rust/compiler/rustc_builtin_macros/src/proc_macro_harness.rs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

387 lines
13 KiB
Rust
Raw Normal View History

use crate::errors;
use rustc_ast::ptr::P;
use rustc_ast::visit::{self, Visitor};
use rustc_ast::{self as ast, attr, NodeId};
use rustc_ast_pretty::pprust;
use rustc_expand::base::{parse_macro_name_and_helper_attrs, ExtCtxt, ResolverExpand};
use rustc_expand::expand::{AstFragment, ExpansionConfig};
use rustc_feature::Features;
use rustc_session::Session;
use rustc_span::hygiene::AstPass;
use rustc_span::source_map::SourceMap;
2020-04-19 13:00:18 +02:00
use rustc_span::symbol::{kw, sym, Ident, Symbol};
use rustc_span::{Span, DUMMY_SP};
use smallvec::smallvec;
use std::mem;
use thin_vec::{thin_vec, ThinVec};
struct ProcMacroDerive {
id: NodeId,
2020-04-19 13:00:18 +02:00
trait_name: Symbol,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
function_name: Ident,
span: Span,
2020-04-19 13:00:18 +02:00
attrs: Vec<Symbol>,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
struct ProcMacroDef {
id: NodeId,
function_name: Ident,
span: Span,
}
enum ProcMacro {
Derive(ProcMacroDerive),
Attr(ProcMacroDef),
Bang(ProcMacroDef),
}
struct CollectProcMacros<'a> {
macros: Vec<ProcMacro>,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
in_root: bool,
dcx: &'a rustc_errors::DiagCtxt,
source_map: &'a SourceMap,
is_proc_macro_crate: bool,
is_test_crate: bool,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
pub fn inject(
krate: &mut ast::Crate,
sess: &Session,
features: &Features,
2020-06-27 23:51:28 +03:00
resolver: &mut dyn ResolverExpand,
is_proc_macro_crate: bool,
has_proc_macro_decls: bool,
is_test_crate: bool,
2023-12-18 10:15:45 +11:00
dcx: &rustc_errors::DiagCtxt,
) {
let ecfg = ExpansionConfig::default("proc_macro".to_string(), features);
2020-03-16 00:43:37 +01:00
let mut cx = ExtCtxt::new(sess, ecfg, resolver, None);
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
let mut collect = CollectProcMacros {
macros: Vec::new(),
in_root: true,
2023-12-18 10:15:45 +11:00
dcx,
source_map: sess.source_map(),
is_proc_macro_crate,
is_test_crate,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
};
if has_proc_macro_decls || is_proc_macro_crate {
visit::walk_crate(&mut collect, krate);
}
let macros = collect.macros;
if !is_proc_macro_crate {
return;
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
if is_test_crate {
return;
}
2021-07-16 22:22:08 +02:00
let decls = mk_decls(&mut cx, &macros);
krate.items.push(decls);
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
impl<'a> CollectProcMacros<'a> {
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
fn check_not_pub_in_root(&self, vis: &ast::Visibility, sp: Span) {
if self.is_proc_macro_crate && self.in_root && vis.kind.is_pub() {
self.dcx.emit_err(errors::ProcMacro { span: sp });
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
}
fn collect_custom_derive(&mut self, item: &'a ast::Item, attr: &'a ast::Attribute) {
2022-02-19 00:48:49 +01:00
let Some((trait_name, proc_attrs)) =
parse_macro_name_and_helper_attrs(self.dcx, attr, "derive")
2022-02-19 00:48:49 +01:00
else {
return;
};
if self.in_root && item.vis.kind.is_pub() {
self.macros.push(ProcMacro::Derive(ProcMacroDerive {
id: item.id,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
span: item.span,
trait_name,
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
function_name: item.ident,
attrs: proc_attrs,
}));
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
} else {
let msg = if !self.in_root {
"functions tagged with `#[proc_macro_derive]` must \
currently reside in the root of the crate"
} else {
"functions tagged with `#[proc_macro_derive]` must be `pub`"
};
self.dcx.span_err(self.source_map.guess_head_span(item.span), msg);
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}
}
fn collect_attr_proc_macro(&mut self, item: &'a ast::Item) {
if self.in_root && item.vis.kind.is_pub() {
self.macros.push(ProcMacro::Attr(ProcMacroDef {
id: item.id,
span: item.span,
function_name: item.ident,
}));
} else {
let msg = if !self.in_root {
"functions tagged with `#[proc_macro_attribute]` must \
currently reside in the root of the crate"
} else {
"functions tagged with `#[proc_macro_attribute]` must be `pub`"
};
self.dcx.span_err(self.source_map.guess_head_span(item.span), msg);
}
}
fn collect_bang_proc_macro(&mut self, item: &'a ast::Item) {
if self.in_root && item.vis.kind.is_pub() {
self.macros.push(ProcMacro::Bang(ProcMacroDef {
id: item.id,
span: item.span,
function_name: item.ident,
}));
} else {
let msg = if !self.in_root {
"functions tagged with `#[proc_macro]` must \
currently reside in the root of the crate"
} else {
"functions tagged with `#[proc_macro]` must be `pub`"
};
self.dcx.span_err(self.source_map.guess_head_span(item.span), msg);
}
}
}
impl<'a> Visitor<'a> for CollectProcMacros<'a> {
fn visit_item(&mut self, item: &'a ast::Item) {
2019-09-26 17:51:36 +01:00
if let ast::ItemKind::MacroDef(..) = item.kind {
if self.is_proc_macro_crate && attr::contains_name(&item.attrs, sym::macro_export) {
self.dcx.emit_err(errors::ExportMacroRules {
span: self.source_map.guess_head_span(item.span),
});
}
}
// First up, make sure we're checking a bare function. If we're not then
// we're just not interested in this item.
//
// If we find one, try to locate a `#[proc_macro_derive]` attribute on it.
2020-12-24 02:55:21 +01:00
let is_fn = matches!(item.kind, ast::ItemKind::Fn(..));
let mut found_attr: Option<&'a ast::Attribute> = None;
for attr in &item.attrs {
if attr.is_proc_macro_attr() {
if let Some(prev_attr) = found_attr {
let prev_item = prev_attr.get_normal_item();
let item = attr.get_normal_item();
let path_str = pprust::path_to_string(&item.path);
let msg = if item.path.segments[0].ident.name
== prev_item.path.segments[0].ident.name
{
format!(
"only one `#[{path_str}]` attribute is allowed on any given function",
)
} else {
format!(
"`#[{}]` and `#[{}]` attributes cannot both be applied
to the same function",
path_str,
pprust::path_to_string(&prev_item.path),
)
};
self.dcx
.struct_span_err(attr.span, msg)
Make `DiagnosticBuilder::emit` consuming. This works for most of its call sites. This is nice, because `emit` very much makes sense as a consuming operation -- indeed, `DiagnosticBuilderState` exists to ensure no diagnostic is emitted twice, but it uses runtime checks. For the small number of call sites where a consuming emit doesn't work, the commit adds `DiagnosticBuilder::emit_without_consuming`. (This will be removed in subsequent commits.) Likewise, `emit_unless` becomes consuming. And `delay_as_bug` becomes consuming, while `delay_as_bug_without_consuming` is added (which will also be removed in subsequent commits.) All this requires significant changes to `DiagnosticBuilder`'s chaining methods. Currently `DiagnosticBuilder` method chaining uses a non-consuming `&mut self -> &mut Self` style, which allows chaining to be used when the chain ends in `emit()`, like so: ``` struct_err(msg).span(span).emit(); ``` But it doesn't work when producing a `DiagnosticBuilder` value, requiring this: ``` let mut err = self.struct_err(msg); err.span(span); err ``` This style of chaining won't work with consuming `emit` though. For that, we need to use to a `self -> Self` style. That also would allow `DiagnosticBuilder` production to be chained, e.g.: ``` self.struct_err(msg).span(span) ``` However, removing the `&mut self -> &mut Self` style would require that individual modifications of a `DiagnosticBuilder` go from this: ``` err.span(span); ``` to this: ``` err = err.span(span); ``` There are *many* such places. I have a high tolerance for tedious refactorings, but even I gave up after a long time trying to convert them all. Instead, this commit has it both ways: the existing `&mut self -> Self` chaining methods are kept, and new `self -> Self` chaining methods are added, all of which have a `_mv` suffix (short for "move"). Changes to the existing `forward!` macro lets this happen with very little additional boilerplate code. I chose to add the suffix to the new chaining methods rather than the existing ones, because the number of changes required is much smaller that way. This doubled chainging is a bit clumsy, but I think it is worthwhile because it allows a *lot* of good things to subsequently happen. In this commit, there are many `mut` qualifiers removed in places where diagnostics are emitted without being modified. In subsequent commits: - chaining can be used more, making the code more concise; - more use of chaining also permits the removal of redundant diagnostic APIs like `struct_err_with_code`, which can be replaced easily with `struct_err` + `code_mv`; - `emit_without_diagnostic` can be removed, which simplifies a lot of machinery, removing the need for `DiagnosticBuilderState`.
2024-01-03 12:17:35 +11:00
.span_label_mv(prev_attr.span, "previous attribute here")
.emit();
return;
}
found_attr = Some(attr);
}
}
2022-02-19 00:48:49 +01:00
let Some(attr) = found_attr else {
self.check_not_pub_in_root(&item.vis, self.source_map.guess_head_span(item.span));
let prev_in_root = mem::replace(&mut self.in_root, false);
visit::walk_item(self, item);
self.in_root = prev_in_root;
return;
};
if !is_fn {
let msg = format!(
"the `#[{}]` attribute may only be used on bare functions",
pprust::path_to_string(&attr.get_normal_item().path),
);
self.dcx.span_err(attr.span, msg);
return;
}
if self.is_test_crate {
return;
}
if !self.is_proc_macro_crate {
let msg = format!(
"the `#[{}]` attribute is only usable with crates of the `proc-macro` crate type",
pprust::path_to_string(&attr.get_normal_item().path),
);
self.dcx.span_err(attr.span, msg);
return;
}
if attr.has_name(sym::proc_macro_derive) {
self.collect_custom_derive(item, attr);
} else if attr.has_name(sym::proc_macro_attribute) {
self.collect_attr_proc_macro(item);
} else if attr.has_name(sym::proc_macro) {
self.collect_bang_proc_macro(item);
};
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
2018-07-13 01:59:41 +03:00
let prev_in_root = mem::replace(&mut self.in_root, false);
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
visit::walk_item(self, item);
self.in_root = prev_in_root;
}
}
// Creates a new module which looks like:
//
2019-08-25 21:03:24 +01:00
// const _: () = {
// extern crate proc_macro;
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
//
// use proc_macro::bridge::client::ProcMacro;
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
//
// #[rustc_proc_macro_decls]
// #[used]
// #[allow(deprecated)]
// static DECLS: &[ProcMacro] = &[
// ProcMacro::custom_derive($name_trait1, &[], ::$name1);
// ProcMacro::custom_derive($name_trait2, &["attribute_name"], ::$name2);
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
// // ...
// ];
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
// }
2021-07-16 22:22:08 +02:00
fn mk_decls(cx: &mut ExtCtxt<'_>, macros: &[ProcMacro]) -> P<ast::Item> {
let expn_id = cx.resolver.expansion_for_ast_pass(
2019-08-25 21:03:24 +01:00
DUMMY_SP,
AstPass::ProcMacroHarness,
&[sym::rustc_attrs, sym::proc_macro_internals],
None,
);
2021-06-25 20:43:04 +02:00
let span = DUMMY_SP.with_def_site_ctxt(expn_id.to_expn_id());
2019-08-25 21:03:24 +01:00
let proc_macro = Ident::new(sym::proc_macro, span);
let krate = cx.item(span, proc_macro, ast::AttrVec::new(), ast::ItemKind::ExternCrate(None));
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
let bridge = Ident::new(sym::bridge, span);
let client = Ident::new(sym::client, span);
let proc_macro_ty = Ident::new(sym::ProcMacro, span);
let custom_derive = Ident::new(sym::custom_derive, span);
let attr = Ident::new(sym::attr, span);
let bang = Ident::new(sym::bang, span);
2021-07-16 22:22:08 +02:00
// We add NodeIds to 'resolver.proc_macros' in the order
// that we generate expressions. The position of each NodeId
// in the 'proc_macros' Vec corresponds to its position
// in the static array that will be generated
let decls = macros
.iter()
.map(|m| {
let harness_span = span;
let span = match m {
ProcMacro::Derive(m) => m.span,
ProcMacro::Attr(m) | ProcMacro::Bang(m) => m.span,
};
let local_path = |cx: &ExtCtxt<'_>, name| cx.expr_path(cx.path(span, vec![name]));
let proc_macro_ty_method_path = |cx: &ExtCtxt<'_>, method| {
cx.expr_path(cx.path(
span.with_ctxt(harness_span.ctxt()),
vec![proc_macro, bridge, client, proc_macro_ty, method],
))
};
match m {
ProcMacro::Derive(cd) => {
2021-07-16 22:22:08 +02:00
cx.resolver.declare_proc_macro(cd.id);
cx.expr_call(
span,
2021-07-16 22:22:08 +02:00
proc_macro_ty_method_path(cx, custom_derive),
thin_vec![
cx.expr_str(span, cd.trait_name),
cx.expr_array_ref(
span,
cd.attrs
.iter()
.map(|&s| cx.expr_str(span, s))
.collect::<ThinVec<_>>(),
),
local_path(cx, cd.function_name),
],
)
}
ProcMacro::Attr(ca) | ProcMacro::Bang(ca) => {
cx.resolver.declare_proc_macro(ca.id);
let ident = match m {
ProcMacro::Attr(_) => attr,
ProcMacro::Bang(_) => bang,
ProcMacro::Derive(_) => unreachable!(),
};
cx.expr_call(
span,
proc_macro_ty_method_path(cx, ident),
thin_vec![
cx.expr_str(span, ca.function_name.name),
local_path(cx, ca.function_name),
],
)
}
}
})
.collect();
let decls_static = cx
.item_static(
span,
Ident::new(sym::_DECLS, span),
cx.ty_ref(
span,
cx.ty(
2019-12-22 17:42:04 -05:00
span,
ast::TyKind::Slice(
cx.ty_path(cx.path(span, vec![proc_macro, bridge, client, proc_macro_ty])),
),
2019-12-22 17:42:04 -05:00
),
None,
ast::Mutability::Not,
2019-12-22 17:42:04 -05:00
),
ast::Mutability::Not,
cx.expr_array_ref(span, decls),
)
.map(|mut i| {
i.attrs.push(cx.attr_word(sym::rustc_proc_macro_decls, span));
i.attrs.push(cx.attr_word(sym::used, span));
i.attrs.push(cx.attr_nested_word(sym::allow, sym::deprecated, span));
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
i
});
2019-12-22 17:42:04 -05:00
2019-09-05 15:05:58 +01:00
let block = cx.expr_block(
2023-01-30 14:13:27 +11:00
cx.block(span, thin_vec![cx.stmt_item(span, krate), cx.stmt_item(span, decls_static)]),
2019-09-05 15:05:58 +01:00
);
2019-08-25 21:03:24 +01:00
let anon_constant = cx.item_const(
span,
2020-04-19 13:00:18 +02:00
Ident::new(kw::Underscore, span),
cx.ty(span, ast::TyKind::Tup(ThinVec::new())),
2019-08-25 21:03:24 +01:00
block,
);
2019-09-05 15:05:58 +01:00
// Integrate the new item into existing module structures.
2019-08-25 21:03:24 +01:00
let items = AstFragment::Items(smallvec![anon_constant]);
cx.monotonic_expander().fully_expand_fragment(items).make_items().pop().unwrap()
rustc: Implement custom derive (macros 1.1) This commit is an implementation of [RFC 1681] which adds support to the compiler for first-class user-define custom `#[derive]` modes with a far more stable API than plugins have today. [RFC 1681]: https://github.com/rust-lang/rfcs/blob/master/text/1681-macros-1.1.md The main features added by this commit are: * A new `rustc-macro` crate-type. This crate type represents one which will provide custom `derive` implementations and perhaps eventually flower into the implementation of macros 2.0 as well. * A new `rustc_macro` crate in the standard distribution. This crate will provide the runtime interface between macro crates and the compiler. The API here is particularly conservative right now but has quite a bit of room to expand into any manner of APIs required by macro authors. * The ability to load new derive modes through the `#[macro_use]` annotations on other crates. All support added here is gated behind the `rustc_macro` feature gate, both for the library support (the `rustc_macro` crate) as well as the language features. There are a few minor differences from the implementation outlined in the RFC, such as the `rustc_macro` crate being available as a dylib and all symbols are `dlsym`'d directly instead of having a shim compiled. These should only affect the implementation, however, not the public interface. This commit also ended up touching a lot of code related to `#[derive]`, making a few notable changes: * Recognized derive attributes are no longer desugared to `derive_Foo`. Wasn't sure how to keep this behavior and *not* expose it to custom derive. * Derive attributes no longer have access to unstable features by default, they have to opt in on a granular level. * The `derive(Copy,Clone)` optimization is now done through another "obscure attribute" which is just intended to ferry along in the compiler that such an optimization is possible. The `derive(PartialEq,Eq)` optimization was also updated to do something similar. --- One part of this PR which needs to be improved before stabilizing are the errors and exact interfaces here. The error messages are relatively poor quality and there are surprising spects of this such as `#[derive(PartialEq, Eq, MyTrait)]` not working by default. The custom attributes added by the compiler end up becoming unstable again when going through a custom impl. Hopefully though this is enough to start allowing experimentation on crates.io! syntax-[breaking-change]
2016-08-22 17:07:11 -07:00
}