[pull] master from gchq:master

.gitignore

@@ -8,6 +8,7 @@ build
 src/core/config/modules/*
 src/core/config/OperationConfig.json
 src/core/operations/index.mjs
+src/core/lib/HTMLEntities.mjs
 src/node/config/OperationConfig.json
 src/node/index.mjs
 tests/operations/index.mjs

Gruntfile.js

@@ -367,6 +367,7 @@ module.exports = function (grunt) {
                 command: chainCommands([
                     "echo '\n--- Regenerating config files. ---'",
                     "echo [] > src/core/config/OperationConfig.json",
+                    `node ${nodeFlags} src/core/config/scripts/generateHTMLEntities.mjs`,
                     `node ${nodeFlags} src/core/config/scripts/generateOpsIndex.mjs`,
                     `node ${nodeFlags} src/core/config/scripts/generateConfig.mjs`,
                     "echo '--- Config scripts finished. ---\n'"

src/core/config/Categories.json

@@ -83,7 +83,9 @@
             "Rison Decode",
             "To Modhex",
             "From Modhex",
-            "MIME Decoding"
+            "MIME Decoding",
+            "To COBS",
+            "From COBS"
         ]
     },
     {
@@ -173,6 +175,7 @@
             "AES Key Wrap",
             "AES Key Unwrap",
             "Pseudo-Random Number Generator",
+            "Pseudo-Random Prime Generator",
             "Enigma",
             "Bombe",
             "Multiple Bombe",
@@ -241,9 +244,8 @@
             "Subtract",
             "Multiply",
             "Divide",
-            "Modular Exponentiation",
-            "Modular Inverse",
             "Extended GCD",
+            "Modular Inverse",
             "Mean",
             "Median",
             "Standard Deviation",

src/core/config/scripts/generateHTMLEntities.mjs

@@ -0,0 +1,139 @@
+/**
+ * This script automatically generates src/core/lib/HTMLEntities.mjs, the shared
+ * HTML entity lookup tables used by the "To HTML Entity" and "From HTML Entity"
+ * operations.
+ *
+ * The data is derived from the vendored WHATWG named character reference set
+ * (src/core/vendor/htmlEntities/entity.json, from
+ * https://html.spec.whatwg.org/entities.json) so the
+ * two operations cannot drift apart and every entity is spec-conformant:
+ *
+ *   - HTML_ENTITY_REVERSE_LOOKUP (decode) is every single-code-point spec name.
+ *   - HTML_ENTITY_LOOKUP (encode) picks one canonical name per code point via a
+ *     deterministic tiebreak, overridden by htmlEntityOverrides.mjs where a
+ *     specific historical name is preferred.
+ *
+ * @author roberson-io [michaelroberson@gmail.com]
+ * @copyright Crown Copyright 2026
+ * @license Apache-2.0
+ */
+
+/* eslint no-console: ["off"] */
+
+import path from "path";
+import fs from "fs";
+import process from "process";
+import { fileURLToPath } from "url";
+import { HTML_ENTITY_CANONICAL_OVERRIDES } from "./htmlEntityOverrides.mjs";
+
+const scriptDir = path.dirname(fileURLToPath(import.meta.url));
+
+if (!fs.existsSync(path.join(process.cwd(), "src/core/lib"))) {
+    console.log("\nCWD: " + process.cwd());
+    console.log("Error: generateHTMLEntities.mjs should be run from the project root");
+    console.log("Example> node src/core/config/scripts/generateHTMLEntities.mjs");
+    process.exit(1);
+}
+
+const SPEC = JSON.parse(fs.readFileSync(
+    path.join(scriptDir, "..", "..", "vendor", "htmlEntities", "entity.json"), "utf8"));
+
+// Build code point -> [spec names] for single-code-point, semicolon-terminated
+// references (the representable subset; multi-code-point entities are skipped).
+const codePointNames = {};
+for (const [key, val] of Object.entries(SPEC)) {
+    if (!key.endsWith(";") || val.codepoints.length !== 1) continue;
+    const name = key.slice(1, -1); // strip leading "&" and trailing ";"
+    (codePointNames[val.codepoints[0]] ??= []).push(name);
+}
+
+// Deterministic canonical-name tiebreak: prefer a lower-case name, then the
+// shortest, then alphabetical. Overridden per code point where a specific
+// historical name is preferred.
+const isAllUpper = s => s === s.toUpperCase() && s !== s.toLowerCase();
+
+/**
+ * Choose the single canonical entity name for a code point.
+ *
+ * @param {number} codePoint - the Unicode code point being encoded
+ * @param {string[]} names - all valid WHATWG names for that code point
+ * @returns {string} the canonical name to emit when encoding
+ */
+function canonicalName(codePoint, names) {
+    const override = HTML_ENTITY_CANONICAL_OVERRIDES[codePoint];
+    if (override !== undefined) {
+        if (!names.includes(override))
+            throw new Error(`Override &${override}; is not a spec name for code point ${codePoint} (spec: ${names})`);
+        return override;
+    }
+    return [...names].sort((a, b) =>
+        (isAllUpper(a) - isAllUpper(b)) ||
+        (a.length - b.length) ||
+        (a < b ? -1 : a > b ? 1 : 0)
+    )[0];
+}
+
+const forward = {};   // code point -> canonical name (encode)
+const reverse = {};   // name -> code point (decode, every spec name)
+for (const [cpStr, names] of Object.entries(codePointNames)) {
+    const cp = Number(cpStr);
+    forward[cp] = canonicalName(cp, names);
+    for (const name of names) reverse[name] = cp;
+}
+
+// Decode-only aliases = spec names that are not the canonical encode name.
+const aliases = {};
+for (const [name, cp] of Object.entries(reverse)) {
+    if (forward[cp] !== name) aliases[name] = cp;
+}
+
+// --- emit -----------------------------------------------------------------
+const fwdEntries = Object.keys(forward).map(Number).sort((a, b) => a - b)
+    .map(cp => `    ${cp}: "${forward[cp]}",`).join("\n").replace(/,$/, "");
+const aliasEntries = Object.keys(aliases).sort()
+    .map(n => `    ${JSON.stringify(n)}: ${aliases[n]},`).join("\n").replace(/,$/, "");
+
+const code = `/**
+ * THIS FILE IS AUTOMATICALLY GENERATED BY src/core/config/scripts/generateHTMLEntities.mjs
+ *
+ * HTML entity lookup tables shared by the "To HTML Entity" and "From HTML Entity"
+ * operations, derived from the WHATWG named character reference set
+ * (https://html.spec.whatwg.org/entities.json). Do not edit by hand — change the
+ * vendored src/core/vendor/htmlEntities/entity.json or htmlEntityOverrides.mjs
+ * and regenerate.
+ *
+ * @author roberson-io [michaelroberson@gmail.com]
+ * @copyright Crown Copyright ${new Date().getUTCFullYear()}
+ * @license Apache-2.0
+ */
+
+/**
+ * Canonical lookup: Unicode code point -> entity name (without "&" and ";"),
+ * used for ENCODING. One canonical name per code point.
+ */
+export const HTML_ENTITY_LOOKUP = {
+${fwdEntries}
+};
+
+/**
+ * Legacy / alias names that only DECODE (spelling variants, deprecated names,
+ * box-drawing aliases, etc.); not used for encoding.
+ */
+export const HTML_ENTITY_DECODE_ALIASES = {
+${aliasEntries}
+};
+
+/**
+ * Derived reverse lookup: entity name -> code point, used for DECODING. Built
+ * from the canonical lookup plus the legacy aliases.
+ */
+export const HTML_ENTITY_REVERSE_LOOKUP = {...HTML_ENTITY_DECODE_ALIASES};
+for (const codePoint in HTML_ENTITY_LOOKUP) {
+    HTML_ENTITY_REVERSE_LOOKUP[HTML_ENTITY_LOOKUP[codePoint]] = Number(codePoint);
+}
+`;
+
+fs.writeFileSync(path.join(process.cwd(), "src/core/lib/HTMLEntities.mjs"), code);
+console.log(`generateHTMLEntities: ${Object.keys(forward).length} encode names, ` +
+    `${Object.keys(reverse).length} decode names, ${Object.keys(aliases).length} aliases, ` +
+    `${Object.keys(HTML_ENTITY_CANONICAL_OVERRIDES).length} overrides applied.`);

src/core/config/scripts/htmlEntityOverrides.mjs

@@ -0,0 +1,86 @@
+/**
+ * Canonical entity-name overrides for generateHTMLEntities.mjs.
+ *
+ * The WHATWG named character reference set assigns MANY names to some code
+ * points (e.g. U+2211 is both ∑ and ∑), but does not designate a
+ * canonical one. The generator therefore needs a rule to pick a single name per
+ * code point for ENCODING. It uses a deterministic tiebreak (prefer a
+ * lower-case name, then the shortest, then alphabetical), which reproduces the
+ * historically-emitted name for ~1355 of the ~1414 encodable code points.
+ *
+ * This file pins the canonical name for the code points where the tiebreak
+ * would otherwise change the emitted entity. Every value here is still a valid
+ * WHATWG name for that code point (the generator asserts this) — these are
+ * editorial choices, not correctness fixes, kept so "To HTML Entity" output
+ * stays stable. Trim an entry to let the tiebreak decide instead.
+ *
+ * @author roberson-io [michaelroberson@gmail.com]
+ * @copyright Crown Copyright 2026
+ * @license Apache-2.0
+ */
+
+/**
+ * @constant
+ * @type {Object.<number, string>}
+ */
+export const HTML_ENTITY_CANONICAL_OVERRIDES = {
+    124: "verbar",
+    168: "uml",
+    177: "plusmn",
+    189: "frac12",
+    247: "divide",
+    711: "caron",
+    728: "breve",
+    937: "Omega",
+    949: "epsilon",
+    965: "upsilon",
+    977: "thetasym",
+    978: "upsih",
+    981: "straightphi",
+    8208: "hyphen",
+    8214: "Verbar",
+    8230: "hellip",
+    8289: "ApplyFunction",
+    8290: "InvisibleTimes",
+    8291: "InvisibleComma",
+    8459: "hamilt",
+    8461: "quaternions",
+    8463: "planck",
+    8465: "image",
+    8472: "weierp",
+    8474: "rationals",
+    8476: "real",
+    8477: "reals",
+    8484: "integers",
+    8492: "bernou",
+    8499: "phmmat",
+    8500: "order",
+    8501: "alefsym",
+    8518: "DifferentialD",
+    8519: "ExponentialE",
+    8520: "ImaginaryI",
+    8612: "LeftTeeArrow",
+    8613: "UpTeeArrow",
+    8615: "DownTeeArrow",
+    8624: "lsh",
+    8625: "rsh",
+    8660: "hArr",
+    8704: "forall",
+    8711: "nabla",
+    8712: "isin",
+    8721: "sum",
+    8723: "mnplus",
+    8730: "radic",
+    8750: "conint",
+    8768: "wreath",
+    8776: "asymp",
+    8781: "asympeq",
+    8784: "esdot",
+    8788: "colone",
+    8869: "perp",
+    8896: "xwedge",
+    8897: "xvee",
+    8902: "sstarf",
+    10536: "nesear",
+    10537: "seswar"
+};

src/core/lib/BigIntUtils.mjs

@@ -70,4 +70,3 @@ export function modPow(base, exponent, modulus) {
 
     return result;
 }
-

src/core/lib/COBS.mjs

@@ -0,0 +1,86 @@
+/**
+ * @author Imantas Lukenskas [imantas@lukenskas.dev]
+ * @copyright Imantas Lukenskas 2026
+ * @license Apache-2.0
+ */
+
+import OperationError from "../errors/OperationError.mjs";
+
+/**
+ * COBS-encode a byte array
+ * @param {Uint8Array} data
+ * @return {Uint8Array}
+ */
+export function toCobs(data) {
+    if (!data || data.length === 0) {
+        return new Uint8Array();
+    }
+
+    const output = [];
+    data = [0, ...data];
+
+    while (data.length > 0) {
+        const endIndex = data.findIndex((value, index) => value === 0 && index > 0);
+
+        if ((endIndex < 0 || endIndex > 254) && data.length > 254) {
+            output.push(255);
+            output.push(...data.slice(1, 255));
+            data = data.slice(255);
+            if (data.length !== 0) {
+                data = [0, ...data];
+            }
+        } else if (endIndex < 0) {
+            output.push(data.length);
+            output.push(...data.slice(1));
+            data = [];
+        } else {
+            output.push(endIndex);
+            output.push(...data.slice(1, endIndex));
+            data = data.slice(endIndex);
+        }
+    }
+
+    return output;
+}
+
+/**
+ * COBS-decode a byte array
+ * @param {Uint8Array} data
+ * @return {Uint8Array}
+ */
+export function fromCobs(data) {
+    if (!data || data.length === 0) {
+        return new Uint8Array();
+    }
+
+    if (data.findIndex((value) => value === 0) >= 0) {
+        throw new OperationError("Could not decode from COBS: payload must not contain a 0x00 byte");
+    }
+
+    const output = [];
+
+    while (data.length > 0) {
+        if (data[0] === 0xFF) {
+            output.push(...data.slice(1, 255));
+            data = data.slice(255);
+        } else {
+            const nextZeroIndex = data[0];
+            output.push(...data.slice(1, nextZeroIndex));
+            data = data.slice(nextZeroIndex);
+
+            let blockSize = data[0];
+            while (data.length > 0) {
+                output.push(0, ...data.slice(1, blockSize));
+                data = data.slice(blockSize);
+
+                if (blockSize === 0xFF) {
+                    break;
+                }
+
+                blockSize = data[0];
+            }
+        }
+    }
+
+    return output;
+}