refactor(tvix/nix-compat): add text_hash_string function r/6057

Use it to calculate the text_hash_string, which is then used in the
calculate_derivation_path and path_with_references functions.

Relates to b/263.

Change-Id: I7478825e2a23a11224212fea5e3fd06daa97d5e5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8364
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI

4 files changed, 58 insertions, 60 deletions

diff --git a/tvix/nix-compat/src/derivation/mod.rs b/tvix/nix-compat/src/derivation/mod.rs
index 281fe3a3c8..53bdccef6d 100644
--- a/tvix/nix-compat/src/derivation/mod.rs
+++ b/tvix/nix-compat/src/derivation/mod.rs
@@ -1,6 +1,7 @@
 use crate::{
     nixhash::HashAlgo,
     store_path::{self, StorePath},
+    texthash::text_hash_string,
 };
 use serde::{Deserialize, Serialize};
 use sha2::{Digest, Sha256};
@@ -78,26 +79,17 @@ impl Derivation {
 
     /// Returns the drv path of a Derivation struct.
     ///
-    /// The drv path is calculated like this:
-    ///   - Write the fingerprint of the Derivation to the sha256 hash function.
-    ///     This is: `text:`,
-    ///     all d.InputDerivations and d.InputSources (sorted, separated by a `:`),
-    ///     a `:`,
-    ///     the nix string representation of the sha256 sum of the ATerm representation
-    ///     a `:`,
-    ///     the storeDir, followed by a `:`,
-    ///     the name of a derivation,
-    ///     a `.drv`.
-    ///   - Write the .drv A-Term contents to a hash function
-    ///   - Take the digest, run hash.CompressHash(digest, 20) on it.
-    ///   - Encode it with nixbase32
-    ///   - Use it (and the name) to construct a [StorePath].
+    /// The drv path is calculated by calculating the [text_hash_string], using
+    /// the `name` with a `.drv` suffix as name, all d.InputDerivations and d.InputSources as references,
+    /// and the ATerm representation of the Derivation as contents.
+    /// The text_hash_string is then passed to the build_store_path function.
     pub fn calculate_derivation_path(&self, name: &str) -> Result<StorePath, DerivationError> {
-        let mut s = String::from("text:");
+        // append .drv to the name
+        let name_with_suffix = &format!("{}.drv", name);
 
         // collect the list of paths from input_sources and input_derivations
-        // into a (sorted, guaranteed by BTreeSet) list, and join them by :
-        let concat_inputs: BTreeSet<String> = {
+        // into a (sorted, guaranteed by BTreeSet) list of references
+        let references: BTreeSet<String> = {
             let mut inputs = self.input_sources.clone();
             let input_derivation_keys: Vec<String> =
                 self.input_derivations.keys().cloned().collect();
@@ -105,25 +97,9 @@ impl Derivation {
             inputs
         };
 
-        for input in concat_inputs {
-            s.push_str(&input);
-            s.push(':');
-        }
-
-        // calculate the sha256 hash of the ATerm representation, and represent
-        // it as a hex-encoded string (prefixed with sha256:).
-        let aterm_digest = Sha256::new_with_prefix(self.to_aterm_string())
-            .finalize()
-            .to_vec();
-
-        s.push_str(&format!(
-            "{}:{}:{}.drv",
-            NixHash::new(HashAlgo::Sha256, aterm_digest).to_nix_hash_string(),
-            store_path::STORE_DIR,
-            name,
-        ));
+        let text_hash_str = &text_hash_string(name_with_suffix, self.to_aterm_string(), references);
 
-        utils::build_store_path(true, &s, name)
+        utils::build_store_path(true, text_hash_str, name)
     }
 
     /// Returns the FOD digest, if the derivation is fixed-output, or None if
diff --git a/tvix/nix-compat/src/derivation/utils.rs b/tvix/nix-compat/src/derivation/utils.rs
index cbb5d02a21..68ab3e23c9 100644
--- a/tvix/nix-compat/src/derivation/utils.rs
+++ b/tvix/nix-compat/src/derivation/utils.rs
@@ -1,7 +1,7 @@
 use crate::derivation::DerivationError;
 use crate::nixbase32;
-use crate::nixhash::NixHash;
-use crate::store_path::{self, StorePath};
+use crate::store_path::StorePath;
+use crate::texthash::text_hash_string;
 use sha2::{Digest, Sha256};
 
 /// compress_hash takes an arbitrarily long sequence of bytes (usually
@@ -55,27 +55,6 @@ pub fn path_with_references<S: AsRef<str>, I: IntoIterator<Item = S>, C: AsRef<[
     content: C,
     references: I,
 ) -> Result<StorePath, DerivationError> {
-    let mut s = String::from("text");
-
-    for reference in references {
-        s.push(':');
-        s.push_str(reference.as_ref());
-    }
-
-    let content_digest = {
-        let mut hasher = Sha256::new();
-        hasher.update(content);
-        hasher.finalize()
-    };
-
-    let h = NixHash::new(crate::nixhash::HashAlgo::Sha256, content_digest.to_vec());
-
-    s.push_str(&format!(
-        ":{}:{}:{}",
-        h.to_nix_hash_string(),
-        store_path::STORE_DIR,
-        name
-    ));
-
-    build_store_path(false, &s, name)
+    let text_hash_str = text_hash_string(name, content, references);
+    build_store_path(false, &text_hash_str, name)
 }
diff --git a/tvix/nix-compat/src/lib.rs b/tvix/nix-compat/src/lib.rs
index 52b3ea65b9..ec9ddd1ffa 100644
--- a/tvix/nix-compat/src/lib.rs
+++ b/tvix/nix-compat/src/lib.rs
@@ -7,6 +7,7 @@ pub mod nixhash;
 mod nixhash_algos;
 mod nixhash_with_mode;
 pub mod store_path;
+mod texthash;
 
 /// Nix placeholders (i.e. values returned by `builtins.placeholder`)
 /// are used to populate outputs with paths that must be
diff --git a/tvix/nix-compat/src/texthash.rs b/tvix/nix-compat/src/texthash.rs
new file mode 100644
index 0000000000..959a4dc3dc
--- /dev/null
+++ b/tvix/nix-compat/src/texthash.rs
@@ -0,0 +1,42 @@
+use sha2::{Digest, Sha256};
+
+use crate::{nixhash::NixHash, store_path};
+
+/// This contains the Nix logic to create "text hash strings", which are used
+/// in `builtins.toFile`, as well as in Derivation Path calculation.
+///
+/// A text hash is calculated by concatenating the following fields, separated by a `:`:
+///
+///  - text
+///  - references, individually joined by `:`
+///  - 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,
+    references: I,
+) -> String {
+    let mut s = String::from("text:");
+
+    for reference in references {
+        s.push_str(reference.as_ref());
+        s.push(':');
+    }
+
+    // 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!(":{}:{}", store_path::STORE_DIR, name));
+
+    s
+}