(v)snprintf space optimisations

[mark9064]

We can save about 1.5K flash by only including one (v)snprintf implementation. Currently both LVGL's and newlib-nano's are included. The newlib-nano implementation does not support 64 bit integers, so we can't replace the LVGL one with it (notably the calculator app PR depends on it which is slated to be merged soon, there may be other usages I'm not aware of). So as far as I'm aware we'd ideally want to link against the normal newlib (not nano) implementation (no clue how big this is), or use the LVGL one everywhere. I'd propose this in a PR but my CMake skills are nowhere near what's needed, so if anyone wants to give it a go have at it!