Hi Jonas, Thanks for the fix. On a second thought yes, given that you're building the package as "Architecture: all" your build.rs patch probably is the right thing to do.
> I think that patch is generally usable, not only for Debian but also sensible > for upstream to adopt. I agree. They are aiming at portable code including platforms with no SSE2 support (they've been testing the i586 target on windows in CI since at least the commit that first introduced the bug and there are explicit code paths for that, this is one reason why I insisted a bit that the bug is not 100% Debian-specific. I thought they might be interested nonetheless), so they should made the user aware that the library doesn't build optimized on such targets. But in the end, yes, as you said that's their choice. Thanks again, NC
