- added a 'platform' target in Makefile that will build only the platform-specific
zenith binary by default i.e. NVIDIA enabled binary if NVIDIA driver detected
else basic binary
- the 'all' target in Makefile will build both flavors if possible like before
but is no longer the default make target (which is 'platform' now)
- the 'install' target adapts to both cases whether 'platform' binary has
been built or both binaries have been built (latter will switch using the
wrapper script at runtime as before)
- debian make target uses 'all' since platform can be different at runtime
* Support for deb package
- added an icon file (royalty-free, attribution-free) from pngkey
- added debian directory with required build files and desktop file
- zenith.sh wrapper script to dynamically switch between NVIDIA and
normal build depending on the presence of NVIDIA device and library
- Makefile to encapsulate various standard build actions including "install"
that builds both the regular as well as NVIDIA enabled binary, then
copies those with the wrapper script, desktop and icon files
- removed build-linux-static.sh which is now covered via "make linux-static"
* use /usr/local as install path by default
- Makefile now uses /usr/local as default installation path
as per usual convention while debian build changes the PREFIX to /usr
- place the binaries inside /usr/lib/zenith/base and /usr/lib/zenith/nvidia
so that the process name appears as zenith even in monitoring tools
* changed the static build to use zenith in nvidia/base subdirs
following same convention as deb build, the real zenith binary lives
inside subdirectories so that its name in tools appears as just "zenith"
* rename zenith dir to zenith-exec in static package
fix clash of zenith script name with the directory name
* Makefile now detects NVIDIA and does appropriate build
* correct uninstall target
* changes as per PR review comments
- removing -D option to install and instead use explicit "mkdir -p" for MacOS
- upddate linux-static target to use musl-gcc, and normal targets with parameters
instead of duplicating the code
- added more build instructions to README.md including static builds and within VMs
- search LD_LIBRARY_PATH too to determine presence of nvidia library; moved the
code to separate script
* add BUILD_NVIDIA flag in Makefile to explicitly enable/disable
* updated README.md
* set BUILD_NVIDIA to false for linux-static by default
* perform some Makefile actions only for Linux
BUILD_NVIDIA and desktop/icon file install only done for Linux now
* make everything silent in install target