Hi everyone!
I tried compiling an LV2 plugin with faust2lv2 on an arm64 Raspios without the elkpi-sdk. I copied it to a different Raspberry Pi that’s running Elk Audio OS, configured Sushi to run it and, to my surprise, it works!
My question is, are there any downsides to this workflow as compared to compiling LV2 plugins with elkpi-sdk?
Hi gaboose, do you mean that you’re building on a RPi with Raspios? Well, if it works it works :), the architecture is the same, but compatibility with system libraries on ElkPi is not guaranteed, so it may work now and not in the future.
Another downside is that it is slow. You can also build on a RPi running a dev image of elk audio os, but that is roughly the same thing.
My 5 cents is that if you take the time to build and setup the docker image, that is time well spent in the future. As x-compiling on a fast dev machine is so much quicker.
Yes!
I’m curious about this and what changes the sdk does to the resulting plugin library. May I ask what type of compatibilities do you mean?
I’ll try. I’m familiar with docker but less so with the gcc toolkit and not at all with yocto.
It was indeed slower to compile on a Raspberry, but the setup was very easy.
Word of caution for future readers: It wasn’t long before I ran into a problem. I tried compiling a pure data patch (simple reverb/echo with plugdata and dpf) to lv2 the same way (on arm64 without elkpi-sdk). It successfully produced an lv2 library but every time sushi tries to load it it makes the whole pi unresponsive.
That was enough for me to decide not to skip the sdk in the future 