Email Processing

Colporteur processes emails in two stages: parsing the raw MIME message, then sanitizing the HTML body for safe feed consumption.

Parsing

Raw emails are parsed using the mailparse crate. Colporteur extracts:

Field	Source	Fallback
Subject	Subject header	"(no subject)"
From	From header	"unknown"
Date	Date header → DateTime<Utc>	Current time
Message-ID	Message-ID header	Generated hash from from+date+subject
Body (HTML)	text/html part	—
Body (text)	text/plain part	—

For multipart messages, the parser recursively walks all MIME parts and collects the first text/html and text/plain bodies found. HTML is preferred over plain text for feed content.

If only a plain text body is available, it’s converted to HTML by escaping entities and wrapping paragraphs in <p> tags.

Sanitization

HTML sanitization uses the ammonia crate with an allowlist approach — only known-safe tags and attributes are kept. Everything else is stripped.

Allowed tags

p, a, img, ul, ol, li, h1–h6, br, hr, strong, em, b, i, blockquote, pre, code, table, thead, tbody, tr, td, th, div, span, sup, sub

What gets stripped

• <script> and <style> blocks (content and tags)
• Event handler attributes (onclick, onload, etc.)
• Tracking pixels — images with both width="1" and height="1"
• Any tags not in the allowlist

Tracking pixel detection

Newsletter emails commonly embed 1x1 invisible images for open tracking. Colporteur detects these by checking for width="1" and height="1" attributes (supporting both single and double quotes) and removes the entire <img> tag.