Paolo Bonzini <pbonz...@redhat.com> writes:
> On 11/3/24 10:54, Junjie Mao wrote:
>> Paolo Bonzini <pbonz...@redhat.com> writes:
>>
>>> diff --git a/rust/qemu-api-macros/src/lib.rs
>>> b/rust/qemu-api-macros/src/lib.rs
>>> index a4bc5d01ee8..c2ea22101e4 100644
>>> --- a/rust/qemu-api-macros/src/lib.rs
>>> +++ b/rust/qemu-api-macros/src/lib.rs
>>> @@ -3,8 +3,34 @@
>>> // SPDX-License-Identifier: GPL-2.0-or-later
>>>
>>> use proc_macro::TokenStream;
>>> -use quote::quote;
>>> -use syn::{parse_macro_input, DeriveInput};
>>> +use proc_macro2::Span;
>>> +use quote::{quote, quote_spanned};
>>> +use syn::{
>>> + parse_macro_input, parse_quote, punctuated::Punctuated, token::Comma,
>>> Data, DeriveInput, Field,
>>> + Fields, Ident, Type, Visibility,
>>> +};
>>> +
>>> +struct CompileError(String, Span);
>>> +
>>> +impl From<CompileError> for proc_macro2::TokenStream {
>>> + fn from(err: CompileError) -> Self {
>>> + let CompileError(msg, span) = err;
>>> + quote_spanned! { span => compile_error!(#msg); }
>> The documentation [2] says "there should be no space before the =>
>> token" and that is by intention to tell that `span` is "evaluated in the
>> context of proc macro" while those after the arm "in the generated
>> code". Should we follow that convention (even though the extra white
>> space does not impact building)?
>
> Ah, forgot to reply about this. Personally I think it's clear enough with the
> space around both sides of "=>", but if there's agreement on removing the
> space
> I don't oppose it.
>
> Paolo
I don't have any preference, either. I think we can keep it as is and
make sure future calls to quote_spanned! have consistent style. Thanks.
--
Best Regards
Junjie Mao
