From 7a6353a3b3d9c7c0a8cedf34c4aeb2eaacd2da79 Mon Sep 17 00:00:00 2001
From: Thomas von Dein <tom@vondein.org>
Date: Tue, 21 Oct 2025 20:46:05 +0200
Subject: [PATCH] clean svg and cdata

---
 pkg/epub/content.go | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/pkg/epub/content.go b/pkg/epub/content.go
index af2273d..8336b73 100644
--- a/pkg/epub/content.go
+++ b/pkg/epub/content.go
@@ -10,7 +10,8 @@ import (
 var (
 	cleanentitles = regexp.MustCompile(`&[a-z]+;`)
 	empty         = regexp.MustCompile(`(?s)^[\s ]*$`)
-	newlines      = regexp.MustCompile(`[\r\n]+`)
+	newlines      = regexp.MustCompile(`[\r\n\s]+`)
+	cleansvg      = regexp.MustCompile(`(<svg.+</svg>|<!\[CDATA\[.+\]\]>)`)
 	cleanmarkup   = regexp.MustCompile(`<[^<>]+>`)
 )
 
@@ -27,7 +28,8 @@ type Content struct {
 func (c *Content) String(content []byte) error {
 	doc, err := xmlquery.Parse(
 		strings.NewReader(
-			cleanentitles.ReplaceAllString(string(content), " ")))
+			cleansvg.ReplaceAllString(
+				cleanentitles.ReplaceAllString(string(content), " "), "")))
 	if err != nil {
 		return err
 	}