From 96d6d7ee01320071bfb2f99374700b9eac4ccf80 Mon Sep 17 00:00:00 2001
From: "J. King" <jking@jkingweb.ca>
Date: Thu, 27 Jul 2023 12:25:20 -0400
Subject: [PATCH] Fix simple textContent

---
 lib/Microformats/Parser.php  | 21 ++++++++++++---------
 tests/cases/StandardTest.php |  8 +++++---
 2 files changed, 17 insertions(+), 12 deletions(-)

diff --git a/lib/Microformats/Parser.php b/lib/Microformats/Parser.php
index fb2c926..bf0c7b8 100644
--- a/lib/Microformats/Parser.php
+++ b/lib/Microformats/Parser.php
@@ -1398,16 +1398,19 @@ class Parser {
         # replacing any nested <img> elements with their alt attribute, if
         #   present; otherwise their src attribute, if present, adding a
         #   space at the beginning and end, resolving the URL if it’s
-        #   relative;
-        foreach ($copy->getElementsByTagName("img") as $e) {
-            $alt = $e->getAttribute("alt");
-            $src = " ".($e->hasAttribute("src") ? $this->normalizeUrl($e->getAttribute("src")) : "")." ";
-            if ($prefix === "e") {
-                $attr = strlen($alt) ? $alt : $src;
-            } else {
-                $attr = strlen($alt) ? $alt : "";
+        #   relative; [p- and e- only]
+        if (in_array($prefix, ["p", "e"])) {
+            foreach ($copy->getElementsByTagName("img") as $e) {
+                $attr = null;
+                if ($e->hasAttribute("alt")) {
+                    $attr = $alt = $e->getAttribute("alt");
+                } elseif ($e->hasAttribute("src")) {
+                    $attr = " ".($e->hasAttribute("src") ? $this->normalizeUrl($e->getAttribute("src")) : "")." ";
+                }
+                if ($attr !== null) {
+                    $e->parentNode->replaceChild($e->ownerDocument->createTextNode($attr), $e);
+                }
             }
-            $e->parentNode->replaceChild($e->ownerDocument->createTextNode($attr), $e);
         }
         # removing all leading/trailing spaces
         return $this->trim($copy->textContent);
diff --git a/tests/cases/StandardTest.php b/tests/cases/StandardTest.php
index 1aca503..6a4a39c 100644
--- a/tests/cases/StandardTest.php
+++ b/tests/cases/StandardTest.php
@@ -36,14 +36,16 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
         // read expectation data
         $exp = json_decode(file_get_contents($path.".json"), true);
         if ($exp) {
+            $pattern = '#(\bhttps?://[^/ "\']+)($|[ "\'])#';
+            $replace = "$1/$2";
             // fix up expectation where necessary
-            array_walk_recursive($exp, function(&$v) {
+            array_walk_recursive($exp, function(&$v) use ($pattern, $replace) {
                 // URLs differ trivially from output of our normalization library
-                $v = preg_replace('#^https?://[^/]+$#', "$0/", $v);
+                $v = preg_replace($pattern, $replace, $v);
             });
             // URLs also need fixing as keys in rel-urls
             foreach ($exp['rel-urls'] as $k => $v) {
-                $fixed = preg_replace('#^https?://[^/]+$#', "$0/", $k);
+                $fixed = preg_replace($pattern, $replace, $k);
                 $exp['rel-urls'][$fixed] = $v;
                 if ($fixed !== $k) {
                     unset($exp['rel-urls'][$k]);