@stxt-lang/core 0.7.0 → 0.8.0

package/README.md

@@ -7,6 +7,7 @@ STXT is a plain-text format for writing structured, semantic documents: no brace
 - Website and language reference: <https://stxt.dev>
 - VSCode extension: [STXT - Semantic Text](https://marketplace.visualstudio.com/items?itemName=stxt-lang.stxt)
 - Java implementation: [`dev.stxt:stxt-core`](https://central.sonatype.com/artifact/dev.stxt/stxt-core) on Maven Central
+- Python implementation: [`stxt`](https://pypi.org/project/stxt/) on PyPI
 
 ## What STXT looks like
 
@@ -149,7 +150,7 @@ for (const error of result.getErrors()) {
 }
 ```
 
-Available value types: `INLINE`, `BLOCK`, `TEXT`, `BOOLEAN`, `INTEGER`, `NATURAL`, `NUMBER`, `DATE`, `TIMESTAMP`, `EMAIL`, `URL`, `HEXADECIMAL`, `BASE64`, `GROUP`, `ENUM`.
+Available value types: `INLINE`, `BLOCK`, `TEXT`, `MARKDOWN`, `BOOLEAN`, `INTEGER`, `NATURAL`, `NUMBER`, `DATE`, `TIME`, `TIMESTAMP`, `UUID`, `EMAIL`, `URL`, `HEXADECIMAL`, `BINARY`, `BASE64`, `GROUP`, `ENUM`.
 
 ## Finding the schemas: discovery
 

package/out/core/Parser.js

@@ -90,6 +90,12 @@ class Parser {
             // Parse the line
             const line = (0, LineParser_1.parseLine)(lineString, lastNodeText, lastLevel, lineNumber);
             if (line.isComment) {
+                // A comment at the level of an open block node (or shallower) closes the block
+                // (spec 6.1 and 9.1): a block is a literal and cannot be commented from inside.
+                // Only the block closes; the comment does not touch the rest of the hierarchy.
+                if (lastNodeText) {
+                    this.closeToLevel(stack, stack.length - 1, result);
+                }
                 // Hand it over to the observers
                 this.observers.forEach(observer => {
                     observer.onComment(lineNumber, lineString);

package/out/core/StringUtils.d.ts

@@ -1,6 +1,7 @@
 /** String normalization helpers used for names, namespaces and values. */
 export declare class StringUtils {
     private static readonly NODE_NAME;
+    private static readonly NODE_NAME_LETTER_OR_DIGIT;
     private constructor();
     /**
      * Removes the trailing whitespace of a string.

package/out/core/StringUtils.js

@@ -62,7 +62,7 @@ class StringUtils {
      */
     static isValidNodeName(input) {
         const nfc = this.compactSpaces(input).normalize("NFC");
-        return this.NODE_NAME.test(nfc) && this.normalize(nfc).length > 0;
+        return this.NODE_NAME.test(nfc) && this.NODE_NAME_LETTER_OR_DIGIT.test(nfc);
     }
     // Used for the normalized name of the nodes (STXT-SPEC 4.3): NFC + lower case,
     // keeping diacritics and non-Latin alphabets (IDN model)
@@ -87,7 +87,9 @@ class StringUtils {
     }
 }
 exports.StringUtils = StringUtils;
-// STXT-SPEC 4.2 / 4.3: validate the logical name after NFC so that a
+// STXT-SPEC 4.2 / 4.3: letters, decimal digits, combining marks (Mn, Mc) and the three
+// separators, with at least one letter or digit. Validated after NFC so that a
 // decomposed spelling such as "e" + combining acute is accepted as "é".
-StringUtils.NODE_NAME = /^[\p{L}\p{Nd}\-_ ]+$/u;
+StringUtils.NODE_NAME = /^[\p{L}\p{Nd}\p{Mn}\p{Mc}\-_ ]+$/u;
+StringUtils.NODE_NAME_LETTER_OR_DIGIT = /[\p{L}\p{Nd}]/u;
 //# sourceMappingURL=StringUtils.js.map

package/out/runtime/ConditionalValidator.d.ts

@@ -5,6 +5,10 @@ import { ValidationException } from "../exceptions/ValidationException";
 /**
  * Wrapper around a {@link SchemaValidator} that only validates namespaced nodes, so that a
  * document mixing schema-bound and free nodes does not report the free ones as unknown.
+ *
+ * @deprecated since 0.8.0: {@link SchemaValidator} applies this rule itself (STXT-SCHEMA-SPEC 5,
+ * the empty namespace is never validated), so the wrapper adds nothing. Register the
+ * `SchemaValidator` directly. Kept for compatibility; to be removed in 1.0.
  */
 export declare class ConditionalValidator implements Validator {
     private readonly schemaValidator;

package/out/runtime/ConditionalValidator.js

@@ -4,6 +4,10 @@ exports.ConditionalValidator = void 0;
 /**
  * Wrapper around a {@link SchemaValidator} that only validates namespaced nodes, so that a
  * document mixing schema-bound and free nodes does not report the free ones as unknown.
+ *
+ * @deprecated since 0.8.0: {@link SchemaValidator} applies this rule itself (STXT-SCHEMA-SPEC 5,
+ * the empty namespace is never validated), so the wrapper adds nothing. Register the
+ * `SchemaValidator` directly. Kept for compatibility; to be removed in 1.0.
  */
 class ConditionalValidator {
     /**

package/out/schema/SchemaValidator.js

@@ -26,6 +26,18 @@ class SchemaValidator {
         const errors = [];
         // Get the namespace
         const namespace = node.getNamespace();
+        // The empty namespace is never validated (STXT-SCHEMA-SPEC 5): a node that neither
+        // declares nor inherits a namespace is valid by definition, no schema is looked up for
+        // it and SCHEMA_NOT_FOUND is never reported for it. Its children are still walked when
+        // recursive, because one of them may declare a namespace of its own.
+        if (namespace === "") {
+            if (this.recursiveValidation && node instanceof InlineNode_1.InlineNode) {
+                for (const childNode of node.getChildren()) {
+                    errors.push(...this.validate(childNode));
+                }
+            }
+            return errors;
+        }
         const schema = this.schemaProvider.getSchema(namespace);
         if (!schema) {
             errors.push(new ValidationException_1.ValidationException(node.getLine(), "SCHEMA_NOT_FOUND", `Not found schema: ${namespace}`));

package/out/schema/type/EMAIL.d.ts

@@ -1,2 +1,8 @@
-/** `EMAIL` type: checks that the value looks like an e-mail address. */
+/**
+ * `EMAIL` type: checks that the value is an e-mail address, in either of the two forms of
+ * STXT-SCHEMA-SPEC 9.4: the bare address (`user@domain.tld`) or a display name followed by the
+ * address between angle brackets (`Joan Costa <joan@example.com>`). The display name is any
+ * non-empty text without `<` or `>` (quotes are not interpreted) and the space before `<` is
+ * optional; `<`/`>` without a name, unbalanced or followed by anything are rejected.
+ */
 export declare const EMAIL: import("../Type").Type;

package/out/schema/type/EMAIL.js

@@ -2,6 +2,21 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.EMAIL = void 0;
 const regexType_1 = require("./regexType");
-/** `EMAIL` type: checks that the value looks like an e-mail address. */
-exports.EMAIL = (0, regexType_1.regexType)("EMAIL", /^(?=.{1,256}$)(?=.{1,64}@.{1,255}$)(?=.{1,64}@.{1,63}\..{1,63}$)[A-Za-z0-9!#$%&'*+/=?^_`{|}~.-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}$/, "Invalid email");
+/**
+ * The address proper, `local@domain` with the usual length limits, as it reads when it ends the
+ * value (bare form)...
+ */
+const ADDRESS = "(?=.{1,256}$)(?=.{1,64}@.{1,255}$)(?=.{1,64}@.{1,63}\\..{1,63}$)"
+    + "[A-Za-z0-9!#$%&'*+/=?^_`{|}~.-]+@[A-Za-z0-9.-]+\\.[A-Za-z]{2,}";
+/** ...and the same address as it reads between `<` and `>` (the lookaheads stop at the `>`). */
+const BRACKETED = "(?=[^>]{1,256}>$)(?=[^>]{1,64}@[^>]{1,255}>$)(?=[^>]{1,64}@[^>]{1,63}\\.[^>]{1,63}>$)"
+    + "[A-Za-z0-9!#$%&'*+/=?^_`{|}~.-]+@[A-Za-z0-9.-]+\\.[A-Za-z]{2,}";
+/**
+ * `EMAIL` type: checks that the value is an e-mail address, in either of the two forms of
+ * STXT-SCHEMA-SPEC 9.4: the bare address (`user@domain.tld`) or a display name followed by the
+ * address between angle brackets (`Joan Costa <joan@example.com>`). The display name is any
+ * non-empty text without `<` or `>` (quotes are not interpreted) and the space before `<` is
+ * optional; `<`/`>` without a name, unbalanced or followed by anything are rejected.
+ */
+exports.EMAIL = (0, regexType_1.regexType)("EMAIL", new RegExp(`^(?:[^<>]*[^<>\\s]\\s*<${BRACKETED}>|${ADDRESS})$`), "Invalid email");
 //# sourceMappingURL=EMAIL.js.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@stxt-lang/core",
-  "version": "0.7.0",
+  "version": "0.8.0",
   "description": "Parser and schema validator for STXT, an indentation-based structured-text format.",
   "main": "out/all.js",
   "types": "out/all.d.ts",