diff --git a/pkg/analyzer/content.go b/pkg/analyzer/content.go index 72ecfc9..0d53c94 100644 --- a/pkg/analyzer/content.go +++ b/pkg/analyzer/content.go @@ -27,6 +27,7 @@ import ( "net/http" "net/url" "regexp" + "slices" "strings" "time" "unicode" @@ -37,8 +38,9 @@ import ( // ContentAnalyzer analyzes email content (HTML, links, images) type ContentAnalyzer struct { - Timeout time.Duration - httpClient *http.Client + Timeout time.Duration + httpClient *http.Client + listUnsubscribeURLs []string // URLs from List-Unsubscribe header } // NewContentAnalyzer creates a new content analyzer with configurable timeout @@ -110,6 +112,9 @@ func (c *ContentAnalyzer) AnalyzeContent(email *EmailMessage) *ContentResults { results.IsMultipart = len(email.Parts) > 1 + // Parse List-Unsubscribe header URLs for use in link detection + c.listUnsubscribeURLs = email.GetListUnsubscribeURLs() + // Get HTML and text parts htmlParts := email.GetHTMLParts() textParts := email.GetTextParts() @@ -331,6 +336,11 @@ func (c *ContentAnalyzer) getAttr(n *html.Node, key string) string { // isUnsubscribeLink checks if a link is an unsubscribe link func (c *ContentAnalyzer) isUnsubscribeLink(href string, node *html.Node) bool { + // First check: does the href match a URL from the List-Unsubscribe header? + if slices.Contains(c.listUnsubscribeURLs, href) { + return true + } + // Check href for unsubscribe keywords lowerHref := strings.ToLower(href) unsubKeywords := []string{"unsubscribe", "opt-out", "optout", "remove", "list-unsubscribe"} diff --git a/pkg/analyzer/parser.go b/pkg/analyzer/parser.go index ca3cb46..79d8310 100644 --- a/pkg/analyzer/parser.go +++ b/pkg/analyzer/parser.go @@ -301,3 +301,20 @@ func (e *EmailMessage) GetHeaderValue(key string) string { func (e *EmailMessage) HasHeader(key string) bool { return e.Header.Get(key) != "" } + +// GetListUnsubscribeURLs parses the List-Unsubscribe header and returns all URLs. +// The header format is: , , ... +func (e *EmailMessage) GetListUnsubscribeURLs() []string { + value := e.Header.Get("List-Unsubscribe") + if value == "" { + return nil + } + var urls []string + for _, part := range strings.Split(value, ",") { + part = strings.TrimSpace(part) + if strings.HasPrefix(part, "<") && strings.HasSuffix(part, ">") { + urls = append(urls, part[1:len(part)-1]) + } + } + return urls +}