about summary refs log tree commit diff
path: root/tvix
diff options
context:
space:
mode:
authorFlorian Klink <flokli@flokli.de>2023-12-09T10·53+0200
committerclbot <clbot@tvl.fyi>2023-12-11T22·35+0000
commit8486f87e3c180732d04e58154762c016c054d776 (patch)
treed555a1cd622bb4e9836314dd2bcb777f9b20ef82 /tvix
parent92bd69aef23126412a15649e441339c96e001212 (diff)
feat(tvix/build): add derivation_to_build_request r/7154
This function converts from a nix_compat::derivation::Derivation to
a BuildRequest.

In addition to the Derivation itself, it needs two lookup functions to
map input paths to their castore nodes.

Change-Id: I0332982f0bc7933a5fda137fe39d5a850639d929
Reviewed-on: https://cl.tvl.fyi/c/depot/+/10236
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Diffstat (limited to 'tvix')
-rw-r--r--tvix/Cargo.lock3
-rw-r--r--tvix/Cargo.nix13
-rw-r--r--tvix/build/Cargo.toml1
-rw-r--r--tvix/glue/Cargo.toml2
-rw-r--r--tvix/glue/src/lib.rs1
-rw-r--r--tvix/glue/src/tests/ch49594n9avinrf8ip0aslidkc4lxkqv-foo.drv1
-rw-r--r--tvix/glue/src/tvix_build.rs251
7 files changed, 272 insertions, 0 deletions
diff --git a/tvix/Cargo.lock b/tvix/Cargo.lock
index 518845af7b51..5d15a9436d78 100644
--- a/tvix/Cargo.lock
+++ b/tvix/Cargo.lock
@@ -3054,6 +3054,7 @@ checksum = "3528ecfd12c466c6f163363caf2d02a71161dd5e1cc6ae7b34207ea2d42d81ed"
 name = "tvix-build"
 version = "0.1.0"
 dependencies = [
+ "bytes",
  "prost",
  "prost-build",
  "tonic",
@@ -3161,11 +3162,13 @@ name = "tvix-glue"
 version = "0.1.0"
 dependencies = [
  "bytes",
+ "lazy_static",
  "nix-compat",
  "test-case",
  "thiserror",
  "tokio",
  "tracing",
+ "tvix-build",
  "tvix-castore",
  "tvix-eval",
  "tvix-store",
diff --git a/tvix/Cargo.nix b/tvix/Cargo.nix
index 6151fd76d378..826ebf320872 100644
--- a/tvix/Cargo.nix
+++ b/tvix/Cargo.nix
@@ -9370,6 +9370,10 @@ rec {
           else ./build;
         dependencies = [
           {
+            name = "bytes";
+            packageId = "bytes";
+          }
+          {
             name = "prost";
             packageId = "prost";
           }
@@ -9819,6 +9823,11 @@ rec {
             packageId = "tracing";
           }
           {
+            name = "tvix-build";
+            packageId = "tvix-build";
+            usesDefaultFeatures = false;
+          }
+          {
             name = "tvix-castore";
             packageId = "tvix-castore";
           }
@@ -9838,6 +9847,10 @@ rec {
         ];
         devDependencies = [
           {
+            name = "lazy_static";
+            packageId = "lazy_static";
+          }
+          {
             name = "test-case";
             packageId = "test-case";
           }
diff --git a/tvix/build/Cargo.toml b/tvix/build/Cargo.toml
index 6f9d8a34f276..99802fcc9901 100644
--- a/tvix/build/Cargo.toml
+++ b/tvix/build/Cargo.toml
@@ -4,6 +4,7 @@ version = "0.1.0"
 edition = "2021"
 
 [dependencies]
+bytes = "1.4.0"
 prost = "0.12.1"
 tonic = "0.10.2"
 tvix-castore = { path = "../castore" }
diff --git a/tvix/glue/Cargo.toml b/tvix/glue/Cargo.toml
index 4ebfda870399..4469c3bab3a0 100644
--- a/tvix/glue/Cargo.toml
+++ b/tvix/glue/Cargo.toml
@@ -5,6 +5,7 @@ edition = "2021"
 
 [dependencies]
 nix-compat = { path = "../nix-compat" }
+tvix-build = { path = "../build", default-features = false, features = []}
 tvix-eval = { path = "../eval" }
 tvix-castore = { path = "../castore" }
 tvix-store = { path = "../store", default-features = false, features = []}
@@ -17,4 +18,5 @@ thiserror = "1.0.38"
 git = "https://github.com/tvlfyi/wu-manber.git"
 
 [dev-dependencies]
+lazy_static = "1.4.0"
 test-case = "2.2.2"
diff --git a/tvix/glue/src/lib.rs b/tvix/glue/src/lib.rs
index 6b3de0e23555..805f16d04b1e 100644
--- a/tvix/glue/src/lib.rs
+++ b/tvix/glue/src/lib.rs
@@ -1,6 +1,7 @@
 pub mod builtins;
 pub mod known_paths;
 pub mod refscan;
+pub mod tvix_build;
 pub mod tvix_io;
 pub mod tvix_store_io;
 
diff --git a/tvix/glue/src/tests/ch49594n9avinrf8ip0aslidkc4lxkqv-foo.drv b/tvix/glue/src/tests/ch49594n9avinrf8ip0aslidkc4lxkqv-foo.drv
new file mode 100644
index 000000000000..1699c2a75e48
--- /dev/null
+++ b/tvix/glue/src/tests/ch49594n9avinrf8ip0aslidkc4lxkqv-foo.drv
@@ -0,0 +1 @@
+Derive([("out","/nix/store/fhaj6gmwns62s6ypkcldbaj2ybvkhx3p-foo","","")],[("/nix/store/ss2p4wmxijn652haqyd7dckxwl4c7hxx-bar.drv",["out"])],[],":",":",[],[("bar","/nix/store/mp57d33657rf34lzvlbpfa1gjfv5gmpg-bar"),("builder",":"),("name","foo"),("out","/nix/store/fhaj6gmwns62s6ypkcldbaj2ybvkhx3p-foo"),("system",":")])
\ No newline at end of file
diff --git a/tvix/glue/src/tvix_build.rs b/tvix/glue/src/tvix_build.rs
new file mode 100644
index 000000000000..e939abf3cc79
--- /dev/null
+++ b/tvix/glue/src/tvix_build.rs
@@ -0,0 +1,251 @@
+//! This module contains glue code translating from
+//! [nix_compat::derivation::Derivation] to [tvix_build::proto::BuildRequest].
+
+use std::collections::BTreeMap;
+
+use bytes::Bytes;
+use nix_compat::{derivation::Derivation, store_path::StorePathRef};
+use tvix_build::proto::{
+    build_request::{BuildConstraints, EnvVar},
+    BuildRequest,
+};
+use tvix_castore::proto::{NamedNode, Node};
+
+/// These are the environment variables that Nix sets in its sandbox for every
+/// build.
+const NIX_ENVIRONMENT_VARS: [(&str, &str); 12] = [
+    ("HOME", "/homeless-shelter"),
+    ("NIX_BUILD_CORES", "0"), // TODO: make this configurable?
+    ("NIX_BUILD_TOP", "/"),
+    ("NIX_LOG_FD", "2"),
+    ("NIX_STORE", "/nix/store"),
+    ("PATH", "/path-not-set"),
+    ("PWD", "/build"),
+    ("TEMP", "/build"),
+    ("TEMPDIR", "/build"),
+    ("TERM", "xterm-256color"),
+    ("TMP", "/build"),
+    ("TMPDIR", "/build"),
+];
+
+/// Takes a [Derivation] and turns it into a [BuildRequest].
+/// It assumes the Derivation has been validated.
+/// It needs two lookup functions:
+/// - one translating input sources to a castore node
+///   (`fn_input_sources_to_node`)
+/// - one translating input derivations and (a subset of their) output names to
+///   a castore node (`fn_input_drvs_to_node`).
+#[allow(dead_code)]
+fn derivation_to_build_request<FIS, FID>(
+    derivation: &Derivation,
+    fn_input_sources_to_node: FIS,
+    fn_input_drvs_to_node: FID,
+) -> BuildRequest
+where
+    FIS: Fn(StorePathRef) -> Node,
+    FID: Fn(StorePathRef, &[&str]) -> Node,
+{
+    debug_assert!(derivation.validate(true).is_ok(), "drv must validate");
+
+    // produce command_args, which is builder and arguments in a Vec.
+    let mut command_args: Vec<String> = Vec::with_capacity(derivation.arguments.len() + 1);
+    command_args.push(derivation.builder.clone());
+    command_args.extend_from_slice(&derivation.arguments);
+
+    // produce output_paths, which is the basename of each output (sorted)
+    // since Derivation is validated, we know output paths can be parsed.
+    // TODO: b/264 will remove the need to parse them here
+    let mut outputs: Vec<String> = derivation
+        .outputs
+        .values()
+        .map(|output| {
+            let output_storepath = StorePathRef::from_absolute_path(output.path.as_bytes())
+                .expect("invalid output storepath");
+
+            output_storepath.to_string()
+        })
+        .collect();
+
+    // Sort the outputs. We can use sort_unstable, as these are unique strings.
+    outputs.sort_unstable();
+
+    // Produce environment_vars. We use a BTreeMap while producing them, so the
+    // resulting Vec is sorted by key.
+    let mut environment_vars: BTreeMap<String, Bytes> = BTreeMap::new();
+
+    // Start with some the ones that nix magically sets:
+    environment_vars.extend(
+        NIX_ENVIRONMENT_VARS
+            .iter()
+            .map(|(k, v)| (k.to_string(), Bytes::from_static(v.as_bytes()))),
+    );
+
+    // extend / overwrite with the keys set in the derivation environment itself.
+    // TODO: check if this order is correct, and environment vars set in the
+    // *Derivation actually* have priority.
+    environment_vars.extend(
+        derivation
+            .environment
+            .iter()
+            .map(|(k, v)| (k.clone(), Bytes::from(v.to_vec()))),
+    );
+
+    // Turn this into a sorted-by-key Vec<EnvVar>.
+    let environment_vars = Vec::from_iter(
+        environment_vars
+            .into_iter()
+            .map(|(k, v)| EnvVar { key: k, value: v }),
+    );
+
+    // Produce inputs. As we refer to the contents here, not just plain store path strings,
+    // we need to perform lookups.
+    // FUTUREWORK: should we also model input_derivations and input_sources with StorePath?
+    let mut inputs: Vec<Node> = Vec::new();
+
+    // since Derivation is validated, we know input sources can be parsed.
+    for input_source in derivation.input_sources.iter() {
+        let sp = StorePathRef::from_absolute_path(input_source.as_bytes())
+            .expect("invalid input source path");
+        let node = fn_input_sources_to_node(sp);
+        inputs.push(node);
+    }
+
+    // since Derivation is validated, we know input derivations can be parsed.
+    for (input_derivation, output_names) in derivation.input_derivations.iter() {
+        let sp = StorePathRef::from_absolute_path(input_derivation.as_bytes())
+            .expect("invalid input derivation path");
+        let output_names: Vec<&str> = output_names.iter().map(|e| e.as_str()).collect();
+        let node = fn_input_drvs_to_node(sp, output_names.as_slice());
+        inputs.push(node);
+    }
+
+    // validate all nodes are actually Some.
+    debug_assert!(
+        inputs.iter().all(|input| input.node.is_some()),
+        "input nodes must be some"
+    );
+
+    // sort inputs by their name
+    inputs.sort_by(|a, b| {
+        a.node
+            .as_ref()
+            .unwrap()
+            .get_name()
+            .cmp(b.node.as_ref().unwrap().get_name())
+    });
+
+    // Produce constraints. We currently only put platform in here.
+    // Maybe more things need to be added here in the future.
+    let constraints = Some(BuildConstraints {
+        system: derivation.system.clone(),
+        min_memory: 0,
+        available_ro_paths: vec![],
+    });
+
+    BuildRequest {
+        command_args,
+        outputs,
+        environment_vars,
+        inputs,
+        constraints,
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use bytes::Bytes;
+    use nix_compat::derivation::Derivation;
+    use tvix_build::proto::{
+        build_request::{BuildConstraints, EnvVar},
+        BuildRequest,
+    };
+    use tvix_castore::{
+        fixtures::DUMMY_DIGEST,
+        proto::{DirectoryNode, Node},
+    };
+
+    use crate::tvix_build::NIX_ENVIRONMENT_VARS;
+
+    use super::derivation_to_build_request;
+    use lazy_static::lazy_static;
+
+    lazy_static! {
+        static ref INPUT_NODE_FOO: Node = Node {
+            node: Some(tvix_castore::proto::node::Node::Directory(DirectoryNode {
+                name: Bytes::from("mp57d33657rf34lzvlbpfa1gjfv5gmpg-bar"),
+                digest: DUMMY_DIGEST.clone().into(),
+                size: 42,
+            })),
+        };
+    }
+
+    #[test]
+    fn test_derivation_to_build_request() {
+        let aterm_bytes = include_bytes!("tests/ch49594n9avinrf8ip0aslidkc4lxkqv-foo.drv");
+
+        let derivation = Derivation::from_aterm_bytes(aterm_bytes).expect("must parse");
+
+        let build_request = derivation_to_build_request(
+            &derivation,
+            |_| unreachable!(),
+            |input_drv, output_names| {
+                // expected to be called with ss2p4wmxijn652haqyd7dckxwl4c7hxx-bar.drv only
+                if input_drv.to_string() != "ss2p4wmxijn652haqyd7dckxwl4c7hxx-bar.drv" {
+                    panic!("called with unexpected input_drv: {}", input_drv);
+                }
+                // expect to be called with ["out"]
+                if output_names != ["out"] {
+                    panic!("called with unexpected output_names: {:?}", output_names);
+                }
+
+                // all good, reply with INPUT_NODE_FOO
+                INPUT_NODE_FOO.clone()
+            },
+        );
+
+        let mut expected_environment_vars = vec![
+            EnvVar {
+                key: "bar".to_string(),
+                value: Bytes::from("/nix/store/mp57d33657rf34lzvlbpfa1gjfv5gmpg-bar"),
+            },
+            EnvVar {
+                key: "builder".to_string(),
+                value: Bytes::from(":"),
+            },
+            EnvVar {
+                key: "name".to_string(),
+                value: Bytes::from("foo"),
+            },
+            EnvVar {
+                key: "out".to_string(),
+                value: Bytes::from("/nix/store/fhaj6gmwns62s6ypkcldbaj2ybvkhx3p-foo"),
+            },
+            EnvVar {
+                key: "system".to_string(),
+                value: Bytes::from(":"),
+            },
+        ];
+
+        expected_environment_vars.extend(NIX_ENVIRONMENT_VARS.iter().map(|(k, v)| EnvVar {
+            key: k.to_string(),
+            value: Bytes::from_static(v.as_bytes()),
+        }));
+
+        expected_environment_vars.sort_unstable_by_key(|e| e.key.to_owned());
+
+        assert_eq!(
+            BuildRequest {
+                command_args: vec![":".to_string()],
+                outputs: vec!["fhaj6gmwns62s6ypkcldbaj2ybvkhx3p-foo".to_string()],
+                environment_vars: expected_environment_vars,
+                inputs: vec![INPUT_NODE_FOO.clone()],
+                constraints: Some(BuildConstraints {
+                    system: derivation.system.clone(),
+                    min_memory: 0,
+                    available_ro_paths: vec![],
+                }),
+            },
+            build_request
+        );
+    }
+}