On 2024-11-03, Neal Gompa <ngomp...@gmail.com> wrote: > It's not that simple. The reason I suggested MPL-2.0 is because > LGPL-2.1-or-later is effectively the same as GPL-2.0-or-later with > Rust because it's statically linked. > > MPL-2.0 preserves the copyleft at a per source file level, but allows > the binary artifact to have a composition of compatible licenses > (including the GNU ones). This is the least messy for Rust bindings to > LGPL libraries.
On second thoughts, I'm not even sure we need the copyleft protections that much here, since - The rust bindings is probably quite shallow shells - The real 'important' things lives in already (l)gpl licensed components. That also makes the usual Rust MIT/Apache2 dual license useful. We just also need to make sure that people knows the thing effectively is (l)gpl. I think I at least won't object to adding a snippet to licensing policies that Rust code can be MIT/Apache2 /Sune
