From 83a6619c44c7891aa8e291d4620dd9407c9b72a5 Mon Sep 17 00:00:00 2001
From: Anand Krishnamoorthi <anakrish@microsoft.com>
Date: Thu, 9 May 2024 20:44:48 -0700
Subject: [PATCH] no_std support
- Disable default features in dependencies
- Use anyhow::Error::msg to map errors. Note: anyhow will itself be removed later.
- lazy_static/spin_no_std used in no_std environments
- ensure_no_std binary is built to target thumbv7m-none-eabi to ensure that
there are no std dependencies. thumbv7m-none-eabi target has no std support.
- The opa-no-std feature enables only those Regorus features that work with no_std.
- Enable tests with no_std
- Update sizes of regorus binary in README.md
- Ensure that regorus example can be built with only std
- Ensure that regorus example can be built with no_std
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
---
.github/workflows/pr.yml | 8 ++
Cargo.toml | 71 +++++++++++-----
README.md | 8 +-
examples/regorus.rs | 54 +++++++++++--
scripts/pre-push | 18 ++++-
src/builtins/debugging.rs | 53 ------------
src/builtins/encoding.rs | 34 +++++---
src/builtins/mod.rs | 4 -
src/builtins/numbers.rs | 8 +-
src/builtins/opa.rs | 1 +
src/builtins/semver.rs | 4 +-
src/builtins/test.rs | 4 +-
src/builtins/time.rs | 2 +-
src/builtins/time/compat.rs | 3 -
src/interpreter.rs | 138 ++++++++++++++++++++++++--------
src/lib.rs | 19 +++--
src/parser.rs | 21 ++---
src/scheduler.rs | 5 +-
src/tests/interpreter/mod.rs | 29 +++++++
src/value.rs | 1 +
tests/arc.rs | 3 +
tests/ensure_no_std/Cargo.toml | 10 +++
tests/ensure_no_std/src/main.rs | 16 ++++
tests/lexer/mod.rs | 1 +
tests/mod.rs | 3 -
25 files changed, 344 insertions(+), 174 deletions(-)
delete mode 100644 src/builtins/debugging.rs
create mode 100644 tests/ensure_no_std/Cargo.toml
create mode 100644 tests/ensure_no_std/src/main.rs
diff --git a/.github/workflows/pr.yml b/.github/workflows/pr.yml
index ec4b4b9b..1dc0a5a7 100644
--- a/.github/workflows/pr.yml
+++ b/.github/workflows/pr.yml
@@ -18,6 +18,8 @@ jobs:
- uses: actions/checkout@v3
- name: Add musl target
run: rustup target add x86_64-unknown-linux-musl
+ - name: Add no_std target
+ run: rustup target add thumbv7m-none-eabi
- name: Install musl-gcc
run: sudo apt update && sudo apt install -y musl-tools
- name: Format Check
@@ -26,6 +28,12 @@ jobs:
run: cargo build -r --all-features --verbose
- name: Build
run: cargo build -r --verbose
+ - name: Build no_std
+ run: cd tests/ensure_no_std && cargo build -r --target thumbv7m-none-eabi
+ - name: Test no_std
+ run: cargo test -r --no-default-features
+ - name: Build only std
+ run: cargo build -r --example regorus --no-default-features --features "std"
- name: Doc Tests
run: cargo test -r --doc
- name: Run tests
diff --git a/Cargo.toml b/Cargo.toml
index 7d753bbf..6ee7d065 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -6,6 +6,7 @@ members = [
"bindings/wasm",
"bindings/java",
"bindings/ruby/ext/regorusrb",
+ "tests/ensure_no_std",
]
[package]
@@ -15,7 +16,7 @@ version = "0.1.5"
edition = "2021"
license-file = "LICENSE"
repository = "https://github.com/microsoft/regorus"
-keywords = ["interpreter", "opa", "policy-as-code", "rego"]
+keywords = ["interpreter", "no_std", "opa", "policy-as-code", "rego"]
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
@@ -33,14 +34,15 @@ crypto = ["dep:constant_time_eq", "dep:hmac", "dep:hex", "dep:md-5", "dep:sha1",
deprecated = []
hex = ["dep:data-encoding"]
http = []
-jwt = ["dep:jsonwebtoken", "dep:data-encoding"]
glob = ["dep:wax"]
graph = []
jsonschema = ["dep:jsonschema"]
+jwt = ["dep:jsonwebtoken", "dep:data-encoding", "dep:itertools"]
+no_std = ["lazy_static/spin_no_std"]
opa-runtime = []
regex = ["dep:regex"]
semver = ["dep:semver"]
-std = ["serde_json/std"]
+std = ["rand/std", "rand/std_rng", "serde_json/std"]
time = ["dep:chrono", "dep:chrono-tz"]
uuid = ["dep:uuid"]
urlquery = ["dep:url"]
@@ -67,41 +69,62 @@ full-opa = [
"yaml"
]
+# Features that can be used in no_std environments.
+# Note that: the spin_no_std feature in lazy_static must be specified.
+opa-no-std = [
+ "arc",
+ "base64",
+ "base64url",
+ "coverage",
+ "crypto",
+ "deprecated",
+ "graph",
+ "hex",
+ "no_std",
+ "opa-runtime",
+ "regex",
+ "semver",
+ # Configure lazy_static to use spinlocks.
+ "lazy_static/spin_no_std"
+]
+
# This feature enables some testing utils for OPA tests.
opa-testutil = []
+rand = ["dep:rand"]
[dependencies]
-anyhow = { version = "1.0.45", default-features=false }
+anyhow = { version = "1.0.45", default-features = false }
serde = {version = "1.0.150", default-features = false, features = ["derive", "rc"] }
-serde_json = { version = "1.0.89", default-features=false, features = ["alloc"] }
-serde_yaml = {version = "0.9.16", optional = true }
-lazy_static = "1.4.0"
-rand = "0.8.5"
-num = "0.4.1"
+serde_json = { version = "1.0.89", default-features = false, features = ["alloc"] }
+lazy_static = { version = "1.4.0", default-features = false }
# Crypto
-constant_time_eq = {version = "0.3.0", optional = true}
-hmac = {version = "0.12.1", optional = true}
-sha2 = {version= "0.10.8", optional = true}
-hex = {version = "0.4.3", optional = true}
-sha1 = {version = "0.10.6", optional = true}
-md-5 = {version = "0.10.6", optional = true}
-
-data-encoding = { version = "2.4.0", optional = true }
+constant_time_eq = {version = "0.3.0", optional = true, default-features = false }
+hmac = {version = "0.12.1", optional = true, default-features = false}
+sha2 = {version= "0.10.8", optional = true, default-features = false }
+hex = {version = "0.4.3", optional = true, default-features = false, features = ["alloc"] }
+sha1 = {version = "0.10.6", optional = true, default-features = false }
+md-5 = {version = "0.10.6", optional = true, default-features = false }
+
+data-encoding = { version = "2.4.0", optional = true, default-features=false, features = ["alloc"] }
scientific = { version = "0.5.2" }
-regex = {version = "1.10.2", optional = true}
-semver = {version = "1.0.20", optional = true}
+regex = {version = "1.10.2", optional = true, default-features = false }
+semver = {version = "1.0.20", optional = true, default-features = false }
wax = { version = "0.6.0", features = [], default-features = false, optional = true }
url = { version = "2.5.0", optional = true }
-uuid = { version = "1.6.1", features = ["v4", "fast-rng"], optional = true }
+uuid = { version = "1.6.1", default-features = false, features = ["v4", "fast-rng"], optional = true }
jsonschema = { version = "0.17.1", default-features = false, optional = true }
chrono = { version = "0.4.31", optional = true }
chrono-tz = { version = "0.8.5", optional = true }
jsonwebtoken = { version = "9.2.0", optional = true }
-itertools = "0.12.1"
+itertools = { version = "0.12.1", default-features = false, optional = true }
+
+serde_yaml = {version = "0.9.16", default-features = false, optional = true }
+rand = { version = "0.8.5", default-features = false, optional = true }
[dev-dependencies]
+anyhow = "1.0.45"
cfg-if = "1.0.0"
clap = { version = "4.4.7", features = ["derive"] }
prettydiff = { version = "0.6.4", default-features = false }
@@ -131,6 +154,12 @@ name="kata"
harness=false
test=false
+[[example]]
+name="regorus"
+harness=false
+test=false
+doctest=false
+
[package.metadata.docs.rs]
# To build locally:
# RUSTDOCFLAGS="--cfg docsrs" cargo +nightly doc --all-features --no-deps
diff --git a/README.md b/README.md
index 66752b7f..be9adb51 100644
--- a/README.md
+++ b/README.md
@@ -88,14 +88,14 @@ features. By default all features are enabled.
The default build of regorus example program is 6.4M:
```bash
$ cargo build -r --example regorus; strip target/release/examples/regorus; ls -lh target/release/examples/regorus
--rwxr-xr-x 1 anand staff 6.4M Jan 19 11:23 target/release/examples/regorus*
+-rwxr-xr-x 1 anand staff 6.3M May 11 22:03 target/release/examples/regorus*
```
-When all features except for `yaml` are disabled, the binary size drops down to 2.9M.
+When all default features are disabled, the binary size drops down to 1.9M.
```bash
-$ cargo build -r --example regorus --features "yaml" --no-default-features; strip target/release/examples/regorus; ls -lh target/release/examples/regorus
--rwxr-xr-x 1 anand staff 2.9M Jan 19 11:26 target/release/examples/regorus*
+$ cargo build -r --example regorus --no-default-features; strip target/release/examples/regorus; ls -lh target/release/examples/regorus
+-rwxr-xr-x 1 anand staff 1.9M May 11 22:04 target/release/examples/regorus*
```
Regorus passes the [OPA v0.64.0 test-suite](https://www.openpolicyagent.org/docs/latest/ir/#test-suite) barring a few
diff --git a/examples/regorus.rs b/examples/regorus.rs
index 22118da3..4d4b437e 100644
--- a/examples/regorus.rs
+++ b/examples/regorus.rs
@@ -1,7 +1,37 @@
// Copyright (c) Microsoft Corporation.
// Licensed under the MIT License.
-use anyhow::{bail, Result};
+use anyhow::{anyhow, bail, Result};
+
+#[allow(dead_code)]
+fn read_file(path: &String) -> Result<String> {
+ std::fs::read_to_string(path).map_err(|_| anyhow!("could not read {path}"))
+}
+
+#[allow(unused_variables)]
+fn read_value_from_yaml_file(path: &String) -> Result<regorus::Value> {
+ #[cfg(feature = "yaml")]
+ return regorus::Value::from_yaml_file(path);
+
+ #[cfg(not(feature = "yaml"))]
+ bail!("regorus has not been built with yaml support");
+}
+
+fn read_value_from_json_file(path: &String) -> Result<regorus::Value> {
+ #[cfg(feature = "std")]
+ return regorus::Value::from_json_file(path);
+
+ #[cfg(not(feature = "std"))]
+ regorus::Value::from_json_str(&read_file(path)?)
+}
+
+fn add_policy_from_file(engine: &mut regorus::Engine, path: String) -> Result<()> {
+ #[cfg(feature = "std")]
+ return engine.add_policy_from_file(path);
+
+ #[cfg(not(feature = "std"))]
+ engine.add_policy(path.clone(), read_file(&path)?)
+}
fn rego_eval(
bundles: &[String],
@@ -35,7 +65,7 @@ fn rego_eval(
_ => continue,
}
- engine.add_policy_from_file(entry.path())?;
+ add_policy_from_file(&mut engine, entry.path().display().to_string())?;
}
}
@@ -43,15 +73,15 @@ fn rego_eval(
for file in files.iter() {
if file.ends_with(".rego") {
// Read policy file.
- engine.add_policy_from_file(file)?;
+ add_policy_from_file(&mut engine, file.clone())?;
} else {
// Read data file.
let data = if file.ends_with(".json") {
- regorus::Value::from_json_file(file)?
+ read_value_from_json_file(file)?
} else if file.ends_with(".yaml") {
- regorus::Value::from_yaml_file(file)?
+ read_value_from_yaml_file(file)?
} else {
- bail!("Unsupported data file `{file}`. Must be rego, json or yaml.")
+ bail!("Unsupported data file `{file}`. Must be rego, json or yaml.");
};
// Merge given data.
@@ -61,9 +91,9 @@ fn rego_eval(
if let Some(file) = input {
let input = if file.ends_with(".json") {
- regorus::Value::from_json_file(&file)?
+ read_value_from_json_file(&file)?
} else if file.ends_with(".yaml") {
- regorus::Value::from_yaml_file(&file)?
+ read_value_from_yaml_file(&file)?
} else {
bail!("Unsupported input file `{file}`. Must be json or yaml.")
};
@@ -95,8 +125,12 @@ fn rego_lex(file: String, verbose: bool) -> Result<()> {
use regorus::unstable::*;
// Create source.
+ #[cfg(feature = "std")]
let source = Source::from_file(file)?;
+ #[cfg(not(feature = "std"))]
+ let source = Source::from_contents(file.clone(), read_file(&file)?)?;
+
// Create lexer.
let mut lexer = Lexer::new(&source);
@@ -122,8 +156,12 @@ fn rego_parse(file: String) -> Result<()> {
use regorus::unstable::*;
// Create source.
+ #[cfg(feature = "std")]
let source = Source::from_file(file)?;
+ #[cfg(not(feature = "std"))]
+ let source = Source::from_contents(file.clone(), read_file(&file)?)?;
+
// Create a parser and parse the source.
let mut parser = Parser::new(&source)?;
let ast = parser.parse()?;
diff --git a/scripts/pre-push b/scripts/pre-push
index 7a2bec87..d57f1309 100755
--- a/scripts/pre-push
+++ b/scripts/pre-push
@@ -5,17 +5,27 @@
set -eo pipefail
if [ -f Cargo.toml ]; then
- # Run precommit checks
+ # Run precommit checks.
dir=$(dirname "${BASH_SOURCE[0]}")
"$dir/pre-commit"
- # Ensure that the public API works
+ # Ensure that the public API works.
cargo test -r --doc
- # Ensure that we can build with all features
+ # Ensure that no_std build succeeds.
+ # Build for a target that has no std available.
+ if command -v rustup > /dev/null; then
+ rustup target add thumbv7m-none-eabi
+ (cd tests/ensure_no_std; cargo build -r --target thumbv7m-none-eabi)
+ fi
+
+ # Ensure that we can build with only std.
+ cargo build -r --example regorus --no-default-features --features std
+
+ # Ensure that we can build with all features.
cargo build -r --all-features
- # Ensure that all tests pass
+ # Ensure that all tests pass.
cargo test -r
cargo test -r --test aci
cargo test -r --test kata
diff --git a/src/builtins/debugging.rs b/src/builtins/debugging.rs
deleted file mode 100644
index 003e1383..00000000
--- a/src/builtins/debugging.rs
+++ /dev/null
@@ -1,53 +0,0 @@
-// Copyright (c) Microsoft Corporation.
-// Licensed under the MIT License.
-
-use crate::ast::{Expr, Ref};
-use crate::builtins;
-use crate::lexer::Span;
-use crate::value::Value;
-use crate::*;
-
-use anyhow::{bail, Result};
-
-// TODO: Should we avoid this limit?
-const MAX_ARGS: u8 = core::u8::MAX;
-
-pub fn register(m: &mut builtins::BuiltinsMap<&'static str, builtins::BuiltinFcn>) {
- m.insert("print", (print, MAX_ARGS));
-}
-
-pub fn print_to_string(
- span: &Span,
- _params: &[Ref<Expr>],
- args: &[Value],
- _strict: bool,
-) -> Result<String> {
- if args.len() > MAX_ARGS as usize {
- bail!(span.error("print supports up to 100 arguments"));
- }
-
- let mut msg = String::default();
- for a in args {
- match a {
- Value::Undefined => msg += " <undefined>",
- Value::String(s) => msg += &format!(" {s}"),
- _ => msg += &format!(" {a}"),
- };
- }
-
- Ok(msg)
-}
-
-// Symbol analyzer must ensure that vars used by print are defined before
-// the print statement. Scheduler must ensure the above constraint.
-// Additionally interpreter must allow undefined inputs to print.
-fn print(span: &Span, params: &[Ref<Expr>], args: &[Value], strict: bool) -> Result<Value> {
- let msg = print_to_string(span, params, args, strict)?;
-
- #[cfg(feature = "std")]
- if !msg.is_empty() {
- std::eprintln!("{}", &msg[1..]);
- }
-
- Ok(Value::Bool(true))
-}
diff --git a/src/builtins/encoding.rs b/src/builtins/encoding.rs
index 6c05a056..0c89ab8a 100644
--- a/src/builtins/encoding.rs
+++ b/src/builtins/encoding.rs
@@ -63,7 +63,13 @@ fn base64_decode(
ensure_args_count(span, name, params, args, 1)?;
let encoded_str = ensure_string(name, ¶ms[0], &args[0])?;
- let decoded_bytes = data_encoding::BASE64.decode(encoded_str.as_bytes())?;
+ let decoded_bytes = data_encoding::BASE64
+ .decode(encoded_str.as_bytes())
+ .map_err(|e| {
+ params[0]
+ .span()
+ .error(&format!("decode failed\nCaused by\n{e}"))
+ })?;
Ok(Value::String(
String::from_utf8_lossy(&decoded_bytes).into(),
))
@@ -173,7 +179,13 @@ fn hex_decode(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool)
ensure_args_count(span, name, params, args, 1)?;
let encoded_str = ensure_string(name, ¶ms[0], &args[0])?;
- let decoded_bytes = data_encoding::HEXLOWER_PERMISSIVE.decode(encoded_str.as_bytes())?;
+ let decoded_bytes = data_encoding::HEXLOWER_PERMISSIVE
+ .decode(encoded_str.as_bytes())
+ .map_err(|e| {
+ params[0]
+ .span()
+ .error(&format!("decode failure\nCaused by\n{e}"))
+ })?;
Ok(Value::String(
String::from_utf8_lossy(&decoded_bytes).into(),
))
@@ -361,11 +373,9 @@ fn json_is_valid(
fn json_marshal(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool) -> Result<Value> {
let name = "json.marshal";
ensure_args_count(span, name, params, args, 1)?;
- Ok(Value::String(
- serde_json::to_string(&args[0])
- .with_context(|| span.error("could not serialize to json"))?
- .into(),
- ))
+ Ok(Value::from(serde_json::to_string(&args[0]).map_err(
+ |e| span.error(&format!("could not serialize to json\nCaused by\n{e}")),
+ )?))
}
fn json_marshal_with_options(
@@ -406,15 +416,13 @@ fn json_marshal_with_options(
}
if !pretty || options.is_empty() {
- return Ok(Value::String(
- serde_json::to_string(&args[0])
- .with_context(|| span.error("could not serialize to json"))?
- .into(),
- ));
+ return Ok(Value::from(serde_json::to_string(&args[0]).map_err(
+ |e| span.error(&format!("could not serialize to json\nCaused by\n{e}")),
+ )?));
}
let lines: Vec<String> = serde_json::to_string_pretty(&args[0])
- .with_context(|| span.error("could not serialize to json"))?
+ .map_err(|e| span.error(&format!("could not serialize to json\nCaused by\n{e}")))?
.split('\n')
.map(|line| {
let mut line = line.to_string();
diff --git a/src/builtins/mod.rs b/src/builtins/mod.rs
index 9abfd589..adf6c89c 100644
--- a/src/builtins/mod.rs
+++ b/src/builtins/mod.rs
@@ -9,7 +9,6 @@ mod conversions;
#[cfg(feature = "crypto")]
mod crypto;
-mod debugging;
#[cfg(feature = "deprecated")]
pub mod deprecated;
mod encoding;
@@ -54,8 +53,6 @@ use lazy_static::lazy_static;
pub type BuiltinFcn = (fn(&Span, &[Ref<Expr>], &[Value], bool) -> Result<Value>, u8);
-pub use debugging::print_to_string;
-
#[cfg(feature = "deprecated")]
pub use deprecated::DEPRECATED;
@@ -104,7 +101,6 @@ lazy_static! {
//rego::register(&mut m);
#[cfg(feature = "opa-runtime")]
opa::register(&mut m);
- debugging::register(&mut m);
tracing::register(&mut m);
units::register(&mut m);
diff --git a/src/builtins/numbers.rs b/src/builtins/numbers.rs
index a3bf55cc..87675386 100644
--- a/src/builtins/numbers.rs
+++ b/src/builtins/numbers.rs
@@ -3,13 +3,15 @@
use crate::ast::{ArithOp, Expr, Ref};
use crate::builtins;
-use crate::builtins::utils::{ensure_args_count, ensure_numeric, ensure_string};
+use crate::builtins::utils::{ensure_args_count, ensure_numeric};
use crate::lexer::Span;
use crate::number::Number;
use crate::value::Value;
use crate::*;
use anyhow::{bail, Result};
+
+#[cfg(feature = "std")]
use rand::{thread_rng, Rng};
pub fn register(m: &mut builtins::BuiltinsMap<&'static str, builtins::BuiltinFcn>) {
@@ -18,6 +20,7 @@ pub fn register(m: &mut builtins::BuiltinsMap<&'static str, builtins::BuiltinFcn
m.insert("floor", (floor, 1));
m.insert("numbers.range", (range, 2));
m.insert("numbers.range_step", (range_step, 3));
+ #[cfg(feature = "std")]
m.insert("rand.intn", (intn, 2));
m.insert("round", (round, 1));
}
@@ -155,10 +158,11 @@ fn round(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool) -> Re
))
}
+#[cfg(feature = "std")]
fn intn(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool) -> Result<Value> {
let fcn = "rand.intn";
ensure_args_count(span, fcn, params, args, 2)?;
- let _ = ensure_string(fcn, ¶ms[0], &args[0])?;
+ let _ = crate::builtins::utils::ensure_string(fcn, ¶ms[0], &args[0])?;
let n = ensure_numeric(fcn, ¶ms[0], &args[1])?;
Ok(match n.as_u64() {
diff --git a/src/builtins/opa.rs b/src/builtins/opa.rs
index fc388650..b427b779 100644
--- a/src/builtins/opa.rs
+++ b/src/builtins/opa.rs
@@ -38,6 +38,7 @@ fn opa_runtime(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool)
);
// Emitting environment variables could lead to confidential data being leaked.
+ #[cfg(feature = "std")]
if false {
obj.insert(
Value::String("env".into()),
diff --git a/src/builtins/semver.rs b/src/builtins/semver.rs
index 8154fd38..6f4c916d 100644
--- a/src/builtins/semver.rs
+++ b/src/builtins/semver.rs
@@ -24,8 +24,8 @@ fn compare(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool) ->
let v1 = ensure_string(name, ¶ms[0], &args[0])?;
let v2 = ensure_string(name, ¶ms[1], &args[1])?;
- let version1 = Version::parse(&v1)?;
- let version2 = Version::parse(&v2)?;
+ let version1 = Version::parse(&v1).map_err(|_| params[0].span().error("invalid semver"))?;
+ let version2 = Version::parse(&v2).map_err(|_| params[0].span().error("invalid semver"))?;
let result = match version1.cmp_precedence(&version2) {
Ordering::Less => -1,
Ordering::Equal => 0,
diff --git a/src/builtins/test.rs b/src/builtins/test.rs
index 0fe42d06..7a6d5264 100644
--- a/src/builtins/test.rs
+++ b/src/builtins/test.rs
@@ -7,6 +7,7 @@ use crate::builtins::time;
use crate::builtins::utils::{ensure_args_count, ensure_string};
use crate::lexer::Span;
use crate::value::Value;
+use crate::*;
use std::thread;
@@ -21,7 +22,8 @@ fn sleep(span: &Span, params: &[Ref<Expr>], args: &[Value], _strict: bool) -> Re
ensure_args_count(span, name, params, args, 1)?;
let val = ensure_string(name, ¶ms[0], &args[0])?;
- let dur = time::compat::parse_duration(val.as_ref())?;
+ let dur = time::compat::parse_duration(val.as_ref())
+ .map_err(|e| params[0].span().error(&format!("{e}")))?;
thread::sleep(dur.to_std()?);
diff --git a/src/builtins/time.rs b/src/builtins/time.rs
index 12cf4743..8bc2ffa6 100644
--- a/src/builtins/time.rs
+++ b/src/builtins/time.rs
@@ -147,7 +147,7 @@ fn parse_duration_ns(
ensure_args_count(span, name, params, args, 1)?;
let value = ensure_string(name, ¶ms[0], &args[0])?;
- let dur = compat::parse_duration(value.as_ref())?;
+ let dur = compat::parse_duration(value.as_ref()).map_err(anyhow::Error::msg)?;
safe_timestamp_nanos(span, strict, dur.num_nanoseconds())
}
diff --git a/src/builtins/time/compat.rs b/src/builtins/time/compat.rs
index e2e0b440..0d491d04 100644
--- a/src/builtins/time/compat.rs
+++ b/src/builtins/time/compat.rs
@@ -34,7 +34,6 @@
use crate::*;
use core::fmt;
use core::iter;
-use std::error::Error;
use chrono::TimeZone;
use chrono::{
@@ -72,8 +71,6 @@ impl fmt::Display for ParseDurationError {
}
}
-impl Error for ParseDurationError {}
-
// Parses a duration string in the form of `10h12m45s`.
//
// Adapted from Go's `time.ParseDuration`:
diff --git a/src/interpreter.rs b/src/interpreter.rs
index caa0a299..a6a4233c 100644
--- a/src/interpreter.rs
+++ b/src/interpreter.rs
@@ -61,6 +61,7 @@ pub struct Interpreter {
builtins_cache: BTreeMap<(&'static str, Vec<Value>), Value>,
no_rules_lookup: bool,
traces: Option<Vec<Rc<str>>>,
+ #[cfg(feature = "deprecated")]
allow_deprecated: bool,
strict_builtin_errors: bool,
imports: BTreeMap<String, Ref<Expr>>,
@@ -186,6 +187,7 @@ impl Interpreter {
builtins_cache: BTreeMap::new(),
no_rules_lookup: false,
traces: None,
+ #[cfg(feature = "deprecated")]
allow_deprecated: true,
strict_builtin_errors: true,
imports: BTreeMap::default(),
@@ -1208,7 +1210,7 @@ impl Interpreter {
let mut target = path.join(".");
let mut target_is_function = self.lookup_function_by_name(&target).is_some()
- || matches!(self.lookup_builtin(wm.refr.span(), &target), Ok(Some(_)));
+ || self.is_builtin(wm.refr.span(), &target);
if !target_is_function
&& !target.starts_with("data.")
@@ -1217,7 +1219,7 @@ impl Interpreter {
{
// target must be a function.
if self.lookup_function_by_name(&target).is_none()
- && !matches!(self.lookup_builtin(wm.refr.span(), &target), Ok(Some(_)))
+ && !self.is_builtin(wm.refr.span(), &target)
{
// Prefix target with current module path.
target = self.current_module_path.clone() + "." + ⌖
@@ -1244,10 +1246,7 @@ impl Interpreter {
// Lookup without current module path prefixed.
function_path = get_path_string(&wm.r#as, None)?;
if self.lookup_function_by_name(&function_path).is_none()
- && !matches!(
- self.lookup_builtin(wm.r#as.span(), &function_path),
- Ok(Some(_))
- )
+ && !self.is_builtin(wm.r#as.span(), &function_path)
{
// bail!(wm.r#as.span().error("could not evaluate expression"));
skip_exec = true;
@@ -1743,9 +1742,9 @@ impl Interpreter {
span.col,
format!(
"value for key `{}` generated multiple times: `{}` and `{}`",
- serde_json::to_string_pretty(&key)?,
- serde_json::to_string_pretty(&pv)?,
- serde_json::to_string_pretty(&value)?,
+ serde_json::to_string_pretty(&key).map_err(anyhow::Error::msg)?,
+ serde_json::to_string_pretty(&pv).map_err(anyhow::Error::msg)?,
+ serde_json::to_string_pretty(&value).map_err(anyhow::Error::msg)?,
)
.as_str(),
));
@@ -2121,27 +2120,11 @@ impl Interpreter {
name: &str,
builtin: builtins::BuiltinFcn,
params: &[ExprRef],
+ args: Vec<Value>,
) -> Result<Value> {
- let mut args = vec![];
- let is_print = name == "print"; // TODO: with modifier
- let allow_undefined = is_print;
- for p in params {
- match self.eval_expr(p)? {
- // If any argument is undefined, then the call is undefined.
- Value::Undefined if !allow_undefined => return Ok(Value::Undefined),
- p => args.push(p),
- }
- }
-
- if is_print && self.gather_prints {
- // Do not print to stderr. Instead, gather.
- let msg =
- builtins::print_to_string(span, params, &args[..], self.strict_builtin_errors)?;
-
- // Prefix location information.
- self.prints
- .push(format!("{}:{}: {msg}", span.source.file(), span.line));
- return Ok(Value::Bool(true));
+ // If any argument is undefined, then the call is undefined.
+ if args.iter().any(|a| a == &Value::Undefined) {
+ return Ok(Value::Undefined);
}
let cache = builtins::must_cache(name);
@@ -2173,6 +2156,7 @@ impl Interpreter {
Ok(v)
}
+ #[allow(unused_variables)]
fn lookup_builtin(&self, span: &Span, path: &str) -> Result<Option<&BuiltinFcn>> {
if let Some(builtin) = builtins::BUILTINS.get(path) {
return Ok(Some(builtin));
@@ -2187,12 +2171,92 @@ impl Interpreter {
return Ok(Some(builtin));
}
- // Mark as used when deprecated feature is not enabled.
- core::convert::identity((span, self.allow_deprecated));
-
Ok(None)
}
+ fn is_builtin(&self, span: &Span, path: &str) -> bool {
+ path == "print" || matches!(self.lookup_builtin(span, path), Ok(Some(_)))
+ }
+
+ fn to_printable(v: &Value, s: &mut String) {
+ match v {
+ Value::Array(array) => {
+ s.push('[');
+ for (idx, e) in array.iter().enumerate() {
+ if idx > 0 {
+ s.push_str(", ");
+ }
+ Self::to_printable(e, s);
+ }
+ s.push(']');
+ }
+ Value::Set(set) => {
+ s.push('{');
+ for (idx, e) in set.iter().enumerate() {
+ if idx > 0 {
+ s.push_str(", ");
+ }
+ Self::to_printable(e, s);
+ }
+ s.push('}');
+ }
+ Value::Object(map) => {
+ s.push('{');
+ for (idx, (k, v)) in map.iter().enumerate() {
+ if idx > 0 {
+ s.push_str(", ");
+ }
+ Self::to_printable(k, s);
+ s.push_str(": ");
+ Self::to_printable(v, s);
+ }
+ s.push('}');
+ }
+ v => s.push_str(&format!("{v}")),
+ }
+ }
+
+ fn eval_print(&mut self, span: &Span, params: &[ExprRef], args: Vec<Value>) -> Result<Value> {
+ const MAX_ARGS: u8 = 100;
+ if args.len() > MAX_ARGS as usize {
+ bail!(span.error(&format!("print supports upto {MAX_ARGS} arguments")));
+ }
+
+ // If not compiling for std target, return early if gathering is not
+ // requested.
+ #[cfg(not(feature = "std"))]
+ if !self.gather_prints {
+ return Ok(Value::Bool(true));
+ }
+
+ let mut msg = String::default();
+ for (i, p) in params.iter().enumerate() {
+ if i > 0 {
+ msg.push(' ');
+ }
+ match self.eval_expr(p)? {
+ Value::Undefined => msg.push_str("<undefined>"),
+ // Do not print quotes for string values.
+ Value::String(s) => msg.push_str(&format!("{s}")),
+ a => Self::to_printable(&a, &mut msg),
+ }
+ }
+
+ if self.gather_prints {
+ // Prefix location information.
+ self.prints
+ .push(format!("{}:{}: {msg}", span.source.file(), span.line));
+ }
+
+ // Print to stderr only if not gathering.
+ #[cfg(feature = "std")]
+ if !self.gather_prints {
+ std::eprintln!("{msg}");
+ }
+
+ Ok(Value::Bool(true))
+ }
+
fn eval_call_impl(
&mut self,
span: &Span,
@@ -2261,10 +2325,18 @@ impl Interpreter {
else if let Some(ext) = self.extensions.get_mut(&fcn_path) {
extension = Some(ext);
(&empty, None)
+ } else if fcn_path == "print" {
+ return self.eval_print(span, params, param_values);
}
// Look up builtin function.
else if let Some(builtin) = self.lookup_builtin(span, &fcn_path)? {
- let r = self.eval_builtin_call(span, &fcn_path.clone(), *builtin, params);
+ let r = self.eval_builtin_call(
+ span,
+ &fcn_path.clone(),
+ *builtin,
+ params,
+ param_values,
+ );
if let Some(with_functions) = with_functions_saved {
self.with_functions = with_functions;
}
diff --git a/src/lib.rs b/src/lib.rs
index f30546c1..a63befbe 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -404,30 +404,29 @@ pub mod coverage {
/// <img src="https://github.com/microsoft/regorus/blob/main/docs/coverage.png?raw=true">
pub fn to_colored_string(&self) -> anyhow::Result<String> {
- use std::io::Write;
- let mut s = Vec::new();
- writeln!(&mut s, "COVERAGE REPORT:")?;
+ let mut s = String::default();
+ s.push_str("COVERAGE REPORT:\n");
for file in self.files.iter() {
if file.not_covered.is_empty() {
- writeln!(&mut s, "{} has full coverage", file.path)?;
+ s.push_str(&format!("{} has full coverage\n", file.path));
continue;
}
- writeln!(&mut s, "{}:", file.path)?;
+ s.push_str(&format!("{}:", file.path));
for (line, code) in file.code.split('\n').enumerate() {
let line = line as u32 + 1;
if file.not_covered.contains(&line) {
- writeln!(&mut s, "\x1b[31m {line:4} {code}\x1b[0m")?;
+ s.push_str(&format!("\x1b[31m {line:4} {code}\x1b[0m\n"));
} else if file.covered.contains(&line) {
- writeln!(&mut s, "\x1b[32m {line:4} {code}\x1b[0m")?;
+ s.push_str(&format!("\x1b[32m {line:4} {code}\x1b[0m\n"));
} else {
- writeln!(&mut s, " {line:4} {code}")?;
+ s.push_str(&format!(" {line:4} {code}\n"));
}
}
}
- writeln!(&mut s)?;
- Ok(core::str::from_utf8(&s)?.to_string())
+ s.push('\n');
+ Ok(s)
}
}
}
diff --git a/src/parser.rs b/src/parser.rs
index 3f34eab0..45d45b8a 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -70,16 +70,19 @@ impl<'source> Parser<'source> {
}
pub fn warn_future_keyword(&self) {
- let kw = self.token_text();
- let msg = format!(
- "`{kw}` will be treated as identifier due to missing `import future.keywords.{kw}`"
- );
#[cfg(feature = "std")]
- std::println!(
- "{}",
- self.source
- .message(self.tok.1.line, self.tok.1.col, "warning", &msg)
- );
+ {
+ let kw = self.token_text();
+ let msg = format!(
+ "`{kw}` will be treated as identifier due to missing `import future.keywords.{kw}`"
+ );
+
+ std::println!(
+ "{}",
+ self.source
+ .message(self.tok.1.line, self.tok.1.col, "warning", &msg)
+ );
+ }
}
pub fn set_future_keyword(&mut self, kw: &str, span: &Span) -> Result<()> {
diff --git a/src/scheduler.rs b/src/scheduler.rs
index 98f96c52..3645996a 100644
--- a/src/scheduler.rs
+++ b/src/scheduler.rs
@@ -47,7 +47,6 @@ pub fn schedule<Str: Clone + cmp::Ord + fmt::Debug>(
empty: &Str,
) -> Result<SortResult> {
let num_statements = infos.len();
- let orig_infos: Vec<&StmtInfo<Str>> = infos.iter().collect();
// Mapping from each var to the list of statements that define it.
let mut defining_stmts: BTreeMap<Str, Vec<usize>> = BTreeMap::new();
@@ -198,7 +197,7 @@ pub fn schedule<Str: Clone + cmp::Ord + fmt::Debug>(
if order.len() != num_statements {
#[cfg(feature = "std")]
- std::eprintln!("could not schedule all statements {order:?} {orig_infos:?}");
+ std::eprintln!("could not schedule all statements {order:?}");
return Ok(SortResult::Order(
(0..num_statements).map(|i| i as u16).collect(),
));
@@ -633,8 +632,8 @@ impl Analyzer {
) -> Result<(Vec<SourceStr>, Vec<Ref<Expr>>)> {
let mut used_vars = vec![];
let mut comprs = vec![];
+ #[cfg(feature = "deprecated")]
let full_expr = expr;
- core::convert::identity(&full_expr);
traverse(expr, &mut |e| match e.as_ref() {
Var(v) if !matches!(v.0.text(), "_" | "input" | "data") => {
let name = v.0.source_str();
diff --git a/src/tests/interpreter/mod.rs b/src/tests/interpreter/mod.rs
index a68701ef..0613f413 100644
--- a/src/tests/interpreter/mod.rs
+++ b/src/tests/interpreter/mod.rs
@@ -277,6 +277,35 @@ fn yaml_test_impl(file: &str) -> Result<()> {
let yaml_str = std::fs::read_to_string(file)?;
let test: YamlTest = serde_yaml::from_str(&yaml_str)?;
+ #[cfg(not(feature = "std"))]
+ {
+ // Skip tests that depend on bultins that need std feature.
+ let skip = [
+ "intn.yaml",
+ "is_valid.yaml",
+ "add_date.yaml",
+ "date.yaml",
+ "clock.yaml",
+ "compare.yaml",
+ "diff.yaml",
+ "format.yaml",
+ "now_ns.yaml",
+ "parse_duration_ns.yaml",
+ "parse_ns.yaml",
+ "parse_rfc3339_ns.yaml",
+ "weekday.yaml",
+ "generate.yaml",
+ "parse.yaml",
+ "tests.yaml",
+ ];
+ for s in skip {
+ if file.contains(s) {
+ std::println!("skipped {file} in no_std mode.");
+ return Ok(());
+ }
+ }
+ }
+
std::println!("running {file}");
for case in test.cases {
diff --git a/src/value.rs b/src/value.rs
index 212aee38..758368ca 100644
--- a/src/value.rs
+++ b/src/value.rs
@@ -406,6 +406,7 @@ impl Value {
/// Deserialize a value from a file containing YAML.
/// Note: Deserialization from YAML does not support arbitrary precision numbers.
+ #[cfg(feature = "std")]
#[cfg(feature = "yaml")]
pub fn from_yaml_file(path: &String) -> Result<Value> {
match std::fs::read_to_string(path) {
diff --git a/tests/arc.rs b/tests/arc.rs
index 528671b7..47d3bdb4 100644
--- a/tests/arc.rs
+++ b/tests/arc.rs
@@ -1,11 +1,13 @@
// Copyright (c) Microsoft Corporation.
// Licensed under the MIT License.
+#![allow(unused)]
use lazy_static::lazy_static;
use std::sync::Mutex;
use regorus::*;
+#[cfg(feature = "arc")]
// Ensure that types can be s
lazy_static! {
static ref VALUE: Value = Value::Null;
@@ -14,6 +16,7 @@ lazy_static! {
}
#[test]
+#[cfg(feature = "arc")]
fn shared_engine() -> anyhow::Result<()> {
let e_guard = ENGINE.lock();
let mut engine = e_guard.expect("failed to lock engine");
diff --git a/tests/ensure_no_std/Cargo.toml b/tests/ensure_no_std/Cargo.toml
new file mode 100644
index 00000000..bcf7745f
--- /dev/null
+++ b/tests/ensure_no_std/Cargo.toml
@@ -0,0 +1,10 @@
+[package]
+name = "ensure_no_std"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+anyhow = { version = "1.0.83", default-features = false }
+regorus = { path = "../..", default-features = false, features = ["opa-no-std"] }
diff --git a/tests/ensure_no_std/src/main.rs b/tests/ensure_no_std/src/main.rs
new file mode 100644
index 00000000..d4aff8a4
--- /dev/null
+++ b/tests/ensure_no_std/src/main.rs
@@ -0,0 +1,16 @@
+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT License.
+#![no_std]
+#![no_main]
+
+use core::panic::PanicInfo;
+
+#[panic_handler]
+fn panic(_info: &PanicInfo) -> ! {
+ loop {}
+}
+
+#[no_mangle]
+pub extern "C" fn _start() -> ! {
+ loop {}
+}
diff --git a/tests/lexer/mod.rs b/tests/lexer/mod.rs
index 909709f9..ec8f8c64 100644
--- a/tests/lexer/mod.rs
+++ b/tests/lexer/mod.rs
@@ -225,6 +225,7 @@ fn invalid_line() -> Result<()> {
}
#[test]
+#[cfg(feature = "std")]
fn file_more_than_64_kb_size() -> Result<()> {
let source = Source::from_file("tests/kata/data/large.rego")?;
let mut lexer = Lexer::new(&source);
diff --git a/tests/mod.rs b/tests/mod.rs
index b7ed10cb..e0a04c4e 100644
--- a/tests/mod.rs
+++ b/tests/mod.rs
@@ -8,6 +8,3 @@ mod engine;
mod lexer;
mod parser;
mod value;
-
-#[cfg(feature = "arc")]
-mod arc;