refactor(nix-compat): Properly encapsulate store path construction r/6088

Before there was code scattered about (e.g. text hashing module and
derivation output computation) constructing store paths from low level
building blocks --- there was some duplication and it was easy to make
nonsense store paths.

Now, we have roughly the same "safe-ish" ways of constructing them as
C++ Nix, and only those are exposed:

- Make text hashed content-addressed store paths

- Make other content-addressed store paths

- Make input-addressed fixed output hashes

Change-Id: I122a3ee0802b4f45ae386306b95b698991be89c8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8411
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI

2 files changed, 188 insertions, 50 deletions

diff --git a/tvix/nix-compat/src/store_path/mod.rs b/tvix/nix-compat/src/store_path/mod.rs
index 0e004ccd7a..7fe425d8ed 100644
--- a/tvix/nix-compat/src/store_path/mod.rs
+++ b/tvix/nix-compat/src/store_path/mod.rs
@@ -4,10 +4,7 @@ use thiserror::Error;
 
 mod utils;
 
-pub use utils::{
-    build_store_path_from_fingerprint, build_store_path_from_references, compress_hash,
-    hash_placeholder,
-};
+pub use utils::*;
 
 pub const DIGEST_SIZE: usize = 20;
 // lazy_static doesn't allow us to call NIXBASE32.encode_len(), so we ran it
@@ -19,19 +16,32 @@ pub const ENCODED_DIGEST_SIZE: usize = 32;
 pub const STORE_DIR: &str = "/nix/store";
 pub const STORE_DIR_WITH_SLASH: &str = "/nix/store/";
 
-/// Errors that can occur during the validation of name characters.
+/// Errors that can occur when parsing a literal store path
 #[derive(Debug, PartialEq, Eq, Error)]
 pub enum Error {
     #[error("Dash is missing between hash and name")]
     MissingDash(),
     #[error("Hash encoding is invalid: {0}")]
     InvalidHashEncoding(Nixbase32DecodeError),
-    #[error("Invalid name: {0}")]
-    InvalidName(String),
+    #[error("{0}")]
+    InvalidName(NameError),
     #[error("Tried to parse an absolute path which was missing the store dir prefix.")]
     MissingStoreDir(),
 }
 
+/// Errors that can occur during the validation of name characters.
+#[derive(Debug, PartialEq, Eq, Error)]
+pub enum NameError {
+    #[error("Invalid name: {0}")]
+    InvalidName(String),
+}
+
+impl From<NameError> for Error {
+    fn from(e: NameError) -> Self {
+        Self::InvalidName(e)
+    }
+}
+
 /// Represents a path in the Nix store (a direct child of [STORE_DIR]).
 ///
 /// It starts with a digest (20 bytes), [crate::nixbase32]-encoded,
@@ -56,7 +66,7 @@ impl StorePath {
         // - 1 dash
         // - 1 character for the name
         if s.len() < ENCODED_DIGEST_SIZE + 2 {
-            return Err(Error::InvalidName("".to_string()));
+            Err(NameError::InvalidName("".to_string()))?;
         }
 
         let digest = match nixbase32::decode(s[..ENCODED_DIGEST_SIZE].as_bytes()) {
@@ -92,7 +102,7 @@ impl StorePath {
     }
 
     /// Checks a given &str to match the restrictions for store path names.
-    pub fn validate_name(s: &str) -> Result<(), Error> {
+    pub fn validate_name(s: &str) -> Result<(), NameError> {
         for c in s.chars() {
             if c.is_ascii_alphanumeric()
                 || c == '-'
@@ -105,7 +115,7 @@ impl StorePath {
                 continue;
             }
 
-            return Err(Error::InvalidName(s.to_string()));
+            return Err(NameError::InvalidName(s.to_string()));
         }
 
         Ok(())
diff --git a/tvix/nix-compat/src/store_path/utils.rs b/tvix/nix-compat/src/store_path/utils.rs
index fa1908ee35..c5fc48a94f 100644
--- a/tvix/nix-compat/src/store_path/utils.rs
+++ b/tvix/nix-compat/src/store_path/utils.rs
@@ -1,9 +1,24 @@
 use crate::nixbase32;
-use crate::nixhash::NixHash;
+use crate::nixhash::{HashAlgo, NixHash, NixHashWithMode};
 use crate::store_path::StorePath;
 use sha2::{Digest, Sha256};
+use thiserror::Error;
 
-use super::Error;
+use super::{NameError, STORE_DIR};
+
+/// Errors that can occur when creating a content-addressed store path.
+///
+/// This wraps the main [Error] which is just about invalid store path names.
+#[derive(Debug, PartialEq, Eq, Error)]
+pub enum BuildStorePathError {
+    #[error("{0}")]
+    InvalidName(NameError),
+    /// This error occurs when we have references outside the SHA-256 +
+    /// Recursive case. The restriction comes from upstream Nix. It may be
+    /// lifted at some point but there isn't a pressing need to anticipate that.
+    #[error("References were not supported as much as requested")]
+    InvalidReference(),
+}
 
 /// compress_hash takes an arbitrarily long sequence of bytes (usually
 /// a hash digest), and returns a sequence of bytes of length
@@ -27,33 +42,107 @@ pub fn compress_hash<const OUTPUT_SIZE: usize>(input: &[u8]) -> [u8; OUTPUT_SIZE
 
 /// This builds a store path, by calculating the text_hash_string of either a
 /// derivation or a literal text file that may contain references.
-pub fn build_store_path_from_references<
-    S: AsRef<str>,
-    I: IntoIterator<Item = S>,
-    C: AsRef<[u8]>,
->(
+pub fn build_text_path<S: AsRef<str>, I: IntoIterator<Item = S>, C: AsRef<[u8]>>(
     name: &str,
     content: C,
     references: I,
-) -> Result<StorePath, Error> {
-    let text_hash_str = text_hash_string(name, content, references);
-    build_store_path_from_fingerprint(name, &text_hash_str)
+) -> Result<StorePath, NameError> {
+    build_store_path_from_fingerprint_parts(
+        &make_type("text", references, false),
+        // the nix_hash_string representation of the sha256 digest of some contents
+        &{
+            let content_digest = {
+                let hasher = Sha256::new_with_prefix(content);
+                hasher.finalize()
+            };
+            NixHash::new(crate::nixhash::HashAlgo::Sha256, content_digest.to_vec())
+        },
+        name,
+    )
+}
+
+/// This builds a more "regular" content-addressed store path
+pub fn build_regular_ca_path<S: AsRef<str>, I: IntoIterator<Item = S>>(
+    name: &str,
+    hash_with_mode: &NixHashWithMode,
+    references: I,
+    self_reference: bool,
+) -> Result<StorePath, BuildStorePathError> {
+    match &hash_with_mode {
+        NixHashWithMode::Recursive(
+            ref hash @ NixHash {
+                algo: HashAlgo::Sha256,
+                ..
+            },
+        ) => build_store_path_from_fingerprint_parts(
+            &make_type("source", references, self_reference),
+            hash,
+            name,
+        )
+        .map_err(BuildStorePathError::InvalidName),
+        _ => {
+            if references.into_iter().next().is_some() {
+                return Err(BuildStorePathError::InvalidReference());
+            }
+            if self_reference {
+                return Err(BuildStorePathError::InvalidReference());
+            }
+            build_store_path_from_fingerprint_parts(
+                "output:out",
+                &{
+                    let content_digest = {
+                        let mut hasher = Sha256::new_with_prefix("fixed:out:");
+                        hasher.update(hash_with_mode.mode().prefix());
+                        hasher.update(hash_with_mode.digest().algo.to_string());
+                        hasher.update(":");
+                        hasher.update(
+                            &data_encoding::HEXLOWER.encode(&hash_with_mode.digest().digest),
+                        );
+                        hasher.update(":");
+                        hasher.finalize()
+                    };
+                    NixHash::new(crate::nixhash::HashAlgo::Sha256, content_digest.to_vec())
+                },
+                name,
+            )
+            .map_err(BuildStorePathError::InvalidName)
+        }
+    }
 }
 
-/// This builds a store path from a fingerprint.
-/// Usually, that function is used from [build_store_path_from_references] and
+/// This builds an input-addressed store path
+///
+/// Input-addresed store paths are always derivation outputs, the "input" in question is the
+/// derivation and its closure.
+pub fn build_output_path(
+    drv_hash: &NixHash,
+    output_name: &str,
+    output_path_name: &str,
+) -> Result<StorePath, NameError> {
+    build_store_path_from_fingerprint_parts(
+        &(String::from("output:") + output_name),
+        drv_hash,
+        &output_path_name,
+    )
+}
+
+/// This builds a store path from fingerprint parts.
+/// Usually, that function is used from [build_text_path] and
 /// passed a "text hash string" (starting with "text:" as fingerprint),
 /// but other fingerprints starting with "output:" are also used in Derivation
 /// output path calculation.
 ///
 /// The fingerprint is hashed with sha256, its digest is compressed to 20 bytes,
 /// and nixbase32-encoded (32 characters).
-pub fn build_store_path_from_fingerprint(
+fn build_store_path_from_fingerprint_parts(
+    ty: &str,
+    hash: &NixHash,
     name: &str,
-    fingerprint: &str,
-) -> Result<StorePath, Error> {
+) -> Result<StorePath, NameError> {
+    let fingerprint =
+        String::from(ty) + ":" + &hash.to_nix_hash_string() + ":" + STORE_DIR + ":" + name;
     let digest = {
-        let hasher = Sha256::new_with_prefix(fingerprint);
+        let hasher = Sha256::new_with_prefix(&fingerprint);
         hasher.finalize()
     };
     let compressed = compress_hash::<20>(&digest);
@@ -74,31 +163,21 @@ pub fn build_store_path_from_fingerprint(
 ///  - the nix_hash_string representation of the sha256 digest of some contents
 ///  - the value of `storeDir`
 ///  - the name
-pub fn text_hash_string<S: AsRef<str>, I: IntoIterator<Item = S>, C: AsRef<[u8]>>(
-    name: &str,
-    content: C,
+fn make_type<S: AsRef<str>, I: IntoIterator<Item = S>>(
+    ty: &str,
     references: I,
+    self_ref: bool,
 ) -> String {
-    let mut s = String::from("text:");
+    let mut s = String::from(ty);
 
     for reference in references {
-        s.push_str(reference.as_ref());
         s.push(':');
+        s.push_str(reference.as_ref());
     }
 
-    // the nix_hash_string representation of the sha256 digest of some contents
-    s.push_str(
-        &{
-            let content_digest = {
-                let hasher = Sha256::new_with_prefix(content);
-                hasher.finalize()
-            };
-            NixHash::new(crate::nixhash::HashAlgo::Sha256, content_digest.to_vec())
-        }
-        .to_nix_hash_string(),
-    );
-
-    s.push_str(&format!(":{}:{}", crate::store_path::STORE_DIR, name));
+    if self_ref {
+        s.push_str(":self");
+    }
 
     s
 }
@@ -121,16 +200,17 @@ pub fn hash_placeholder(name: &str) -> String {
 
 #[cfg(test)]
 mod test {
-    use crate::store_path::build_store_path_from_references;
+    use super::*;
+    use crate::nixhash::{NixHash, NixHashWithMode};
 
     #[test]
-    fn build_store_path_with_zero_references() {
+    fn build_text_path_with_zero_references() {
         // This hash should match `builtins.toFile`, e.g.:
         //
         // nix-repl> builtins.toFile "foo" "bar"
         // "/nix/store/vxjiwkjkn7x4079qvh1jkl5pn05j2aw0-foo"
 
-        let store_path = build_store_path_from_references("foo", "bar", Vec::<String>::new())
+        let store_path = build_text_path("foo", "bar", Vec::<String>::new())
             .expect("build_store_path() should succeed");
 
         assert_eq!(
@@ -140,17 +220,17 @@ mod test {
     }
 
     #[test]
-    fn build_store_path_with_non_zero_references() {
+    fn build_text_path_with_non_zero_references() {
         // This hash should match:
         //
         // nix-repl> builtins.toFile "baz" "${builtins.toFile "foo" "bar"}"
         // "/nix/store/5xd714cbfnkz02h2vbsj4fm03x3f15nf-baz"
 
-        let inner = build_store_path_from_references("foo", "bar", Vec::<String>::new())
+        let inner = build_text_path("foo", "bar", Vec::<String>::new())
             .expect("path_with_references() should succeed");
         let inner_path = inner.to_absolute_path();
 
-        let outer = build_store_path_from_references("baz", &inner_path, vec![inner_path.as_str()])
+        let outer = build_text_path("baz", &inner_path, vec![inner_path.as_str()])
             .expect("path_with_references() should succeed");
 
         assert_eq!(
@@ -158,4 +238,52 @@ mod test {
             "/nix/store/5xd714cbfnkz02h2vbsj4fm03x3f15nf-baz"
         );
     }
+
+    #[test]
+    fn build_sha1_path() {
+        let outer = build_regular_ca_path(
+            "bar",
+            &NixHashWithMode::Recursive(NixHash {
+                algo: HashAlgo::Sha1,
+                digest: data_encoding::HEXLOWER
+                    .decode(b"0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33")
+                    .expect("hex should decode"),
+            }),
+            Vec::<String>::new(),
+            false,
+        )
+        .expect("path_with_references() should succeed");
+
+        assert_eq!(
+            outer.to_absolute_path().as_str(),
+            "/nix/store/mp57d33657rf34lzvlbpfa1gjfv5gmpg-bar"
+        );
+    }
+
+    #[test]
+    fn build_store_path_with_non_zero_references() {
+        // This hash should match:
+        //
+        // nix-repl> builtins.toFile "baz" "${builtins.toFile "foo" "bar"}"
+        // "/nix/store/5xd714cbfnkz02h2vbsj4fm03x3f15nf-baz"
+        //
+        // $ nix store make-content-addressed /nix/store/5xd714cbfnkz02h2vbsj4fm03x3f15nf-baz
+        // rewrote '/nix/store/5xd714cbfnkz02h2vbsj4fm03x3f15nf-baz' to '/nix/store/s89y431zzhmdn3k8r96rvakryddkpv2v-baz'
+        let outer = build_regular_ca_path(
+            "baz",
+            &NixHashWithMode::Recursive(NixHash {
+                algo: HashAlgo::Sha256,
+                digest: nixbase32::decode(b"1xqkzcb3909fp07qngljr4wcdnrh1gdam1m2n29i6hhrxlmkgkv1")
+                    .expect("hex should decode"),
+            }),
+            vec!["/nix/store/dxwkwjzdaq7ka55pkk252gh32bgpmql4-foo"],
+            false,
+        )
+        .expect("path_with_references() should succeed");
+
+        assert_eq!(
+            outer.to_absolute_path().as_str(),
+            "/nix/store/s89y431zzhmdn3k8r96rvakryddkpv2v-baz"
+        );
+    }
 }