Use redirected URL in UnfurlResult

saket · Mar 20, 2022 · b11a15f · b11a15f
1 parent bd1d011
commit b11a15f
Show file tree

Hide file tree

Showing 3 changed files with 13 additions and 4 deletions.
diff --git a/cli/src/main/kotlin/me/saket/unfurl/cmd/UnfurlCommand.kt b/cli/src/main/kotlin/me/saket/unfurl/cmd/UnfurlCommand.kt
@@ -41,6  41,7 @@ class UnfurlCommand : CliktCommand(name = "unfurl") {
     } else {
       echo("")
       with(unfurled) {
         echo("Url: $url")
         when (title) {
           null -> echo("Title: null")
           else -> echo("Title: \"$title\"")

diff --git a/unfurl/src/main/kotlin/me/saket/unfurl/UnfurlResult.kt b/unfurl/src/main/kotlin/me/saket/unfurl/UnfurlResult.kt
@@ -2,6  2,11 @@ package me.saket.unfurl
 
 import okhttp3.HttpUrl
 
 /**
  * @param url May or may not be equal to the original URL used with [Unfurler.unfurl].
  * This can happen in situations where HTTP 3xx redirects are followed. For example,
  * `https://youtu.be/foo` will redirect to `https://www.youtube.com/watch?v=foo`.
  */
 data class UnfurlResult(
   val url: HttpUrl,
   val title: String?,

diff --git a/unfurl/src/main/kotlin/me/saket/unfurl/delegates/html/HtmlTagsBasedUnfurler.kt b/unfurl/src/main/kotlin/me/saket/unfurl/delegates/html/HtmlTagsBasedUnfurler.kt
@@ -4,6  4,7 @@ import me.saket.unfurl.UnfurlResult
 import me.saket.unfurl.delegates.UnfurlerDelegate
 import me.saket.unfurl.delegates.UnfurlerDelegateScope
 import okhttp3.HttpUrl
 import okhttp3.HttpUrl.Companion.toHttpUrl
 import okhttp3.MediaType
 import okhttp3.Request
 import org.jsoup.Jsoup
@@ -15,7  16,7 @@ class HtmlTagsBasedUnfurler(
   private val parsers = parsers   DefaultHtmlMetadataParser()
 
   override fun UnfurlerDelegateScope.unfurl(url: HttpUrl): UnfurlResult? {
-    return downloadHtml(url)?.extractMetadata(url)
     return downloadHtml(url)?.extractMetadata()
   }
 
   private fun UnfurlerDelegateScope.downloadHtml(url: HttpUrl): Document? {
@@ -33,12  34,14 @@ class HtmlTagsBasedUnfurler(
     return try {
       httpClient.newCall(request).execute().use { response ->
         val body = response.body
         val redirectedUrl = response.request.url
 
         if (body != null && body.contentType().isHtmlText()) {
           // TODO: stream the HTML body only until a "</head>" is received instead of streaming the entire HTML body.
           Jsoup.parse(
             /* in */ body.source().inputStream(),
             /* charsetName */ null,
-            /* baseUri */ url.toString()
             /* baseUri */ redirectedUrl.toString()
           )
         } else {
           null
@@ -50,9  53,9 @@ class HtmlTagsBasedUnfurler(
     }
   }
 
-  private fun Document.extractMetadata(url: HttpUrl): UnfurlResult? {
   private fun Document.extractMetadata(): UnfurlResult? {
     return parsers.asSequence()
-      .map { it.parse(url, document = this) }
       .map { it.parse(url = baseUri().toHttpUrl(), document = this) }
       .firstOrNull()
   }