feat: regex related builtins

nix-js/runtime-ts/src/builtins/index.ts

@@ -175,6 +175,8 @@ export const builtins: any = {
   substring: mkPrimop(string.substring, "substring", 3),
   concatStringsSep: mkPrimop(string.concatStringsSep, "concatStringsSep", 2),
   baseNameOf: mkPrimop(string.baseNameOf, "baseNameOf", 1),
+  match: mkPrimop(string.match, "match", 2),
+  split: mkPrimop(string.split, "split", 2),
 
   seq: mkPrimop(functional.seq, "seq", 2),
   deepSeq: mkPrimop(functional.deepSeq, "deepSeq", 2),
@@ -234,14 +236,12 @@ export const builtins: any = {
   functionArgs: mkPrimop(misc.functionArgs, "functionArgs", 1),
   genericClosure: mkPrimop(misc.genericClosure, "genericClosure", 1),
   getFlake: mkPrimop(misc.getFlake, "getFlake", 1),
-  match: mkPrimop(misc.match, "match", 2),
   outputOf: mkPrimop(misc.outputOf, "outputOf", 2),
   parseDrvName: mkPrimop(misc.parseDrvName, "parseDrvName", 1),
   parseFlakeName: mkPrimop(misc.parseFlakeName, "parseFlakeName", 1),
   parseFlakeRef: mkPrimop(misc.parseFlakeRef, "parseFlakeRef", 1),
   placeholder: mkPrimop(misc.placeholder, "placeholder", 1),
   replaceStrings: mkPrimop(misc.replaceStrings, "replaceStrings", 3),
-  split: mkPrimop(misc.split, "split", 2),
   splitVersion: mkPrimop(misc.splitVersion, "splitVersion", 1),
   traceVerbose: mkPrimop(misc.traceVerbose, "traceVerbose", 2),
   tryEval: mkPrimop(misc.tryEval, "tryEval", 1),
@@ -259,6 +259,6 @@ export const builtins: any = {
 
   langVersion: 6,
   nixPath: [],
-  nixVersion: "NIX_JS_VERSION",
+  nixVersion: "2.31.2",
   storeDir: "/nix/store",
 };

nix-js/runtime-ts/src/builtins/misc.ts

@@ -160,12 +160,6 @@ export const getFlake = (attrs: NixValue): never => {
   throw new Error("Not implemented: getFlake");
 };
 
-export const match =
-  (regex: NixValue) =>
-  (str: NixValue): never => {
-    throw new Error("Not implemented: match");
-  };
-
 export const outputOf =
   (drv: NixValue) =>
   (out: NixValue): never => {
@@ -191,16 +185,77 @@ export const placeholder = (output: NixValue): never => {
 export const replaceStrings =
   (from: NixValue) =>
   (to: NixValue) =>
-  (s: NixValue): never => {
-    throw new Error("Not implemented: replaceStrings");
+  (s: NixValue): NixValue => {
+    const fromList = forceList(from);
+    const toList = forceList(to);
+    const inputStr = forceString(s);
+
+    if (fromList.length !== toList.length) {
+      throw new Error(
+        "'from' and 'to' arguments passed to builtins.replaceStrings have different lengths"
+      );
+    }
+
+    const toCache = new Map<number, string>();
+
+    let result = "";
+    let pos = 0;
+
+    while (pos <= inputStr.length) {
+      let found = false;
+
+      for (let i = 0; i < fromList.length; i++) {
+        const pattern = forceString(fromList[i]);
+
+        if (inputStr.substring(pos).startsWith(pattern)) {
+          found = true;
+
+          if (!toCache.has(i)) {
+            toCache.set(i, forceString(toList[i]));
+          }
+          const replacement = toCache.get(i)!;
+
+          result += replacement;
+
+          if (pattern.length === 0) {
+            if (pos < inputStr.length) {
+              result += inputStr[pos];
+            }
+            pos++;
+          } else {
+            pos += pattern.length;
+          }
+          break;
+        }
+      }
+
+      if (!found) {
+        if (pos < inputStr.length) {
+          result += inputStr[pos];
+        }
+        pos++;
+      }
+    }
+
+    return result;
   };
 
-export const split = (regex: NixValue, str: NixValue): never => {
-  throw new Error("Not implemented: split");
-};
 
-export const splitVersion = (s: NixValue): never => {
-  throw new Error("Not implemented: splitVersion");
+export const splitVersion = (s: NixValue): NixValue => {
+  const version = forceString(s);
+  const components: string[] = [];
+  let idx = 0;
+
+  while (idx < version.length) {
+    const result = nextComponent(version, idx);
+    if (result.component === "") {
+      break;
+    }
+    components.push(result.component);
+    idx = result.nextIndex;
+  }
+
+  return components;
 };
 
 export const traceVerbose = (e1: NixValue, e2: NixValue): never => {

nix-js/runtime-ts/src/builtins/string.ts

@@ -42,3 +42,122 @@ export const baseNameOf = (x: NixValue): string => {
 
   return str.substring(pos, last + 1);
 };
+
+const POSIX_CLASSES: Record<string, string> = {
+  alnum: "a-zA-Z0-9",
+  alpha: "a-zA-Z",
+  blank: " \\t",
+  digit: "0-9",
+  lower: "a-z",
+  upper: "A-Z",
+  space: "\\s",
+  xdigit: "0-9A-Fa-f",
+  punct: "\\-!\"#$%&'()*+,./:;<=>?@[\\\\\\]^_`{|}~",
+};
+
+function posixToJsRegex(pattern: string, fullMatch: boolean = false): RegExp {
+  let jsPattern = pattern;
+
+  jsPattern = jsPattern.replace(/\[(\^?)(?:\[:(\w+):\])+\]/g, (match) => {
+    const isNegated = match[1] === "^";
+    const classNames = [...match.matchAll(/\[:(\w+):\]/g)].map((m) => m[1]);
+
+    const combined = classNames
+      .map((className) => {
+        const replacement = POSIX_CLASSES[className];
+        if (!replacement) {
+          throw new Error(`Unknown POSIX character class: ${className}`);
+        }
+        return replacement;
+      })
+      .join("");
+
+    return isNegated ? `[^${combined}]` : `[${combined}]`;
+  });
+
+  jsPattern = jsPattern.replace(/\[:(\w+):\]/g, (_match, className) => {
+    const replacement = POSIX_CLASSES[className];
+    if (!replacement) {
+      throw new Error(`Unknown POSIX character class: ${className}`);
+    }
+    return replacement;
+  });
+
+  if (fullMatch) {
+    if (!jsPattern.startsWith("^")) {
+      jsPattern = "^" + jsPattern;
+    }
+    if (!jsPattern.endsWith("$")) {
+      jsPattern = jsPattern + "$";
+    }
+  }
+
+  return new RegExp(jsPattern, "u");
+}
+
+
+export const match =
+  (regex: NixValue) =>
+  (str: NixValue): NixValue => {
+    const regexStr = forceString(regex);
+    const inputStr = forceString(str);
+
+    try {
+      const re = posixToJsRegex(regexStr, true);
+      const result = inputStr.match(re);
+
+      if (!result) {
+        return null;
+      }
+
+      const groups: NixValue[] = [];
+      for (let i = 1; i < result.length; i++) {
+        groups.push(result[i] !== undefined ? result[i] : null);
+      }
+
+      return groups;
+    } catch (e) {
+      throw new Error(`Invalid regular expression '${regexStr}': ${e}`);
+    }
+  };
+
+export const split =
+  (regex: NixValue) =>
+  (str: NixValue): NixValue => {
+    const regexStr = forceString(regex);
+    const inputStr = forceString(str);
+
+    try {
+      const re = posixToJsRegex(regexStr);
+      const reGlobal = new RegExp(re.source, re.flags + "g");
+
+      const result: NixValue[] = [];
+      let lastIndex = 0;
+      let match: RegExpExecArray | null;
+
+      while ((match = reGlobal.exec(inputStr)) !== null) {
+        result.push(inputStr.substring(lastIndex, match.index));
+
+        const groups: NixValue[] = [];
+        for (let i = 1; i < match.length; i++) {
+          groups.push(match[i] !== undefined ? match[i] : null);
+        }
+        result.push(groups);
+
+        lastIndex = match.index + match[0].length;
+
+        if (match[0].length === 0) {
+          reGlobal.lastIndex++;
+        }
+      }
+
+      if (lastIndex === 0) {
+        return [inputStr];
+      }
+
+      result.push(inputStr.substring(lastIndex));
+      return result;
+    } catch (e) {
+      throw new Error(`Invalid regular expression '${regexStr}': ${e}`);
+    }
+  };

nix-js/src/codegen.rs

@@ -78,7 +78,7 @@ impl<Ctx: CodegenContext> Compile<Ctx> for Ir {
             &Ir::Assert(Assert { assertion, expr }) => {
                 let assertion = ctx.get_ir(assertion).compile(ctx);
                 let expr = ctx.get_ir(expr).compile(ctx);
-                format!("({assertion})?({expr}):(()=>{{throw \"assertion failed\"}})()")
+                format!("({assertion})?({expr}):(()=>{{throw new Error(\"assertion failed\")}})()")
             }
         }
     }

nix-js/tests/regex.rs

@@ -0,0 +1,316 @@
+mod utils;
+
+use nix_js::value::{List, Value};
+use utils::eval;
+
+#[test]
+fn test_match_exact_full_string() {
+    assert_eq!(
+        eval(r#"builtins.match "foobar" "foobar""#),
+        Value::List(List::new(vec![]))
+    );
+}
+
+#[test]
+fn test_match_partial_returns_null() {
+    assert_eq!(eval(r#"builtins.match "foo" "foobar""#), Value::Null);
+}
+
+#[test]
+fn test_match_with_capture_groups() {
+    assert_eq!(
+        eval(r#"builtins.match "(.*)\\.nix" "foobar.nix""#),
+        Value::List(List::new(vec![Value::String("foobar".into())]))
+    );
+}
+
+#[test]
+fn test_match_multiple_capture_groups() {
+    assert_eq!(
+        eval(r#"builtins.match "((.*)/)?([^/]*)\\.nix" "foobar.nix""#),
+        Value::List(List::new(vec![
+            Value::Null,
+            Value::Null,
+            Value::String("foobar".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_match_with_path() {
+    assert_eq!(
+        eval(r#"builtins.match "((.*)/)?([^/]*)\\.nix" "/path/to/foobar.nix""#),
+        Value::List(List::new(vec![
+            Value::String("/path/to/".into()),
+            Value::String("/path/to".into()),
+            Value::String("foobar".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_match_posix_space_class() {
+    assert_eq!(
+        eval(r#"builtins.match "[[:space:]]+([^[:space:]]+)[[:space:]]+" "  foo   ""#),
+        Value::List(List::new(vec![Value::String("foo".into())]))
+    );
+}
+
+#[test]
+fn test_match_posix_upper_class() {
+    assert_eq!(eval(r#"builtins.match "[[:space:]]+([[:upper:]]+)[[:space:]]+" "  foo   ""#), Value::Null);
+
+    assert_eq!(
+        eval(r#"builtins.match "[[:space:]]+([[:upper:]]+)[[:space:]]+" "  FOO   ""#),
+        Value::List(List::new(vec![Value::String("FOO".into())]))
+    );
+}
+
+#[test]
+fn test_match_quantifiers() {
+    assert_eq!(
+        eval(r#"builtins.match "fo*" "f""#),
+        Value::List(List::new(vec![]))
+    );
+
+    assert_eq!(eval(r#"builtins.match "fo+" "f""#), Value::Null);
+
+    assert_eq!(
+        eval(r#"builtins.match "fo{1,2}" "foo""#),
+        Value::List(List::new(vec![]))
+    );
+
+    assert_eq!(eval(r#"builtins.match "fo{1,2}" "fooo""#), Value::Null);
+}
+
+#[test]
+fn test_split_non_capturing() {
+    assert_eq!(
+        eval(r#"builtins.split "foobar" "foobar""#),
+        Value::List(List::new(vec![
+            Value::String("".into()),
+            Value::List(List::new(vec![])),
+            Value::String("".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_no_match() {
+    assert_eq!(
+        eval(r#"builtins.split "fo+" "f""#),
+        Value::List(List::new(vec![Value::String("f".into())]))
+    );
+}
+
+#[test]
+fn test_split_with_capture_group() {
+    assert_eq!(
+        eval(r#"builtins.split "(fo*)" "foobar""#),
+        Value::List(List::new(vec![
+            Value::String("".into()),
+            Value::List(List::new(vec![Value::String("foo".into())])),
+            Value::String("bar".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_multiple_matches() {
+    assert_eq!(
+        eval(r#"builtins.split "(b)" "foobarbaz""#),
+        Value::List(List::new(vec![
+            Value::String("foo".into()),
+            Value::List(List::new(vec![Value::String("b".into())])),
+            Value::String("ar".into()),
+            Value::List(List::new(vec![Value::String("b".into())])),
+            Value::String("az".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_with_multiple_groups() {
+    assert_eq!(
+        eval(r#"builtins.split "(f)(o*)" "foo""#),
+        Value::List(List::new(vec![
+            Value::String("".into()),
+            Value::List(List::new(vec![
+                Value::String("f".into()),
+                Value::String("oo".into())
+            ])),
+            Value::String("".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_with_optional_groups() {
+    assert_eq!(
+        eval(r#"builtins.split "(a)|(c)" "abc""#),
+        Value::List(List::new(vec![
+            Value::String("".into()),
+            Value::List(List::new(vec![Value::String("a".into()), Value::Null])),
+            Value::String("b".into()),
+            Value::List(List::new(vec![Value::Null, Value::String("c".into())])),
+            Value::String("".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_greedy_matching() {
+    assert_eq!(
+        eval(r#"builtins.split "(o+)" "oooofoooo""#),
+        Value::List(List::new(vec![
+            Value::String("".into()),
+            Value::List(List::new(vec![Value::String("oooo".into())])),
+            Value::String("f".into()),
+            Value::List(List::new(vec![Value::String("oooo".into())])),
+            Value::String("".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_posix_classes() {
+    assert_eq!(
+        eval(r#"builtins.split "([[:upper:]]+)" "  FOO   ""#),
+        Value::List(List::new(vec![
+            Value::String("  ".into()),
+            Value::List(List::new(vec![Value::String("FOO".into())])),
+            Value::String("   ".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_replace_basic() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["o"] ["a"] "foobar""#),
+        Value::String("faabar".into())
+    );
+}
+
+#[test]
+fn test_replace_with_empty() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["o"] [""] "foobar""#),
+        Value::String("fbar".into())
+    );
+}
+
+#[test]
+fn test_replace_multiple_patterns() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["oo" "a"] ["a" "oo"] "foobar""#),
+        Value::String("faboor".into())
+    );
+}
+
+#[test]
+fn test_replace_first_match_wins() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["oo" "oo"] ["u" "i"] "foobar""#),
+        Value::String("fubar".into())
+    );
+}
+
+#[test]
+fn test_replace_empty_pattern() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings [""] ["X"] "abc""#),
+        Value::String("XaXbXcX".into())
+    );
+}
+
+#[test]
+fn test_replace_empty_pattern_empty_string() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings [""] ["X"] """#),
+        Value::String("X".into())
+    );
+}
+
+#[test]
+fn test_replace_simple_char() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["-"] ["_"] "a-b""#),
+        Value::String("a_b".into())
+    );
+}
+
+#[test]
+fn test_replace_longer_pattern() {
+    assert_eq!(
+        eval(r#"builtins.replaceStrings ["oo"] ["u"] "foobar""#),
+        Value::String("fubar".into())
+    );
+}
+
+#[test]
+fn test_replace_different_lengths() {
+    let result = std::panic::catch_unwind(|| {
+        eval(r#"builtins.replaceStrings ["a" "b"] ["x"] "test""#)
+    });
+    assert!(result.is_err());
+}
+
+#[test]
+fn test_split_version_simple() {
+    assert_eq!(
+        eval(r#"builtins.splitVersion "1.2.3""#),
+        Value::List(List::new(vec![
+            Value::String("1".into()),
+            Value::String("2".into()),
+            Value::String("3".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_version_with_pre() {
+    assert_eq!(
+        eval(r#"builtins.splitVersion "2.3.0pre1234""#),
+        Value::List(List::new(vec![
+            Value::String("2".into()),
+            Value::String("3".into()),
+            Value::String("0".into()),
+            Value::String("pre".into()),
+            Value::String("1234".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_version_with_letters() {
+    assert_eq!(
+        eval(r#"builtins.splitVersion "2.3a""#),
+        Value::List(List::new(vec![
+            Value::String("2".into()),
+            Value::String("3".into()),
+            Value::String("a".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_version_with_dashes() {
+    assert_eq!(
+        eval(r#"builtins.splitVersion "2.3-beta1""#),
+        Value::List(List::new(vec![
+            Value::String("2".into()),
+            Value::String("3".into()),
+            Value::String("beta".into()),
+            Value::String("1".into())
+        ]))
+    );
+}
+
+#[test]
+fn test_split_version_empty() {
+    assert_eq!(
+        eval(r#"builtins.splitVersion """#),
+        Value::List(List::new(vec![]))
+    );
+}