feat(stdlib): add optional psl argument to parse_etld

Cargo.toml

@@ -88,6 +88,8 @@ stdlib = [
  "dep:prost",
  "dep:prost-reflect",
  "dep:psl",
+ "dep:psl-types",
+ "dep:publicsuffix",
  "dep:quoted_printable",
  "dep:rand",
  "dep:roxmltree",
@@ -154,6 +156,8 @@ prettytable-rs = { version = "0.10", default-features = false, optional = true }
 quickcheck = { version = "1", optional = true }
 quoted_printable = {version = "0.5", optional = true }
 psl = { version = "2", optional = true }
+psl-types = { version = "2", optional = true }
+publicsuffix = { version = "2", optional = true }
 rand = { version = "0.8", optional = true }
 regex = { version = "1", default-features = false, optional = true, features = ["std", "perf", "unicode"] }
 roxmltree = { version = "0.19", optional = true }

LICENSE-3rdparty.csv

@@ -171,6 +171,7 @@ prost-reflect,https://github.com/andrewhickman/prost-reflect,MIT OR Apache-2.0,A
 psl,https://github.com/addr-rs/psl,MIT OR Apache-2.0,rushmorem <rushmore@webenchanter.com>
 psl-types,https://github.com/addr-rs/psl-types,MIT OR Apache-2.0,rushmorem <rushmore@webenchanter.com>
 ptr_meta,https://github.com/djkoloski/ptr_meta,MIT,David Koloski <djkoloski@gmail.com>
+publicsuffix,https://github.com/rushmorem/publicsuffix,MIT OR Apache-2.0,rushmorem <rushmore@webenchanter.com>
 quanta,https://github.com/metrics-rs/quanta,MIT,Toby Lawrence <toby@nuclearfurnace.com>
 quote,https://github.com/dtolnay/quote,MIT OR Apache-2.0,David Tolnay <dtolnay@gmail.com>
 quoted_printable,https://github.com/staktrace/quoted-printable,0BSD,Kartikaya Gupta <kats@seldon.staktrace.com>

changelog.d/851.feature.md

@@ -0,0 +1,3 @@
+Added `psl` argument to the `parse_etld` function. It enables customizing used public suffix list. If none is provided the default (https://publicsuffix.org/list/public_suffix_list.dat) is used, which is that was used before this change.
+
+authors: esensar

lib/tests/tests/functions/custom_public_suffix_list.dat

@@ -0,0 +1,3 @@
+// ===BEGIN ICANN DOMAINS===
+
+customdev

lib/tests/tests/functions/parse_etld/custom_psl_file.vrl

@@ -0,0 +1,4 @@
+# object: { "host": "vector.customdev" }
+# result: { "etld": "customdev", "etld_plus": "vector.customdev", "known_suffix": true }
+
+parse_etld!(.host, plus_parts: 1, psl: "lib/tests/tests/functions/custom_public_suffix_list.dat")

lib/tests/tests/functions/parse_etld/custom_psl_file_does_not_exist.vrl

@@ -0,0 +1,18 @@
+# object: { "host": "vector.customdev" }
+# result:
+#
+# error[E610]: function compilation error: error[E403] invalid argument
+# ┌─ :2:1
+# │
+# 2 │ parse_etld!(.host, plus_parts: 1, psl: "lib/tests/tests/functions/definitelydoesnot.exist")
+# │ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+# │ │
+# │ invalid argument "psl"
+# │ error: Unable to read psl file
+# │ received: "\"lib/tests/tests/functions/definitelydoesnot.exist\""
+# │
+# = learn more about error code 403 at https://errors.vrl.dev/403
+# = see language documentation at https://vrl.dev
+# = try your code in the VRL REPL, learn more at https://vrl.dev/examples
+
+parse_etld!(.host, plus_parts: 1, psl: "lib/tests/tests/functions/definitelydoesnot.exist")

lib/tests/tests/functions/parse_etld/custom_psl_file_wrong_format.vrl

@@ -0,0 +1,18 @@
+# object: { "host": "vector.customdev" }
+# result:
+#
+# error[E610]: function compilation error: error[E403] invalid argument
+# ┌─ :2:1
+# │
+# 2 │ parse_etld!(.host, plus_parts: 1, psl: "lib/tests/tests/functions/parse_groks_alias_source.json")
+# │ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+# │ │
+# │ invalid argument "psl"
+# │ error: Unable to parse psl file
+# │ received: "\"lib/tests/tests/functions/parse_groks_alias_source.json\""
+# │
+# = learn more about error code 403 at https://errors.vrl.dev/403
+# = see language documentation at https://vrl.dev
+# = try your code in the VRL REPL, learn more at https://vrl.dev/examples
+
+parse_etld!(.host, plus_parts: 1, psl: "lib/tests/tests/functions/parse_groks_alias_source.json")

lib/tests/tests/functions/parse_etld/etld_plus_one.vrl

@@ -1,5 +1,4 @@
 # object: { "host": "vector.dev" }
 # result: { "etld": "dev", "etld_plus": "vector.dev", "known_suffix": true }
 
-etld_result = parse_etld!(.host, plus_parts: 1)
-etld_result
+parse_etld!(.host, plus_parts: 1)

lib/tests/tests/functions/parse_etld/etld_plus_ten.vrl

@@ -1,5 +1,4 @@
 # object: { "host": "vector.dev" }
 # result: { "etld": "dev", "etld_plus": "vector.dev", "known_suffix": true }
 
-etld_result = parse_etld!(.host, plus_parts: 10)
-etld_result
+parse_etld!(.host, plus_parts: 10)

src/stdlib/parse_etld.rs

@@ -1,5 +1,8 @@
+use psl::Psl;
+use publicsuffix::List;
+
 use crate::compiler::prelude::*;
-use std::collections::BTreeMap;
+use std::{collections::BTreeMap, path::Path};
 
 #[derive(Clone, Copy, Debug)]
 pub struct ParseEtld;
@@ -21,6 +24,11 @@ impl Function for ParseEtld {
                 kind: kind::INTEGER,
                 required: false,
             },
+            Parameter {
+                keyword: "psl",
+                kind: kind::BYTES,
+                required: false,
+            },
         ]
     }
 
@@ -64,21 +72,62 @@ impl Function for ParseEtld {
 
     fn compile(
         &self,
-        _state: &state::TypeState,
+        state: &state::TypeState,
         _ctx: &mut FunctionCompileContext,
         arguments: ArgumentList,
     ) -> Compiled {
         let value = arguments.required("value");
         let plus_parts = arguments.optional("plus_parts").unwrap_or_else(|| expr!(0));
 
-        Ok(ParseEtldFn { value, plus_parts }.as_expr())
+        let psl_expr = arguments.optional_expr("psl");
+        let mut psl: Option<List> = None;
+        if let Some(psl_expr) = psl_expr {
+            let psl_location = psl_expr
+                .clone()
+                .resolve_constant(state)
+                .ok_or(function::Error::ExpectedStaticExpression {
+                    keyword: "psl",
+                    expr: psl_expr.clone(),
+                })?
+                .try_bytes_utf8_lossy()
+                .map_err(|_| function::Error::InvalidArgument {
+                    keyword: "psl",
+                    value: format!("{psl_expr:?}").into(),
+                    error: "psl should be a string",
+                })?
+                .into_owned();
+
+            let path = Path::new(&psl_location);
+            psl = Some(
+                std::fs::read_to_string(path)
+                    .map_err(|_| function::Error::InvalidArgument {
+                        keyword: "psl",
+                        value: format!("{path:?}").into(),
+                        error: "Unable to read psl file",
+                    })?
+                    .parse()
+                    .map_err(|_| function::Error::InvalidArgument {
+                        keyword: "psl",
+                        value: format!("{path:?}").into(),
+                        error: "Unable to parse psl file",
+                    })?,
+            );
+        }
+
+        Ok(ParseEtldFn {
+            value,
+            plus_parts,
+            psl,
+        }
+        .as_expr())
     }
 }
 
 #[derive(Debug, Clone)]
 struct ParseEtldFn {
     value: Box<dyn Expression>,
     plus_parts: Box<dyn Expression>,
+    psl: Option<List>,
 }
 
 impl FunctionExpression for ParseEtldFn {
@@ -91,8 +140,12 @@ impl FunctionExpression for ParseEtldFn {
             x => x as usize,
         };
 
-        let etld = psl::suffix(string.as_bytes())
-            .ok_or(format!("unable to determine eTLD for {string}"))?;
+        let suffix_result = if let Some(list) = &self.psl {
+            list.suffix(string.as_bytes())
+        } else {
+            psl::suffix(string.as_bytes())
+        };
+        let etld = suffix_result.ok_or(format!("unable to determine eTLD for {string}"))?;
         let etld_string = core::str::from_utf8(etld.as_bytes())
             .map_err(|err| format!("could not convert eTLD to UTF8 {err}"))?;
 
@@ -236,5 +289,11 @@ mod tests {
             want: Err("unable to determine eTLD for "),
             tdef: TypeDef::object(inner_kind()).fallible(),
         }
+
+        bad_psl_file {
+            args: func_args![value: value!("vector.dev"), psl: value!("definitelynotafile")],
+            want: Err("invalid argument"),
+            tdef: TypeDef::object(inner_kind()).fallible(),
+        }
     ];
 }