From 6771e8916a8858b3f2ee921d7fff730db44f86b2 Mon Sep 17 00:00:00 2001
From: "J. King" <jking@jkingweb.ca>
Date: Tue, 23 May 2017 22:15:57 -0400
Subject: [PATCH] Fixed more Feed bugs

- Two items with differing IDs but identical hashes were matching in the deduplicator; they shouldn't
- They would also match in the database matcher, and shouldn't
- The second-pass database matcher was overaggressively finding items as new
---
 lib/Feed.php                                  |  9 +++--
 tests/Feed/TestFeed.php                       |  3 ++
 .../Feed/Deduplication/IdenticalHashes.php    | 37 +++++++++++++++++++
 3 files changed, 46 insertions(+), 3 deletions(-)
 create mode 100644 tests/docroot/Feed/Deduplication/IdenticalHashes.php

diff --git a/lib/Feed.php b/lib/Feed.php
index 7c89dcf..3c48f54 100644
--- a/lib/Feed.php
+++ b/lib/Feed.php
@@ -139,6 +139,8 @@ class Feed {
         $out = [];
         foreach($items as $item) {
             foreach($out as $index => $check) {
+                // if the two items both have IDs and they differ, they do not match, regardless of hashes
+                if($item->id && $check->id && $item->id != $check->id) continue;
                 // if the two items have the same ID or any one hash matches, they are two versions of the same item
                 if(
                     ($item->id && $check->id && $item->id == $check->id) ||
@@ -185,6 +187,8 @@ class Feed {
         foreach($items as $index => $i) {
             $found = false;
             foreach($articles as $a) {
+                // if the item has an ID and it doesn't match the article ID, the two don't match, regardless of hashes
+                if($i->id && $i->id !== $a['guid']) continue;
                 if(
                     // the item matches if the GUID matches...
                     ($i->id && $i->id === $a['guid']) ||
@@ -228,6 +232,8 @@ class Feed {
                 $i = $items[$index];
                 $found = false;
                 foreach($articles as $a) {
+                    // if the item has an ID and it doesn't match the article ID, the two don't match, regardless of hashes
+                    if($i->id && $i->id !== $a['guid']) continue;
                     if(
                         // the item matches if the GUID matches...
                         ($i->id && $i->id === $a['guid']) ||
@@ -252,9 +258,6 @@ class Feed {
                             $found = true;
                             break;
                         }
-                    } else {
-                        // if we don't have a match, add the item to the definite new list
-                        $new[] = $index;
                     }
                 }
                 if(!$found) $new[] = $index;
diff --git a/tests/Feed/TestFeed.php b/tests/Feed/TestFeed.php
index dd294a2..0bd9bef 100644
--- a/tests/Feed/TestFeed.php
+++ b/tests/Feed/TestFeed.php
@@ -27,6 +27,9 @@ class TestFeed extends \PHPUnit\Framework\TestCase {
         $f = new Feed(null, $this->base."Deduplication/ID-Dates");
         $this->assertCount(2, $f->newItems);
         $this->assertTime($t, $f->newItems[0]->updatedDate);
+        $f = new Feed(null, $this->base."Deduplication/IdenticalHashes");
+        $this->assertCount(2, $f->newItems);
+        $this->assertTime($t, $f->newItems[0]->updatedDate);
     }
 
     function testHandleCacheHeadersOn304() {
diff --git a/tests/docroot/Feed/Deduplication/IdenticalHashes.php b/tests/docroot/Feed/Deduplication/IdenticalHashes.php
new file mode 100644
index 0000000..a512631
--- /dev/null
+++ b/tests/docroot/Feed/Deduplication/IdenticalHashes.php
@@ -0,0 +1,37 @@
+<?php return [
+    'mime'    => "application/rss+xml",
+    'content' => <<<MESSAGE_BODY
+<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
+<channel>
+    <title>Test feed</title>
+    <link>http://example.com/</link>
+    <description>A basic feed for testing</description>
+
+    <item>
+        <guid>1</guid>
+        <description>Sample article 2</description>
+        <pubDate>Sun, 19 May 2002 15:21:36 GMT</pubDate> <!-- This is the correct item and date -->
+        <atom:updated>2002-04-19T15:21:36Z</atom:updated>
+    </item>
+    <item>
+        <guid>1</guid>
+        <description>Sample article 2</description>
+        <pubDate>Sun, 19 May 2002 15:21:36 GMT</pubDate>
+        <atom:updated>2002-04-19T15:21:36Z</atom:updated>
+    </item>
+    <item>
+        <guid>1</guid>
+        <description>Sample article 2</description>
+        <pubDate>Sun, 19 May 2002 15:21:36 GMT</pubDate>
+        <atom:updated>2002-04-19T15:21:36Z</atom:updated>
+    </item>
+    <item>
+        <guid>2</guid>
+        <description>Sample article 2</description>
+        <pubDate>Sun, 19 May 2002 15:21:36 GMT</pubDate>
+        <atom:updated>2002-04-19T15:21:36Z</atom:updated>
+    </item>
+</channel>
+</rss>
+MESSAGE_BODY
+];
\ No newline at end of file