Content Normalization: Paste, Whitespace, and DOM Hygiene

Overview

Every browser inserts different HTML when a user pastes or hits “Enter.” A robust editor must normalize this “Browser Soup” into a predictable internal schema to prevent data corruption and layout breakage.

Critical Normalization Zones

1. Paste Filter & Cleansing

When pasting from external sources (Word, Excel, Web), browsers inject massive amounts of hidden meta-data and proprietary CSS inside <style> blocks. Strict sanitization is required to strip non-standard attributes.

2. Whitespace &   Management

Browsers follow HTML rules, which collapse consecutive spaces into one. To maintain visual fidelity, editors often use Non-breaking spaces ( ).

• Contamination:   blocks CSS line-wrapping, causing layout overflows. This is severe in plaintext-only mode.
• Conversion: Chrome/Edge frequently convert non-breaking spaces back to regular spaces during editing, causing intended alignment to collapse.

3. Empty Node Pruning

Rapid editing often leaves empty <span>, <b>, or <div> tags in the DOM. These “Ghost Tags” don’t affect visuals but break selection logic and node-count based features.

Normalization Strategy

The Parser Pipeline

Interrupt the paste or beforeinput event and run the incoming HTML through a DOMParser. Apply a strict whitelist of tags and attributes before allowing the insertion.

Whitespace Preservation (CSS over entities)

Prefer white-space: pre-wrap for preserving layouts rather than relying on   chains. If manual intervention is required, use a beforeinput handler to insert \u00A0 only when a trailing space is detected.

Related Cases

• ce-0572: plaintext-only nbsp bug
• ce-0153: nbsp line break prevention
• ce-0102: consecutive spaces collapsed
• ce-0111: empty elements accumulate