From 2d652306a6b84f87ae55f03d855a38ff9f7bafe6 Mon Sep 17 00:00:00 2001
From: Ndevu12 <niyokwizerwajeanpaulelisa@email.com>
Date: Sun, 29 Mar 2026 14:18:25 +0000
Subject: [PATCH] fix(core): make isContentEmpty robust to malformed and
 unclosed tags

Replace the regex-based strip (<[^>]*>) with a linear scan over angle brackets. The previous approach could misclassify strings with an unclosed '<' (e.g. literal text or partial tags) because the pattern left odd fragments or dropped visible characters in edge cases.

- Treat text after '<' with no following '>' as visible content, so partial tags are not treated as empty.
- Preserve existing behavior for normal HTML: empty/whitespace-only paragraphs and plain whitespace still count as empty.

Tests cover unclosed brackets, bare tags, multi-line ASCII whitespace inside tags, and tagless whitespace-only input.
---
 src/core/model.ts             | 19 +++++++++++++++++--
 tests/unit/core/model.test.ts | 16 ++++++++++++++++
 2 files changed, 33 insertions(+), 2 deletions(-)

diff --git a/src/core/model.ts b/src/core/model.ts
index f6ed6b0..2e3065a 100644
--- a/src/core/model.ts
+++ b/src/core/model.ts
@@ -28,6 +28,21 @@ export function createEmptyDoc(): string {
  * Strips all tags and checks whether any visible text remains.
  */
 export function isContentEmpty(html: string): boolean {
-  const stripped = html.replace(/<[^>]*>/g, '').trim();
-  return stripped.length === 0;
+  let i = 0;
+  let stripped = '';
+  while (i < html.length) {
+    const lt = html.indexOf('<', i);
+    if (lt === -1) {
+      stripped += html.slice(i);
+      break;
+    }
+    stripped += html.slice(i, lt);
+    const gt = html.indexOf('>', lt + 1);
+    if (gt === -1) {
+      stripped += html.slice(lt);
+      break;
+    }
+    i = gt + 1;
+  }
+  return stripped.trim().length === 0;
 }
diff --git a/tests/unit/core/model.test.ts b/tests/unit/core/model.test.ts
index 911e211..15b431c 100644
--- a/tests/unit/core/model.test.ts
+++ b/tests/unit/core/model.test.ts
@@ -47,5 +47,21 @@ describe('model', () => {
     it('returns true for empty string', () => {
       expect(isContentEmpty('')).toBe(true);
     });
+
+    it('returns false when < is unclosed and there is visible text after it', () => {
+      expect(isContentEmpty('< world')).toBe(false);
+    });
+
+    it('returns false for a bare unclosed tag (no closing >)', () => {
+      expect(isContentEmpty('<foo')).toBe(false);
+    });
+
+    it('returns true for paragraph with only HTML ASCII whitespace between tags', () => {
+      expect(isContentEmpty('<p>\n\t\r\f </p>')).toBe(true);
+    });
+
+    it('returns true for a string with no tags that is only whitespace', () => {
+      expect(isContentEmpty('   \n  ')).toBe(true);
+    });
   });
 });