Only mark link as broken when it matches the configured status (#45)

* Only mark link as broken when it matches the configured status * Fix styling --------- Co-authored-by: Baspa <[email protected]>
vormkracht10 · Sep 27, 2023 · 0a1437f · 0a1437f
1 parent b4e29ac
commit 0a1437f
Show file tree

Hide file tree

Showing 4 changed files with 32 additions and 8 deletions.
diff --git a/config/seo.php b/config/seo.php
@@ -178,4 +178,15 @@
  |
  */
  'javascript' => false,
+
+ /*
+ |--------------------------------------------------------------------------
+ | Check specific options
+ |--------------------------------------------------------------------------
+ |
+ */
+ 'broken_link_check' => [
+ // Add status codes that should be considered as broken links. Empty array means all status codes starting with a 4, 5 or 0.
+ 'status_codes' => [],
+ ],
 ];
diff --git a/src/Checks/Content/BrokenLinkCheck.php b/src/Checks/Content/BrokenLinkCheck.php
@@ -60,13 +60,7 @@ public function validateContent(Crawler $crawler): bool
  return $link;
  })
  ->filter(function ($link) {
- $statusCode = (string) getRemoteStatus($link);
-
- if (str_starts_with($statusCode, '4') || str_starts_with($statusCode, '5') || $statusCode === '0') {
- return $link;
- }
-
- return false;
+ return isBrokenLink($link) ? $link : false;
  })->map(function ($link) {
  return [
  'url' => $link,

diff --git a/src/helpers.php b/src/helpers.php
@@ -10,7 +10,11 @@ function isBrokenLink(string $url): bool
  {
  $statusCode = (string) getRemoteStatus($url);
 
- if (str_starts_with($statusCode, '4') || str_starts_with($statusCode, '5') || $statusCode === '0') {
+ if (! empty(config('seo.broken_link_check.status_codes'))) {
+ return in_array($statusCode, config('seo.broken_link_check.status_codes'));
+ }
+
+ if (str_starts_with($statusCode, '4') || str_starts_with($statusCode, '5') || $statusCode == '0') {
  return true;
  }
 

diff --git a/tests/Checks/Content/BrokenLinkCheckTest.php b/tests/Checks/Content/BrokenLinkCheckTest.php
@@ -91,3 +91,18 @@
 
  $this->assertTrue($check->check(Http::get('vormkracht10.nl'), $crawler));
 });
+
+it('can check if link is broken by checking on configured status codes', function () {
+ $check = new BrokenLinkCheck();
+ $crawler = new Crawler();
+
+ config(['seo.broken_link_check.status_codes' => ['403']]);
+
+ Http::fake([
+ 'vormkracht10.nl' => Http::response('<html><head></head><body><a href="https://vormkracht10.nl/404">Vormkracht10</a></body></html>', 200),
+ ]);
+
+ $crawler->addHtmlContent(Http::get('vormkracht10.nl')->body());
+
+ $this->assertTrue($check->check(Http::get('vormkracht10.nl/admin/dashboard'), $crawler));
+});