From c51c8584040f62756d55fadb39eee084c43f9367 Mon Sep 17 00:00:00 2001
From: Sasikanth Miriyampalli <hello@sasikanth.dev>
Date: Sat, 16 Sep 2023 09:56:34 +0530
Subject: [PATCH] Try to fetch feed link from HTML when XML parse error occurs

There are scenarios where XML parsers fails to properly identify if the given document starts with HTML tag and just fails immediately with XML parse error. So, we will just try one final time to see if it's a HTML using Ksoup and try and fetch the URL
---
 .../rss/reader/network/FeedFetcher.kt         | 41 +++++++++++++------
 1 file changed, 29 insertions(+), 12 deletions(-)

diff --git a/shared/src/commonMain/kotlin/dev/sasikanth/rss/reader/network/FeedFetcher.kt b/shared/src/commonMain/kotlin/dev/sasikanth/rss/reader/network/FeedFetcher.kt
index b737aabba..f3f47700d 100644
--- a/shared/src/commonMain/kotlin/dev/sasikanth/rss/reader/network/FeedFetcher.kt
+++ b/shared/src/commonMain/kotlin/dev/sasikanth/rss/reader/network/FeedFetcher.kt
@@ -17,6 +17,7 @@ package dev.sasikanth.rss.reader.network
 
 import com.mohamedrejeb.ksoup.html.parser.KsoupHtmlHandler
 import com.mohamedrejeb.ksoup.html.parser.KsoupHtmlParser
+import dev.sasikanth.rss.reader.utils.XmlParsingError
 import io.ktor.client.HttpClient
 import io.ktor.client.request.get
 import io.ktor.client.statement.HttpResponse
@@ -98,21 +99,33 @@ class FeedFetcher(private val httpClient: HttpClient, private val feedParser: Fe
     return try {
       val feedPayload = feedParser.parse(xmlContent = responseContent, feedUrl = url)
       FeedFetchResult.Success(feedPayload)
-    } catch (e: HtmlContentException) {
-      val newUrl = fetchFeedLinkFromHtmlIfExists(responseContent)
-      val host = URLBuilder(url).build().host
-      val rootUrl = "https://$host"
-      val feedUrl = FeedParser.safeUrl(rootUrl, newUrl)
-
-      if (!feedUrl.isNullOrBlank()) {
-        fetch(feedUrl)
-      } else {
-        throw UnsupportedOperationException()
+    } catch (e: Exception) {
+      when (e) {
+        // There are situation where XML parsers fail to identify if it's
+        // a HTML document and fail, so trying to fetch link with HTML one
+        // last time just to be safe if it fails with XML parsing issue.
+        is HtmlContentException,
+        is XmlParsingError -> {
+          val feedUrl = fetchFeedLinkFromHtmlIfExists(responseContent, url)
+          if (!feedUrl.isNullOrBlank()) {
+            fetch(feedUrl)
+          } else {
+            if (e is XmlParsingError) {
+              throw e
+            } else {
+              throw UnsupportedOperationException()
+            }
+          }
+        }
+        else -> throw e
       }
     }
   }
 
-  private suspend fun fetchFeedLinkFromHtmlIfExists(htmlContent: String): String? {
+  private suspend fun fetchFeedLinkFromHtmlIfExists(
+    htmlContent: String,
+    originalUrl: String
+  ): String? {
     return suspendCoroutine { continuation ->
       var link: String? = null
       KsoupHtmlParser(
@@ -134,7 +147,11 @@ class FeedFetcher(private val httpClient: HttpClient, private val feedParser: Fe
               }
 
               override fun onEnd() {
-                continuation.resume(link)
+                val host = URLBuilder(originalUrl).build().host
+                val rootUrl = "https://$host"
+                val feedUrl = FeedParser.safeUrl(rootUrl, link)
+
+                continuation.resume(feedUrl)
               }
             }
         )