feat(tvix/cli): add known_paths module r/5669

This module implements types used to track the set of known paths in
the context of an evaluation.

These are used to determine the build references of a derivation.

Change-Id: I81e15ae33632784e699128916485751613b231a3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7816
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>

3 files changed, 143 insertions, 5 deletions

diff --git a/tvix/cli/src/known_paths.rs b/tvix/cli/src/known_paths.rs
new file mode 100644
index 0000000000..10191c967e
--- /dev/null
+++ b/tvix/cli/src/known_paths.rs
@@ -0,0 +1,114 @@
+//! This module implements logic required for persisting known paths
+//! during an evaluation.
+//!
+//! Tvix needs to be able to keep track of each Nix store path that it
+//! knows about during the scope of a single evaluation and its
+//! related builds.
+//!
+//! This data is required to scan derivation inputs for the build
+//! references (the "build closure") that they make use of.
+//!
+//! Please see //tvix/eval/docs/build-references.md for more
+//! information.
+
+use crate::refscan::ReferenceScanner;
+use std::{
+    collections::{hash_map, BTreeSet, HashMap},
+    ops::Index,
+};
+
+#[derive(Debug, PartialEq)]
+pub enum PathType {
+    /// A literal derivation (`.drv`-file), and the *names* of its outputs.
+    Derivation { output_names: BTreeSet<String> },
+
+    /// An output of a derivation, its name, and the path of its derivation.
+    Output { name: String, derivation: String },
+
+    /// A plain store path (e.g. source files copied to the store).
+    Plain,
+}
+
+pub struct KnownPaths {
+    /// All known paths, and their associated [`PathType`].
+    paths: HashMap<String, PathType>,
+}
+
+impl Index<&str> for KnownPaths {
+    type Output = PathType;
+
+    fn index(&self, index: &str) -> &Self::Output {
+        &self.paths[index]
+    }
+}
+
+impl KnownPaths {
+    /// Mark a plain path as known.
+    pub fn plain<S: ToString>(&mut self, path: S) {
+        self.paths.insert(path.to_string(), PathType::Plain);
+    }
+
+    /// Mark a derivation as known.
+    pub fn drv<P: ToString, O: ToString>(&mut self, path: P, outputs: &[O]) {
+        match self.paths.entry(path.to_string()) {
+            hash_map::Entry::Occupied(mut entry) => {
+                for output in outputs {
+                    match entry.get_mut() {
+                        PathType::Derivation {
+                            ref mut output_names,
+                        } => {
+                            output_names.insert(output.to_string());
+                        }
+
+                        // Branches like this explicitly panic right now to find odd
+                        // situations where something unexpected is done with the
+                        // same path being inserted twice as different types.
+                        _ => panic!(
+                            "bug: {} is already a known path, but not a derivation!",
+                            path.to_string()
+                        ),
+                    }
+                }
+            }
+
+            hash_map::Entry::Vacant(entry) => {
+                let output_names = outputs.iter().map(|o| o.to_string()).collect();
+                entry.insert(PathType::Derivation { output_names });
+            }
+        }
+    }
+
+    /// Mark a derivation output path as known.
+    pub fn output<P: ToString, N: ToString, D: ToString>(
+        &mut self,
+        output_path: P,
+        name: N,
+        drv_path: D,
+    ) {
+        match self.paths.entry(output_path.to_string()) {
+            hash_map::Entry::Occupied(entry) => {
+                /* nothing to do, really! */
+                debug_assert!(
+                    *entry.get()
+                        == PathType::Output {
+                            name: name.to_string(),
+                            derivation: drv_path.to_string(),
+                        }
+                );
+            }
+
+            hash_map::Entry::Vacant(entry) => {
+                entry.insert(PathType::Output {
+                    name: name.to_string(),
+                    derivation: drv_path.to_string(),
+                });
+            }
+        }
+    }
+
+    /// Create a reference scanner from the current set of known paths.
+    pub fn reference_scanner<'a>(&'a self) -> ReferenceScanner<'a> {
+        let candidates: Vec<&'a str> = self.paths.keys().map(|s| s.as_str()).collect();
+        ReferenceScanner::new(candidates)
+    }
+}
diff --git a/tvix/cli/src/main.rs b/tvix/cli/src/main.rs
index 0f837b346c..42297d365b 100644
--- a/tvix/cli/src/main.rs
+++ b/tvix/cli/src/main.rs
@@ -1,3 +1,4 @@
+mod known_paths;
 mod nix_compat;
 mod refscan;
 
diff --git a/tvix/cli/src/refscan.rs b/tvix/cli/src/refscan.rs
index 31fccb797a..6ef486385d 100644
--- a/tvix/cli/src/refscan.rs
+++ b/tvix/cli/src/refscan.rs
@@ -13,17 +13,34 @@ use std::io;
 
 /// Represents a "primed" reference scanner with an automaton that knows the set
 /// of store paths to scan for.
-pub struct ReferenceScanner<'c, 's> {
-    candidates: &'c [&'s str],
+pub struct ReferenceScanner<'s> {
+    candidates: Vec<&'s str>,
     searcher: AhoCorasick,
     matches: BTreeSet<&'s str>,
 }
 
-impl<'c, 's> ReferenceScanner<'c, 's> {
+pub trait ToOwnedVec<T> {
+    fn to_owned_vec(self) -> Vec<T>;
+}
+
+impl<T: Clone> ToOwnedVec<T> for &[T] {
+    fn to_owned_vec(self) -> Vec<T> {
+        self.to_vec()
+    }
+}
+
+impl<T> ToOwnedVec<T> for Vec<T> {
+    fn to_owned_vec(self) -> Vec<T> {
+        self
+    }
+}
+
+impl<'s> ReferenceScanner<'s> {
     /// Construct a new `ReferenceScanner` that knows how to scan for the given
     /// candidate store paths.
-    pub fn new(candidates: &'c [&'s str]) -> Self {
-        let searcher = AhoCorasick::new_auto_configured(candidates);
+    pub fn new<V: ToOwnedVec<&'s str>>(candidates: V) -> Self {
+        let candidates = candidates.to_owned_vec();
+        let searcher = AhoCorasick::new_auto_configured(&candidates);
 
         ReferenceScanner {
             searcher,
@@ -70,6 +87,12 @@ mod tests {
     // The actual derivation of `nixpkgs.hello`.
     const HELLO_DRV: &'static str = r#"Derive([("out","/nix/store/33l4p0pn0mybmqzaxfkpppyh7vx1c74p-hello-2.12.1","","")],[("/nix/store/6z1jfnqqgyqr221zgbpm30v91yfj3r45-bash-5.1-p16.drv",["out"]),("/nix/store/ap9g09fxbicj836zm88d56dn3ff4clxl-stdenv-linux.drv",["out"]),("/nix/store/pf80kikyxr63wrw56k00i1kw6ba76qik-hello-2.12.1.tar.gz.drv",["out"])],["/nix/store/9krlzvny65gdc8s7kpb6lkx8cd02c25b-default-builder.sh"],"x86_64-linux","/nix/store/4xw8n979xpivdc46a9ndcvyhwgif00hz-bash-5.1-p16/bin/bash",["-e","/nix/store/9krlzvny65gdc8s7kpb6lkx8cd02c25b-default-builder.sh"],[("buildInputs",""),("builder","/nix/store/4xw8n979xpivdc46a9ndcvyhwgif00hz-bash-5.1-p16/bin/bash"),("cmakeFlags",""),("configureFlags",""),("depsBuildBuild",""),("depsBuildBuildPropagated",""),("depsBuildTarget",""),("depsBuildTargetPropagated",""),("depsHostHost",""),("depsHostHostPropagated",""),("depsTargetTarget",""),("depsTargetTargetPropagated",""),("doCheck","1"),("doInstallCheck",""),("mesonFlags",""),("name","hello-2.12.1"),("nativeBuildInputs",""),("out","/nix/store/33l4p0pn0mybmqzaxfkpppyh7vx1c74p-hello-2.12.1"),("outputs","out"),("patches",""),("pname","hello"),("propagatedBuildInputs",""),("propagatedNativeBuildInputs",""),("src","/nix/store/pa10z4ngm0g83kx9mssrqzz30s84vq7k-hello-2.12.1.tar.gz"),("stdenv","/nix/store/cp65c8nk29qq5cl1wyy5qyw103cwmax7-stdenv-linux"),("strictDeps",""),("system","x86_64-linux"),("version","2.12.1")])"#;
 
+    impl<T: Clone, const N: usize> ToOwnedVec<T> for &[T; N] {
+        fn to_owned_vec(self) -> Vec<T> {
+            self.to_vec()
+        }
+    }
+
     #[test]
     fn test_empty() {
         let mut scanner = ReferenceScanner::new(&[]);