From 8481fadbd025a640387aacba6aa7a3c3d93b2068 Mon Sep 17 00:00:00 2001
From: Koitharu <nvasya95@gmail.com>
Date: Thu, 16 Jan 2025 08:51:48 +0200
Subject: [PATCH] Fix absolute url resolving

---
 .../kotlin/org/koitharu/kotatsu/parsers/util/Jsoup.kt |  2 +-
 .../kotlin/org/koitharu/kotatsu/parsers/util/Parse.kt | 11 +++++++----
 2 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Jsoup.kt b/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Jsoup.kt
index a0d7aa07..47f38e1b 100644
--- a/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Jsoup.kt
+++ b/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Jsoup.kt
@@ -79,7 +79,7 @@ public fun Element.attrAsAbsoluteUrlOrNull(attributeKey: String): String? {
 	if (attr.isEmpty() || attr.startsWith("data:")) {
 		return null
 	}
-	return (baseUri().toHttpUrlOrNull()?.newBuilder(attr) ?: return null).toString()
+	return (baseUri().toHttpUrlOrNull()?.resolve(attr) ?: return null).toString()
 }
 
 /**
diff --git a/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Parse.kt b/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Parse.kt
index ac7b4f77..1e96b1aa 100644
--- a/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Parse.kt
+++ b/src/main/kotlin/org/koitharu/kotatsu/parsers/util/Parse.kt
@@ -12,6 +12,9 @@ import org.jsoup.nodes.Document
 import org.koitharu.kotatsu.parsers.InternalParsersApi
 import java.text.DateFormat
 
+private val REGEX_SCHEME_PREFIX = Regex("^\\w{2,6}://", RegexOption.IGNORE_CASE)
+private const val SCHEME_HTTPS = "https"
+
 /**
  * Parse [Response] body as html document using Jsoup
  * @see [parseJson]
@@ -76,10 +79,10 @@ public fun String.toRelativeUrl(domain: String): String {
  * @return an absolute url with [domain] if this is relative
  */
 public fun String.toAbsoluteUrl(domain: String): String = when {
-	this.startsWith("//") -> "https:$this"
-	this.startsWith('/') -> "https://$domain$this"
-	this.startsWith("https://") -> this
-	else -> "https://$domain/$this"
+	startsWith("//") -> "$SCHEME_HTTPS:$this"
+	startsWith('/') -> "$SCHEME_HTTPS://$domain$this"
+	REGEX_SCHEME_PREFIX.containsMatchIn(this) -> this
+	else -> "$SCHEME_HTTPS://$domain/$this"
 }
 
 public fun concatUrl(host: String, path: String): String {