Merge #767

767: Extract project model to separate crate r=matklad a=flodiebold

I'm looking into creating a separate crate that would allow getting a HIR db for a project for 'batch' analyses, and this seems to be an obvious first step. We'd probably want to change the error handling to not rely on failure, though, right?

Co-authored-by: Florian Diebold <[email protected]>

4 files changed, 459 insertions, 0 deletions

diff --git a/crates/ra_project_model/Cargo.toml b/crates/ra_project_model/Cargo.toml
new file mode 100644
index 000000000..90f8b8398
--- /dev/null
+++ b/crates/ra_project_model/Cargo.toml
@@ -0,0 +1,21 @@
+[package]
+edition = "2018"
+name = "ra_project_model"
+version = "0.1.0"
+authors = ["Aleksey Kladov <[email protected]>"]
+
+[dependencies]
+log = "0.4.5"
+rustc-hash = "1.0"
+
+failure = "0.1.4"
+
+walkdir = "2.2.7"
+
+cargo_metadata = "0.7.0"
+
+ra_arena = { path = "../ra_arena" }
+ra_db = { path = "../ra_db" }
+
+[dev-dependencies]
+test_utils = { path = "../test_utils" }
diff --git a/crates/ra_project_model/src/cargo_workspace.rs b/crates/ra_project_model/src/cargo_workspace.rs
new file mode 100644
index 000000000..8adf463a6
--- /dev/null
+++ b/crates/ra_project_model/src/cargo_workspace.rs
@@ -0,0 +1,172 @@
+use std::path::{Path, PathBuf};
+
+use cargo_metadata::{MetadataCommand, CargoOpt};
+use ra_arena::{Arena, RawId, impl_arena_id};
+use rustc_hash::FxHashMap;
+use failure::format_err;
+
+use crate::Result;
+
+/// `CargoWorkspace` represents the logical structure of, well, a Cargo
+/// workspace. It pretty closely mirrors `cargo metadata` output.
+///
+/// Note that internally, rust analyzer uses a different structure:
+/// `CrateGraph`. `CrateGraph` is lower-level: it knows only about the crates,
+/// while this knows about `Pacakges` & `Targets`: purely cargo-related
+/// concepts.
+#[derive(Debug, Clone)]
+pub struct CargoWorkspace {
+    packages: Arena<Package, PackageData>,
+    targets: Arena<Target, TargetData>,
+}
+
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
+pub struct Package(RawId);
+impl_arena_id!(Package);
+
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
+pub struct Target(RawId);
+impl_arena_id!(Target);
+
+#[derive(Debug, Clone)]
+struct PackageData {
+    name: String,
+    manifest: PathBuf,
+    targets: Vec<Target>,
+    is_member: bool,
+    dependencies: Vec<PackageDependency>,
+}
+
+#[derive(Debug, Clone)]
+pub struct PackageDependency {
+    pub pkg: Package,
+    pub name: String,
+}
+
+#[derive(Debug, Clone)]
+struct TargetData {
+    pkg: Package,
+    name: String,
+    root: PathBuf,
+    kind: TargetKind,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum TargetKind {
+    Bin,
+    Lib,
+    Example,
+    Test,
+    Bench,
+    Other,
+}
+
+impl TargetKind {
+    fn new(kinds: &[String]) -> TargetKind {
+        for kind in kinds {
+            return match kind.as_str() {
+                "bin" => TargetKind::Bin,
+                "test" => TargetKind::Test,
+                "bench" => TargetKind::Bench,
+                "example" => TargetKind::Example,
+                _ if kind.contains("lib") => TargetKind::Lib,
+                _ => continue,
+            };
+        }
+        TargetKind::Other
+    }
+}
+
+impl Package {
+    pub fn name(self, ws: &CargoWorkspace) -> &str {
+        ws.packages[self].name.as_str()
+    }
+    pub fn root(self, ws: &CargoWorkspace) -> &Path {
+        ws.packages[self].manifest.parent().unwrap()
+    }
+    pub fn targets<'a>(self, ws: &'a CargoWorkspace) -> impl Iterator<Item = Target> + 'a {
+        ws.packages[self].targets.iter().cloned()
+    }
+    #[allow(unused)]
+    pub fn is_member(self, ws: &CargoWorkspace) -> bool {
+        ws.packages[self].is_member
+    }
+    pub fn dependencies<'a>(
+        self,
+        ws: &'a CargoWorkspace,
+    ) -> impl Iterator<Item = &'a PackageDependency> + 'a {
+        ws.packages[self].dependencies.iter()
+    }
+}
+
+impl Target {
+    pub fn package(self, ws: &CargoWorkspace) -> Package {
+        ws.targets[self].pkg
+    }
+    pub fn name(self, ws: &CargoWorkspace) -> &str {
+        ws.targets[self].name.as_str()
+    }
+    pub fn root(self, ws: &CargoWorkspace) -> &Path {
+        ws.targets[self].root.as_path()
+    }
+    pub fn kind(self, ws: &CargoWorkspace) -> TargetKind {
+        ws.targets[self].kind
+    }
+}
+
+impl CargoWorkspace {
+    pub fn from_cargo_metadata(cargo_toml: &Path) -> Result<CargoWorkspace> {
+        let mut meta = MetadataCommand::new();
+        meta.manifest_path(cargo_toml).features(CargoOpt::AllFeatures);
+        if let Some(parent) = cargo_toml.parent() {
+            meta.current_dir(parent);
+        }
+        let meta = meta.exec().map_err(|e| format_err!("cargo metadata failed: {}", e))?;
+        let mut pkg_by_id = FxHashMap::default();
+        let mut packages = Arena::default();
+        let mut targets = Arena::default();
+
+        let ws_members = &meta.workspace_members;
+
+        for meta_pkg in meta.packages {
+            let is_member = ws_members.contains(&meta_pkg.id);
+            let pkg = packages.alloc(PackageData {
+                name: meta_pkg.name.into(),
+                manifest: meta_pkg.manifest_path.clone(),
+                targets: Vec::new(),
+                is_member,
+                dependencies: Vec::new(),
+            });
+            let pkg_data = &mut packages[pkg];
+            pkg_by_id.insert(meta_pkg.id.clone(), pkg);
+            for meta_tgt in meta_pkg.targets {
+                let tgt = targets.alloc(TargetData {
+                    pkg,
+                    name: meta_tgt.name.into(),
+                    root: meta_tgt.src_path.clone(),
+                    kind: TargetKind::new(meta_tgt.kind.as_slice()),
+                });
+                pkg_data.targets.push(tgt);
+            }
+        }
+        let resolve = meta.resolve.expect("metadata executed with deps");
+        for node in resolve.nodes {
+            let source = pkg_by_id[&node.id];
+            for dep_node in node.deps {
+                let dep =
+                    PackageDependency { name: dep_node.name.into(), pkg: pkg_by_id[&dep_node.pkg] };
+                packages[source].dependencies.push(dep);
+            }
+        }
+
+        Ok(CargoWorkspace { packages, targets })
+    }
+
+    pub fn packages<'a>(&'a self) -> impl Iterator<Item = Package> + 'a {
+        self.packages.iter().map(|(id, _pkg)| id)
+    }
+
+    pub fn target_by_root(&self, root: &Path) -> Option<Target> {
+        self.packages().filter_map(|pkg| pkg.targets(self).find(|it| it.root(self) == root)).next()
+    }
+}
diff --git a/crates/ra_project_model/src/lib.rs b/crates/ra_project_model/src/lib.rs
new file mode 100644
index 000000000..3b1e07149
--- /dev/null
+++ b/crates/ra_project_model/src/lib.rs
@@ -0,0 +1,130 @@
+mod cargo_workspace;
+mod sysroot;
+
+use std::path::{Path, PathBuf};
+
+use failure::bail;
+use rustc_hash::FxHashMap;
+
+use ra_db::{CrateGraph, FileId};
+
+pub use crate::{
+    cargo_workspace::{CargoWorkspace, Package, Target, TargetKind},
+    sysroot::Sysroot,
+};
+
+// TODO use proper error enum
+pub type Result<T> = ::std::result::Result<T, ::failure::Error>;
+
+#[derive(Debug, Clone)]
+pub struct ProjectWorkspace {
+    pub cargo: CargoWorkspace,
+    pub sysroot: Sysroot,
+}
+
+impl ProjectWorkspace {
+    pub fn discover(path: &Path) -> Result<ProjectWorkspace> {
+        let cargo_toml = find_cargo_toml(path)?;
+        let cargo = CargoWorkspace::from_cargo_metadata(&cargo_toml)?;
+        let sysroot = Sysroot::discover(&cargo_toml)?;
+        let res = ProjectWorkspace { cargo, sysroot };
+        Ok(res)
+    }
+
+    pub fn to_crate_graph(&self, load: &mut dyn FnMut(&Path) -> Option<FileId>) -> CrateGraph {
+        let mut crate_graph = CrateGraph::default();
+        let mut sysroot_crates = FxHashMap::default();
+        for krate in self.sysroot.crates() {
+            if let Some(file_id) = load(krate.root(&self.sysroot)) {
+                sysroot_crates.insert(krate, crate_graph.add_crate_root(file_id));
+            }
+        }
+        for from in self.sysroot.crates() {
+            for to in from.deps(&self.sysroot) {
+                let name = to.name(&self.sysroot);
+                if let (Some(&from), Some(&to)) =
+                    (sysroot_crates.get(&from), sysroot_crates.get(&to))
+                {
+                    if let Err(_) = crate_graph.add_dep(from, name.into(), to) {
+                        log::error!("cyclic dependency between sysroot crates")
+                    }
+                }
+            }
+        }
+
+        let libstd = self.sysroot.std().and_then(|it| sysroot_crates.get(&it).map(|&it| it));
+
+        let mut pkg_to_lib_crate = FxHashMap::default();
+        let mut pkg_crates = FxHashMap::default();
+        // Next, create crates for each package, target pair
+        for pkg in self.cargo.packages() {
+            let mut lib_tgt = None;
+            for tgt in pkg.targets(&self.cargo) {
+                let root = tgt.root(&self.cargo);
+                if let Some(file_id) = load(root) {
+                    let crate_id = crate_graph.add_crate_root(file_id);
+                    if tgt.kind(&self.cargo) == TargetKind::Lib {
+                        lib_tgt = Some(crate_id);
+                        pkg_to_lib_crate.insert(pkg, crate_id);
+                    }
+                    pkg_crates.entry(pkg).or_insert_with(Vec::new).push(crate_id);
+                }
+            }
+
+            // Set deps to the std and to the lib target of the current package
+            for &from in pkg_crates.get(&pkg).into_iter().flatten() {
+                if let Some(to) = lib_tgt {
+                    if to != from {
+                        if let Err(_) = crate_graph.add_dep(from, pkg.name(&self.cargo).into(), to)
+                        {
+                            log::error!(
+                                "cyclic dependency between targets of {}",
+                                pkg.name(&self.cargo)
+                            )
+                        }
+                    }
+                }
+                if let Some(std) = libstd {
+                    if let Err(_) = crate_graph.add_dep(from, "std".into(), std) {
+                        log::error!("cyclic dependency on std for {}", pkg.name(&self.cargo))
+                    }
+                }
+            }
+        }
+
+        // Now add a dep ednge from all targets of upstream to the lib
+        // target of downstream.
+        for pkg in self.cargo.packages() {
+            for dep in pkg.dependencies(&self.cargo) {
+                if let Some(&to) = pkg_to_lib_crate.get(&dep.pkg) {
+                    for &from in pkg_crates.get(&pkg).into_iter().flatten() {
+                        if let Err(_) = crate_graph.add_dep(from, dep.name.clone().into(), to) {
+                            log::error!(
+                                "cyclic dependency {} -> {}",
+                                pkg.name(&self.cargo),
+                                dep.pkg.name(&self.cargo)
+                            )
+                        }
+                    }
+                }
+            }
+        }
+
+        crate_graph
+    }
+}
+
+fn find_cargo_toml(path: &Path) -> Result<PathBuf> {
+    if path.ends_with("Cargo.toml") {
+        return Ok(path.to_path_buf());
+    }
+    let mut curr = Some(path);
+    while let Some(path) = curr {
+        let candidate = path.join("Cargo.toml");
+        if candidate.exists() {
+            return Ok(candidate);
+        }
+        curr = path.parent();
+    }
+    bail!("can't find Cargo.toml at {}", path.display())
+}
diff --git a/crates/ra_project_model/src/sysroot.rs b/crates/ra_project_model/src/sysroot.rs
new file mode 100644
index 000000000..8b87aa7bd
--- /dev/null
+++ b/crates/ra_project_model/src/sysroot.rs
@@ -0,0 +1,136 @@
+use std::{
+    path::{Path, PathBuf},
+    process::Command,
+};
+
+use ra_arena::{Arena, RawId, impl_arena_id};
+
+use crate::Result;
+
+#[derive(Debug, Clone)]
+pub struct Sysroot {
+    crates: Arena<SysrootCrate, SysrootCrateData>,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
+pub struct SysrootCrate(RawId);
+impl_arena_id!(SysrootCrate);
+
+#[derive(Debug, Clone)]
+struct SysrootCrateData {
+    name: String,
+    root: PathBuf,
+    deps: Vec<SysrootCrate>,
+}
+
+impl Sysroot {
+    pub fn std(&self) -> Option<SysrootCrate> {
+        self.by_name("std")
+    }
+
+    pub fn crates<'a>(&'a self) -> impl Iterator<Item = SysrootCrate> + 'a {
+        self.crates.iter().map(|(id, _data)| id)
+    }
+
+    pub fn discover(cargo_toml: &Path) -> Result<Sysroot> {
+        let rustc_output = Command::new("rustc")
+            .current_dir(cargo_toml.parent().unwrap())
+            .args(&["--print", "sysroot"])
+            .output()?;
+        if !rustc_output.status.success() {
+            failure::bail!("failed to locate sysroot")
+        }
+        let stdout = String::from_utf8(rustc_output.stdout)?;
+        let sysroot_path = Path::new(stdout.trim());
+        let src = sysroot_path.join("lib/rustlib/src/rust/src");
+        if !src.exists() {
+            failure::bail!(
+                "can't load standard library from sysroot\n\
+                 {:?}\n\
+                 try running `rustup component add rust-src`",
+                src,
+            );
+        }
+
+        let mut sysroot = Sysroot { crates: Arena::default() };
+        for name in SYSROOT_CRATES.trim().lines() {
+            let root = src.join(format!("lib{}", name)).join("lib.rs");
+            if root.exists() {
+                sysroot.crates.alloc(SysrootCrateData {
+                    name: name.into(),
+                    root,
+                    deps: Vec::new(),
+                });
+            }
+        }
+        if let Some(std) = sysroot.std() {
+            for dep in STD_DEPS.trim().lines() {
+                if let Some(dep) = sysroot.by_name(dep) {
+                    sysroot.crates[std].deps.push(dep)
+                }
+            }
+        }
+        Ok(sysroot)
+    }
+
+    fn by_name(&self, name: &str) -> Option<SysrootCrate> {
+        self.crates.iter().find(|(_id, data)| data.name == name).map(|(id, _data)| id)
+    }
+}
+
+impl SysrootCrate {
+    pub fn name(self, sysroot: &Sysroot) -> &str {
+        &sysroot.crates[self].name
+    }
+    pub fn root(self, sysroot: &Sysroot) -> &Path {
+        sysroot.crates[self].root.as_path()
+    }
+    pub fn root_dir(self, sysroot: &Sysroot) -> &Path {
+        self.root(sysroot).parent().unwrap()
+    }
+    pub fn deps<'a>(self, sysroot: &'a Sysroot) -> impl Iterator<Item = SysrootCrate> + 'a {
+        sysroot.crates[self].deps.iter().map(|&it| it)
+    }
+}
+
+const SYSROOT_CRATES: &str = "
+std
+core
+alloc
+collections
+libc
+panic_unwind
+proc_macro
+rustc_unicode
+std_unicode
+test
+alloc_jemalloc
+alloc_system
+compiler_builtins
+getopts
+panic_unwind
+panic_abort
+rand
+term
+unwind
+build_helper
+rustc_asan
+rustc_lsan
+rustc_msan
+rustc_tsan
+syntax";
+
+const STD_DEPS: &str = "
+alloc
+alloc_jemalloc
+alloc_system
+core
+panic_abort
+rand
+compiler_builtins
+unwind
+rustc_asan
+rustc_lsan
+rustc_msan
+rustc_tsan
+build_helper";