will try to increase the limit where needed, but you may be asked to do it manually.

By default, binaries are built in the debug variant. To build a release variant, call cerbero.sh with the "-v nodebug" option.

Examples:

  1. Bootstrap the Android build tools and NDK:

    ./cerbero.sh -c cross-android-universal.cbc bootstrap

  2. Build the release version of flexVDI Desktop Client on Windows 32-bit:

    ./cerbero.sh -v nodebug -c cross-win32.cbc build flexvdi-client

Building portable Linux binaries

In order to build Linux binaries that can be run in many other Linux distributions, use the linux-container.sh script. It launches a docker container based on Debian Stretch. Linux binaries built there will be compatible with many Linux distributions of the last ~5 years.

For 64-bit environment:

$ ./linux-container.sh

For 32-bit environment:

$ ./linux-container.sh i686

The script will discover the Cerbero home_dir you configured in cerbero.cbc and mount it as a volume. So, you only have to run cerbero bootstrap once, even if you destroy the container. Inside the container, the cerbero binary is in the $PATH, and the cerbero user is created with the same uid and gid of the user running docker.