Merge #2114

2114: Move macro expansion to a separate crate r=matklad a=matklad



Co-authored-by: Aleksey Kladov <[email protected]>

4 files changed, 386 insertions, 0 deletions

diff --git a/crates/ra_hir_expand/Cargo.toml b/crates/ra_hir_expand/Cargo.toml
new file mode 100644
index 000000000..9bf5b7918
--- /dev/null
+++ b/crates/ra_hir_expand/Cargo.toml
@@ -0,0 +1,15 @@
+[package]
+edition = "2018"
+name = "ra_hir_expand"
+version = "0.1.0"
+authors = ["rust-analyzer developers"]
+
+[dependencies]
+log = "0.4.5"
+
+ra_arena = { path = "../ra_arena" }
+ra_db = { path = "../ra_db" }
+ra_syntax = { path = "../ra_syntax" }
+ra_prof = { path = "../ra_prof" }
+tt = { path = "../ra_tt", package = "ra_tt" }
+mbe = { path = "../ra_mbe", package = "ra_mbe" }
diff --git a/crates/ra_hir_expand/src/ast_id_map.rs b/crates/ra_hir_expand/src/ast_id_map.rs
new file mode 100644
index 000000000..cb464c3ff
--- /dev/null
+++ b/crates/ra_hir_expand/src/ast_id_map.rs
@@ -0,0 +1,106 @@
+//! `AstIdMap` allows to create stable IDs for "large" syntax nodes like items
+//! and macro calls.
+//!
+//! Specifically, it enumerates all items in a file and uses position of a an
+//! item as an ID. That way, id's don't change unless the set of items itself
+//! changes.
+
+use std::{
+    hash::{Hash, Hasher},
+    marker::PhantomData,
+};
+
+use ra_arena::{impl_arena_id, Arena, RawId};
+use ra_syntax::{ast, AstNode, AstPtr, SyntaxNode, SyntaxNodePtr};
+
+/// `AstId` points to an AST node in a specific file.
+#[derive(Debug)]
+pub struct FileAstId<N: AstNode> {
+    raw: ErasedFileAstId,
+    _ty: PhantomData<fn() -> N>,
+}
+
+impl<N: AstNode> Clone for FileAstId<N> {
+    fn clone(&self) -> FileAstId<N> {
+        *self
+    }
+}
+impl<N: AstNode> Copy for FileAstId<N> {}
+
+impl<N: AstNode> PartialEq for FileAstId<N> {
+    fn eq(&self, other: &Self) -> bool {
+        self.raw == other.raw
+    }
+}
+impl<N: AstNode> Eq for FileAstId<N> {}
+impl<N: AstNode> Hash for FileAstId<N> {
+    fn hash<H: Hasher>(&self, hasher: &mut H) {
+        self.raw.hash(hasher);
+    }
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]
+struct ErasedFileAstId(RawId);
+impl_arena_id!(ErasedFileAstId);
+
+/// Maps items' `SyntaxNode`s to `ErasedFileAstId`s and back.
+#[derive(Debug, PartialEq, Eq, Default)]
+pub struct AstIdMap {
+    arena: Arena<ErasedFileAstId, SyntaxNodePtr>,
+}
+
+impl AstIdMap {
+    pub(crate) fn from_source(node: &SyntaxNode) -> AstIdMap {
+        assert!(node.parent().is_none());
+        let mut res = AstIdMap { arena: Arena::default() };
+        // By walking the tree in bread-first order we make sure that parents
+        // get lower ids then children. That is, adding a new child does not
+        // change parent's id. This means that, say, adding a new function to a
+        // trait does not change ids of top-level items, which helps caching.
+        bfs(node, |it| {
+            if let Some(module_item) = ast::ModuleItem::cast(it.clone()) {
+                res.alloc(module_item.syntax());
+            } else if let Some(macro_call) = ast::MacroCall::cast(it) {
+                res.alloc(macro_call.syntax());
+            }
+        });
+        res
+    }
+
+    pub fn ast_id<N: AstNode>(&self, item: &N) -> FileAstId<N> {
+        let raw = self.erased_ast_id(item.syntax());
+        FileAstId { raw, _ty: PhantomData }
+    }
+    fn erased_ast_id(&self, item: &SyntaxNode) -> ErasedFileAstId {
+        let ptr = SyntaxNodePtr::new(item);
+        match self.arena.iter().find(|(_id, i)| **i == ptr) {
+            Some((it, _)) => it,
+            None => panic!(
+                "Can't find {:?} in AstIdMap:\n{:?}",
+                item,
+                self.arena.iter().map(|(_id, i)| i).collect::<Vec<_>>(),
+            ),
+        }
+    }
+
+    pub(crate) fn get<N: AstNode>(&self, id: FileAstId<N>) -> AstPtr<N> {
+        self.arena[id.raw].cast::<N>().unwrap()
+    }
+
+    fn alloc(&mut self, item: &SyntaxNode) -> ErasedFileAstId {
+        self.arena.alloc(SyntaxNodePtr::new(item))
+    }
+}
+
+/// Walks the subtree in bfs order, calling `f` for each node.
+fn bfs(node: &SyntaxNode, mut f: impl FnMut(SyntaxNode)) {
+    let mut curr_layer = vec![node.clone()];
+    let mut next_layer = vec![];
+    while !curr_layer.is_empty() {
+        curr_layer.drain(..).for_each(|node| {
+            next_layer.extend(node.children());
+            f(node);
+        });
+        std::mem::swap(&mut curr_layer, &mut next_layer);
+    }
+}
diff --git a/crates/ra_hir_expand/src/db.rs b/crates/ra_hir_expand/src/db.rs
new file mode 100644
index 000000000..a4ee9a529
--- /dev/null
+++ b/crates/ra_hir_expand/src/db.rs
@@ -0,0 +1,104 @@
+//! Defines database & queries for macro expansion.
+
+use std::sync::Arc;
+
+use mbe::MacroRules;
+use ra_db::{salsa, SourceDatabase};
+use ra_prof::profile;
+use ra_syntax::{AstNode, Parse, SyntaxNode};
+
+use crate::{
+    ast_id_map::AstIdMap, HirFileId, HirFileIdRepr, MacroCallId, MacroCallLoc, MacroDefId,
+    MacroFile, MacroFileKind,
+};
+
+// FIXME: rename to ExpandDatabase
+#[salsa::query_group(AstDatabaseStorage)]
+pub trait AstDatabase: SourceDatabase {
+    fn ast_id_map(&self, file_id: HirFileId) -> Arc<AstIdMap>;
+
+    #[salsa::transparent]
+    fn parse_or_expand(&self, file_id: HirFileId) -> Option<SyntaxNode>;
+
+    #[salsa::interned]
+    fn intern_macro(&self, macro_call: MacroCallLoc) -> MacroCallId;
+    fn macro_arg(&self, id: MacroCallId) -> Option<Arc<tt::Subtree>>;
+    fn macro_def(&self, id: MacroDefId) -> Option<Arc<mbe::MacroRules>>;
+    fn parse_macro(&self, macro_file: MacroFile) -> Option<Parse<SyntaxNode>>;
+    fn macro_expand(&self, macro_call: MacroCallId) -> Result<Arc<tt::Subtree>, String>;
+}
+
+pub(crate) fn ast_id_map(db: &dyn AstDatabase, file_id: HirFileId) -> Arc<AstIdMap> {
+    let map =
+        db.parse_or_expand(file_id).map_or_else(AstIdMap::default, |it| AstIdMap::from_source(&it));
+    Arc::new(map)
+}
+
+pub(crate) fn macro_def(db: &dyn AstDatabase, id: MacroDefId) -> Option<Arc<MacroRules>> {
+    let macro_call = id.ast_id.to_node(db);
+    let arg = macro_call.token_tree()?;
+    let (tt, _) = mbe::ast_to_token_tree(&arg).or_else(|| {
+        log::warn!("fail on macro_def to token tree: {:#?}", arg);
+        None
+    })?;
+    let rules = MacroRules::parse(&tt).ok().or_else(|| {
+        log::warn!("fail on macro_def parse: {:#?}", tt);
+        None
+    })?;
+    Some(Arc::new(rules))
+}
+
+pub(crate) fn macro_arg(db: &dyn AstDatabase, id: MacroCallId) -> Option<Arc<tt::Subtree>> {
+    let loc = db.lookup_intern_macro(id);
+    let macro_call = loc.ast_id.to_node(db);
+    let arg = macro_call.token_tree()?;
+    let (tt, _) = mbe::ast_to_token_tree(&arg)?;
+    Some(Arc::new(tt))
+}
+
+pub(crate) fn macro_expand(
+    db: &dyn AstDatabase,
+    id: MacroCallId,
+) -> Result<Arc<tt::Subtree>, String> {
+    let loc = db.lookup_intern_macro(id);
+    let macro_arg = db.macro_arg(id).ok_or("Fail to args in to tt::TokenTree")?;
+
+    let macro_rules = db.macro_def(loc.def).ok_or("Fail to find macro definition")?;
+    let tt = macro_rules.expand(&macro_arg).map_err(|err| format!("{:?}", err))?;
+    // Set a hard limit for the expanded tt
+    let count = tt.count();
+    if count > 65536 {
+        return Err(format!("Total tokens count exceed limit : count = {}", count));
+    }
+    Ok(Arc::new(tt))
+}
+
+pub(crate) fn parse_or_expand(db: &dyn AstDatabase, file_id: HirFileId) -> Option<SyntaxNode> {
+    match file_id.0 {
+        HirFileIdRepr::FileId(file_id) => Some(db.parse(file_id).tree().syntax().clone()),
+        HirFileIdRepr::MacroFile(macro_file) => {
+            db.parse_macro(macro_file).map(|it| it.syntax_node())
+        }
+    }
+}
+
+pub(crate) fn parse_macro(
+    db: &dyn AstDatabase,
+    macro_file: MacroFile,
+) -> Option<Parse<SyntaxNode>> {
+    let _p = profile("parse_macro_query");
+    let macro_call_id = macro_file.macro_call_id;
+    let tt = db
+        .macro_expand(macro_call_id)
+        .map_err(|err| {
+            // Note:
+            // The final goal we would like to make all parse_macro success,
+            // such that the following log will not call anyway.
+            log::warn!("fail on macro_parse: (reason: {})", err,);
+        })
+        .ok()?;
+    match macro_file.macro_file_kind {
+        MacroFileKind::Items => mbe::token_tree_to_items(&tt).ok().map(Parse::to_syntax),
+        MacroFileKind::Expr => mbe::token_tree_to_expr(&tt).ok().map(Parse::to_syntax),
+    }
+}
diff --git a/crates/ra_hir_expand/src/lib.rs b/crates/ra_hir_expand/src/lib.rs
new file mode 100644
index 000000000..6b3538673
--- /dev/null
+++ b/crates/ra_hir_expand/src/lib.rs
@@ -0,0 +1,161 @@
+//! `ra_hir_expand` deals with macro expansion.
+//!
+//! Specifically, it implements a concept of `MacroFile` -- a file whose syntax
+//! tree originates not from the text of some `FileId`, but from some macro
+//! expansion.
+
+pub mod db;
+pub mod ast_id_map;
+
+use std::hash::{Hash, Hasher};
+
+use ra_db::{salsa, CrateId, FileId};
+use ra_syntax::ast::{self, AstNode};
+
+use crate::{ast_id_map::FileAstId, db::AstDatabase};
+
+/// Input to the analyzer is a set of files, where each file is identified by
+/// `FileId` and contains source code. However, another source of source code in
+/// Rust are macros: each macro can be thought of as producing a "temporary
+/// file". To assign an id to such a file, we use the id of the macro call that
+/// produced the file. So, a `HirFileId` is either a `FileId` (source code
+/// written by user), or a `MacroCallId` (source code produced by macro).
+///
+/// What is a `MacroCallId`? Simplifying, it's a `HirFileId` of a file
+/// containing the call plus the offset of the macro call in the file. Note that
+/// this is a recursive definition! However, the size_of of `HirFileId` is
+/// finite (because everything bottoms out at the real `FileId`) and small
+/// (`MacroCallId` uses the location interner).
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub struct HirFileId(HirFileIdRepr);
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+enum HirFileIdRepr {
+    FileId(FileId),
+    MacroFile(MacroFile),
+}
+
+impl From<FileId> for HirFileId {
+    fn from(id: FileId) -> Self {
+        HirFileId(HirFileIdRepr::FileId(id))
+    }
+}
+
+impl From<MacroFile> for HirFileId {
+    fn from(id: MacroFile) -> Self {
+        HirFileId(HirFileIdRepr::MacroFile(id))
+    }
+}
+
+impl HirFileId {
+    /// For macro-expansion files, returns the file original source file the
+    /// expansion originated from.
+    pub fn original_file(self, db: &dyn AstDatabase) -> FileId {
+        match self.0 {
+            HirFileIdRepr::FileId(file_id) => file_id,
+            HirFileIdRepr::MacroFile(macro_file) => {
+                let loc = db.lookup_intern_macro(macro_file.macro_call_id);
+                loc.ast_id.file_id().original_file(db)
+            }
+        }
+    }
+
+    /// Get the crate which the macro lives in, if it is a macro file.
+    pub fn macro_crate(self, db: &dyn AstDatabase) -> Option<CrateId> {
+        match self.0 {
+            HirFileIdRepr::FileId(_) => None,
+            HirFileIdRepr::MacroFile(macro_file) => {
+                let loc = db.lookup_intern_macro(macro_file.macro_call_id);
+                Some(loc.def.krate)
+            }
+        }
+    }
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub struct MacroFile {
+    macro_call_id: MacroCallId,
+    macro_file_kind: MacroFileKind,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub enum MacroFileKind {
+    Items,
+    Expr,
+}
+
+/// `MacroCallId` identifies a particular macro invocation, like
+/// `println!("Hello, {}", world)`.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub struct MacroCallId(salsa::InternId);
+impl salsa::InternKey for MacroCallId {
+    fn from_intern_id(v: salsa::InternId) -> Self {
+        MacroCallId(v)
+    }
+    fn as_intern_id(&self) -> salsa::InternId {
+        self.0
+    }
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub struct MacroDefId {
+    pub krate: CrateId,
+    pub ast_id: AstId<ast::MacroCall>,
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Hash)]
+pub struct MacroCallLoc {
+    pub def: MacroDefId,
+    pub ast_id: AstId<ast::MacroCall>,
+}
+
+impl MacroCallId {
+    pub fn as_file(self, kind: MacroFileKind) -> HirFileId {
+        let macro_file = MacroFile { macro_call_id: self, macro_file_kind: kind };
+        macro_file.into()
+    }
+}
+
+/// `AstId` points to an AST node in any file.
+///
+/// It is stable across reparses, and can be used as salsa key/value.
+// FIXME: isn't this just a `Source<FileAstId<N>>` ?
+#[derive(Debug)]
+pub struct AstId<N: AstNode> {
+    file_id: HirFileId,
+    file_ast_id: FileAstId<N>,
+}
+
+impl<N: AstNode> Clone for AstId<N> {
+    fn clone(&self) -> AstId<N> {
+        *self
+    }
+}
+impl<N: AstNode> Copy for AstId<N> {}
+
+impl<N: AstNode> PartialEq for AstId<N> {
+    fn eq(&self, other: &Self) -> bool {
+        (self.file_id, self.file_ast_id) == (other.file_id, other.file_ast_id)
+    }
+}
+impl<N: AstNode> Eq for AstId<N> {}
+impl<N: AstNode> Hash for AstId<N> {
+    fn hash<H: Hasher>(&self, hasher: &mut H) {
+        (self.file_id, self.file_ast_id).hash(hasher);
+    }
+}
+
+impl<N: AstNode> AstId<N> {
+    pub fn new(file_id: HirFileId, file_ast_id: FileAstId<N>) -> AstId<N> {
+        AstId { file_id, file_ast_id }
+    }
+
+    pub fn file_id(&self) -> HirFileId {
+        self.file_id
+    }
+
+    pub fn to_node(&self, db: &dyn AstDatabase) -> N {
+        let root = db.parse_or_expand(self.file_id).unwrap();
+        db.ast_id_map(self.file_id).get(self.file_ast_id).to_node(&root)
+    }
+}