jhodges/sitemap
1
0
Fork 0
Code Issues Pull Requests Releases 5 Wiki Activity

fix tests

Browse Source
This commit is contained in:
James 2020-02-23 01:08:14 +00:00
parent ffdef4b0d5
commit 9c0961fd87
2 changed files with 75 additions and 74 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/CrawlObserver.php
View File

@ -124,7 +124,7 @@ class CrawlObserver extends \Spatie\Crawler\CrawlObserver
if( $response=$requestException->getResponse() ){
$this->crawled($url,$response,$foundOnUrl);
}else{
$this->addResult((String)$url,(string)$foundOnUrl);
$this->addResult((String)$url,(string)$foundOnUrl,'---','');
}
}

147
tests/CrawlerTest.php
View File

@ -5,123 +5,124 @@ use JHodges\Sitemap\Crawler;
class CrawlerTest extends TestCase{
public function testCanCrawlSite(){
$crawler=new Crawler('http://localhost:8080');
$crawler->crawl('http://localhost:8080');
public function testFound(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/found');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/' => ['code' => 200 ],
'http://localhost:8080/link1' => ['code' => 200 ],
'http://localhost:8080/link2' => ['code' => 200 ],
'http://localhost:8080/link3' => ['code' => 200 ],
'http://localhost:8080/link4' => ['code' => 302 ],
'http://localhost:8080/notExists' => ['code' => 404 ],
'http://localhost:8080/found' => ['code' => 200],
], print_r($sitemap,true));
}
public function testNotFound(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/notFound');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/notFound' => ['code' => 404],
], print_r($sitemap,true));
}
public function testExternalLink(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/externalLink');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/externalLink' => ['code' => 200],
'http://example.com/' => ['code' => 200],
], print_r($sitemap,true));
}
public function testDeeplink(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/deeplink1');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/deeplink1' => ['code' => 200],
'http://localhost:8080/deeplink2' => ['code' => 200],
'http://localhost:8080/deeplink3' => ['code' => 200],
], print_r($sitemap,true));
}
public function testInterlinked(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/interlinked1');
//$crawler->crawl('http://localhost:8080/page4'); // TODO!!! this ensures the order or results for the URL tracking test 3PARTS.
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/interlinked1' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/interlinked1' => 1,
'http://localhost:8080/interlinked2' => 1,
'http://localhost:8080/interlinked3' => 1,
]],
'http://localhost:8080/interlinked2' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/interlinked1' => 1,
'http://localhost:8080/interlinked2' => 1,
'http://localhost:8080/interlinked3' => 1,
]],
'http://localhost:8080/interlinked3' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/interlinked1' => 1,
'http://localhost:8080/interlinked2' => 1,
'http://localhost:8080/interlinked3' => 1,
]],
]);
}
public function testCollectsAllFoundOnUrls(){
$crawler=new Crawler('http://localhost:8080');
$crawler->crawl('http://localhost:8080/page1');
$crawler->crawl('http://localhost:8080/page4'); // this ensures the order or results for the URL tracking test 3PARTS.
$sitemap=$crawler->getResults();
print_r($sitemap);
$this->assertTreeContains($sitemap,[
'http://localhost:8080/page1' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
]],
'http://localhost:8080/page2' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
]],
'http://localhost:8080/page3' => ['code' => 200 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
]],
'http://localhost:8080/notFound1' => ['code' => 404 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
]],
'http://localhost:8080/notFound2' => ['code' => 404 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
'http://localhost:8080/page4' => 1,
]],
'http://localhost:8080/redirectToNotFound' => ['code' => 302 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
'http://localhost:8080/page4' => 1,
]],
'http://localhost:8080/redirectToRedirectToNotFound' => ['code' => 302 , 'foundOn' => [
'http://localhost:8080/page1' => 1,
'http://localhost:8080/page2' => 1,
'http://localhost:8080/page3' => 1,
'http://localhost:8080/page4' => 1,
]],
]);
}
public function testCanFollowRedirectToFound(){
$crawler=new Crawler('http://localhost:8080');
public function testRedirectToFound(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/redirectToFound');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/redirectToFound' => ['code' => 302 ],
'http://localhost:8080/' => ['code' => 200 ],
'http://localhost:8080/found' => ['code' => 200 ],
]);
}
public function testCanFollowRedirectToNotFound(){
$crawler=new Crawler('http://localhost:8080');
public function testRedirectToNotFound(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/redirectToNotFound');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/redirectToNotFound' => ['code' => 302 ],
'http://localhost:8080/notFound2' => ['code' => 404 ],
'http://localhost:8080/notFound' => ['code' => 404 ],
]);
}
public function testCanFollowRedirectToRedirectToNotFound(){
$crawler=new Crawler('http://localhost:8080');
public function testRedirectToRedirectToNotFound(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/redirectToRedirectToNotFound');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/redirectToRedirectToNotFound' => ['code' => 302 ],
'http://localhost:8080/redirectToNotFound' => ['code' => 302 ],
'http://localhost:8080/notFound2' => ['code' => 404 ],
'http://localhost:8080/notFound' => ['code' => 404 ],
]);
}
public function testCanFollowTwoRedirectsToSameLocation(){
$crawler=new Crawler('http://localhost:8080');
public function testTwoRedirectsToSameLocation(){
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/twoRedirectsToSameLocation');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/twoRedirectsToSameLocation' => ['code' => 200 ],
'http://localhost:8080/redirect1' => ['code' => 302 ],
'http://localhost:8080/redirect2' => ['code' => 302 ],
'http://localhost:8080/link1' => ['code' => 200 ],
'http://localhost:8080/found' => ['code' => 200 ],
]);
}
public function testTimeout(){
$crawler=new Crawler('http://localhost:8080');
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/timeout');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[
'http://localhost:8080/timeout' => ['code' => '' ],
'http://localhost:8080/timeout' => ['code' => '---' ],
]);
}
public function testInternalServerError(){
$crawler=new Crawler('http://localhost:8080');
$crawler=new Crawler();
$crawler->crawl('http://localhost:8080/internalServerError');
$sitemap=$crawler->getResults();
$this->assertTreeContains($sitemap,[