On Thu, Oct 24, 2024 at 4:35 AM Junjie Mao <junjie....@hotmail.com> wrote: > > > Paolo Bonzini <pbonz...@redhat.com> writes: > > > Replay the configuration that would be computed by build.rs when compiling > > on a 1.63.0 compiler. > > > > Signed-off-by: Paolo Bonzini <pbonz...@redhat.com> > > Reviewed-by: Junjie Mao <junjie....@hotmail.com> > > One question below. > > > --- > > subprojects/packagefiles/proc-macro2-1-rs/meson.build | 4 +++- > > 1 file changed, 3 insertions(+), 1 deletion(-) > > > > diff --git a/subprojects/packagefiles/proc-macro2-1-rs/meson.build > > b/subprojects/packagefiles/proc-macro2-1-rs/meson.build > > index 818ec59336b..8e601b50ccc 100644 > > --- a/subprojects/packagefiles/proc-macro2-1-rs/meson.build > > +++ b/subprojects/packagefiles/proc-macro2-1-rs/meson.build > > @@ -15,7 +15,9 @@ _proc_macro2_rs = static_library( > > rust_abi: 'rust', > > rust_args: [ > > '--cfg', 'feature="proc-macro"', > > - '--cfg', 'span_locations', > > 'span_locations' is not subject to any version check in build.rs. Is it > removed because of its performance penalty and the fact that we are not > using it anywhere?
The docs say "When executing in a procedural macro context, the returned line/column are only meaningful if compiled with a nightly toolchain. The stable toolchain does not have this information available. When executing outside of a procedural macro, such as main.rs or build.rs, the line/column are always meaningful regardless of toolchain". So for our case it seems useless. Paolo
