fix(import): scan the KML root delimiter in bounded time

docs/CURRENT-STATE.md

@@ -1,5 +1,11 @@
 # God's Eye View Current State
 
+KML namespace repair locates the opening token once and scans its quoted tag
+delimiter once. Malformed or unterminated tags fail before DOM parsing, so a file
+of repeated incomplete `<kml` prefixes cannot trigger quadratic regex rescans.
+Duplicate namespace repair and missing xsi declarations remain supported; GPX
+continues through its existing XML parsing path.
+
 Shared distance checks use `src/geoDistance.js` with `{ lat, lon }` points and
 the IUGG mean spherical Earth radius, 6371.0088 km. `greatCircleKm` is
 canonical; `greatCircleMeters` is its exact metre conversion. Missing,

src/data/geoFileImport.js

@@ -1040,26 +1040,86 @@ export function detachImportedTime(dataSource) {
   return dataSource;
 }
 
+/** Repair only the first KML opening tag, scanning its delimiter once. */
+export function repairKmlRootNamespaces(value, what = 'KML') {
+  const text = String(value);
+  // Skip XML comments and processing instructions without treating their
+  // literal '<kml' text as an opening root tag.
+  let start = 0;
+  for (;;) {
+    start = text.indexOf('<', start);
+    if (start < 0) return text;
+    if (text.slice(start, start + 9).toUpperCase() === '<!DOCTYPE') {
+      let quote = null;
+      let subsetDepth = 0;
+      let end = start + 9;
+      for (; end < text.length; end++) {
+        const char = text[end];
+        if (quote) {
+          if (char === quote) quote = null;
+        } else if (char === '"' || char === "'") quote = char;
+        else if (char === '[') subsetDepth++;
+        else if (char === ']') subsetDepth--;
+        else if (char === '>' && subsetDepth === 0) break;
+      }
+      if (end === text.length) return text;
+      start = end + 1;
+      continue;
+    }
+    const close = text.startsWith('<!--', start)
+      ? '-->'
+      : text.startsWith('<?', start)
+        ? '?>'
+        : null;
+    if (!close) break;
+    const end = text.indexOf(close, start + 2);
+    if (end < 0) return text; // The XML parser reports an unterminated preamble.
+    start = end + close.length;
+  }
+  const match = /^<kml\b/i.exec(text.slice(start, start + 5));
+  if (!match) return text;
+  let quote = null;
+  let end = -1;
+  for (let index = start + match[0].length; index < text.length; index++) {
+    const char = text[index];
+    if (char === '<') break; // Raw '<' is invalid even inside an XML attribute.
+    if (quote) {
+      if (char === quote) quote = null;
+    } else if (char === '"' || char === "'") {
+      quote = char;
+    } else if (char === '>') {
+      end = index + 1;
+      break;
+    }
+  }
+  if (end < 0)
+    throw new GeoImportError(
+      `That ${what} file is not valid XML.`,
+      msg('That {format} file is not valid XML.'),
+      { format: what },
+    );
+  const seen = new Set();
+  let root = text
+    .slice(start, end)
+    .replace(
+      /\sxmlns:([\w.-]+)\s*=\s*(["'])[^"']*\2/g,
+      (declaration, prefix) => {
+        if (seen.has(prefix)) return '';
+        seen.add(prefix);
+        return declaration;
+      },
+    );
+  if (!seen.has('xsi') && /[<\s]xsi:/.test(text))
+    root = root.replace(
+      /(\/?)>$/,
+      ' xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"$1>',
+    );
+  return text.slice(0, start) + root + text.slice(end);
+}
+
 function parseXml(text, what) {
   // Match the namespace repairs Cesium applies to ordinary exported KML.
-  if (what !== 'GPX')
-    text = String(text).replace(/<kml\b[^>]*>/i, (root) => {
-      const seen = new Set();
-      root = root.replace(
-        /\s+xmlns:([\w.-]+)\s*=\s*(["'])[^"']*\2/g,
-        (declaration, prefix) => {
-          if (seen.has(prefix)) return '';
-          seen.add(prefix);
-          return declaration;
-        },
-      );
-      if (!seen.has('xsi') && /[<\s]xsi:/.test(text))
-        root = root.replace(
-          />$/,
-          ' xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">',
-        );
-      return root;
-    });
+  if (what !== 'GPX') text = repairKmlRootNamespaces(text, what);
   const doc = new DOMParser().parseFromString(
     String(text).replace(/^/, ''),
     'application/xml',

src/data/geoFileImport.test.mjs

@@ -1,9 +1,11 @@
 import test from 'node:test';
 import assert from 'node:assert/strict';
 import { deflateRawSync } from 'node:zlib';
+import { spawnSync } from 'node:child_process';
 import * as Cesium from 'cesium';
 import {
   GeoImportError,
+  repairKmlRootNamespaces,
   loadGeoFile,
   MAX_IMPORT_POSITIONS,
   MAX_IMPORT_ENTITIES,
@@ -1171,3 +1173,54 @@ test('a dated import exposes no clock and draws for all time', () => {
   assert.equal(recorded.getValue(Cesium.JulianDate.now()).length, 2);
   assert.ok(track.polyline.clampToGround.getValue());
 });
+
+test('KML root namespace repair respects quoted delimiters and preserves the document tail', () => {
+  const text =
+    '<KML note="a > b" xmlns:x="first" xmlns:x="second"><Placemark xsi:type="test"/></KML>';
+  const result = repairKmlRootNamespaces(text);
+  assert.match(result, /note="a > b"/);
+  assert.match(result, /xmlns:x="first"/);
+  assert.doesNotMatch(result, /second/);
+  assert.match(
+    result,
+    /xmlns:xsi="http:\/\/www.w3.org\/2001\/XMLSchema-instance"/,
+  );
+  assert.ok(result.endsWith('<Placemark xsi:type="test"/></KML>'));
+  assert.match(
+    repairKmlRootNamespaces('<kml xsi:type="test"/>'),
+    /XMLSchema-instance"\/>$/,
+  );
+  for (const malformed of ['<kml ', '<kml name="unterminated>', '<kml <kml >'])
+    assert.throws(() => repairKmlRootNamespaces(malformed), GeoImportError);
+});
+
+test('repeated incomplete KML roots finish within an isolated process deadline', () => {
+  const source = `import { repairKmlRootNamespaces } from ${JSON.stringify(new URL('./geoFileImport.js', import.meta.url).href)};
+    try { repairKmlRootNamespaces('<kml '.repeat(1000000)); process.exit(1); }
+    catch (error) { if (error.name !== 'GeoImportError') throw error; }`;
+  const child = spawnSync(
+    process.execPath,
+    ['--input-type=module', '-e', source],
+    { timeout: 5000, encoding: 'utf8' },
+  );
+  assert.ifError(child.error);
+  assert.equal(child.status, 0, child.stderr);
+});
+
+test('KML root repair ignores incomplete tag text inside preamble comments', () => {
+  const input =
+    '<?xml version="1.0"?><!-- example <kml incomplete -->\n<kml xmlns:gx="a" xmlns:gx="b"><Document/></kml>';
+  assert.equal(
+    repairKmlRootNamespaces(input),
+    input.replace(' xmlns:gx="b"', ''),
+  );
+});
+
+test('KML namespace repair follows a doctype with a quoted root-like entity', () => {
+  const input =
+    '<!DOCTYPE kml [<!ENTITY example "<kml incomplete >">]><kml xmlns:gx="a" xmlns:gx="b"/>';
+  assert.equal(
+    repairKmlRootNamespaces(input),
+    input.replace(' xmlns:gx="b"', ''),
+  );
+});