CheckExternalLinksTask::runLinksCheck() - Code Metrics - silverstripe/silverstripe-externallinks - Measure and Improve Code Quality continuously with Scrutinizer

CheckExternalLinksTask::runLinksCheck() B
last analyzed 2019-11-05 01:30 UTC

↳ Parent: CheckExternalLinksTask

Complexity

Conditions	7
Paths	16

Size

Total Lines	58
Code Lines	33

Duplication

Lines	0
Ratio	0 %

Importance

Changes	2
Bugs	0	Features	0

Metric	Value
cc	7
eloc	33
c	2
b	0
f	0
nc	16
nop	1
dl	0
loc	58
rs	8.4586

How to fix Long Method

<?php

namespace SilverStripe\ExternalLinks\Tasks;

use DOMNode;
use SilverStripe\CMS\Model\SiteTree;
use SilverStripe\Core\Config\Config;
use SilverStripe\Core\Injector\Injector;
use SilverStripe\Dev\BuildTask;
use SilverStripe\Dev\Debug;
use SilverStripe\ExternalLinks\Model\BrokenExternalLink;
use SilverStripe\ExternalLinks\Model\BrokenExternalPageTrack;
use SilverStripe\ExternalLinks\Model\BrokenExternalPageTrackStatus;
use SilverStripe\ExternalLinks\Tasks\LinkChecker;
use SilverStripe\ORM\DataObject;
use SilverStripe\ORM\DB;
use SilverStripe\ORM\ValidationException;

class CheckExternalLinksTask extends BuildTask
{
    private static $dependencies = [

        'LinkChecker' => '%$' . LinkChecker::class
    ];

    private static $segment = 'CheckExternalLinksTask';


    /**
     * Define a list of HTTP response codes that should not be treated as "broken", where they usually
     * might be.
     *
     * @config
     * @var array
     */
    private static $ignore_codes = [];


    /**
     * @var bool
     */
    protected $silent = false;

    /**
     * @var LinkChecker
     */
    protected $linkChecker;

    protected $title = 'Checking broken External links in the SiteTree';

    protected $description = 'A task that records external broken links in the SiteTree';

    protected $enabled = true;

    /**
     * Log a message
     *
     * @param string $message
     */
    protected function log($message)
    {
        if (!$this->silent) {
            Debug::message($message);
        }
    }

    public function run($request)
    {
        $this->runLinksCheck();
    }
    /**
     * Turn on or off message output
     *
     * @param bool $silent
     */
    public function setSilent($silent)
    {
        $this->silent = $silent;
    }

    /**
     * @param LinkChecker $linkChecker
     */
    public function setLinkChecker(LinkChecker $linkChecker)
    {
        $this->linkChecker = $linkChecker;
    }

    /**
     * @return LinkChecker
     */
    public function getLinkChecker()
    {
        return $this->linkChecker;
    }

    /**
     * Check the status of a single link on a page
     *
     * @param BrokenExternalPageTrack $pageTrack
     * @param DOMNode $link
     */
    protected function checkPageLink(BrokenExternalPageTrack $pageTrack, DOMNode $link)
    {
        $class = $link->getAttribute('class');
        $href = $link->getAttribute('href');
        $markedBroken = preg_match('/\b(ss-broken)\b/', $class);

        // Check link
        $httpCode = $this->linkChecker->checkLink($href);
        if ($httpCode === null) {

            return; // Null link means uncheckable, such as an internal link
        }

        // If this code is broken then mark as such
        if ($foundBroken = $this->isCodeBroken($httpCode)) {
            // Create broken record
            $brokenLink = new BrokenExternalLink();
            $brokenLink->Link = $href;
            $brokenLink->HTTPCode = $httpCode;
            $brokenLink->TrackID = $pageTrack->ID;

            $brokenLink->StatusID = $pageTrack->StatusID; // Slight denormalisation here for performance reasons

            $brokenLink->write();
        }

        // Check if we need to update CSS class, otherwise return
        if ($markedBroken == $foundBroken) {
            return;
        }
        if ($foundBroken) {
            $class .= ' ss-broken';
        } else {
            $class = preg_replace('/\s*\b(ss-broken)\b\s*/', ' ', $class);
        }
        $link->setAttribute('class', trim($class));
    }

    /**
     * Determine if the given HTTP code is "broken"
     *
     * @param int $httpCode
     * @return bool True if this is a broken code
     */
    protected function isCodeBroken($httpCode)
    {
        // Null represents no request attempted
        if ($httpCode === null) {

            return false;
        }

        // do we have any whitelisted codes
        $ignoreCodes = $this->config()->get('ignore_codes');
        if (is_array($ignoreCodes) && in_array($httpCode, $ignoreCodes)) {
            return false;
        }

        // Check if code is outside valid range
        return $httpCode < 200 || $httpCode > 302;
    }

    /**
     * Runs the links checker and returns the track used
     *
     * @param int $limit Limit to number of pages to run, or null to run all
     * @return BrokenExternalPageTrackStatus
     */
    public function runLinksCheck($limit = null)
    {
        // Check the current status
        $status = BrokenExternalPageTrackStatus::get_or_create();

        // Calculate pages to run
        $pageTracks = $status->getIncompleteTracks();
        if ($limit) {
0   == false // true
0   == null  // true
123 == false // false
123 == null  // false

// It is often better to use strict comparison
0 === false // false
0 === null  // false
            $pageTracks = $pageTracks->limit($limit);
        }

        // Check each page
        foreach ($pageTracks as $pageTrack) {
            // Flag as complete
            $pageTrack->Processed = 1;
            $pageTrack->write();

            // Check value of html area
            $page = $pageTrack->Page();
            $this->log("Checking {$page->Title}");
            $htmlValue = Injector::inst()->create('HTMLValue', $page->Content);
            if (!$htmlValue->isValid()) {
                continue;
            }

            // Check each link
            $links = $htmlValue->getElementsByTagName('a');
            foreach ($links as $link) {
                $this->checkPageLink($pageTrack, $link);
            }

            // Update content of page based on link fixes / breakages
            $htmlValue->saveHTML();
            $page->Content = $htmlValue->getContent();
            try {
                $page->write();
            } catch (ValidationException $ex) {
                $this->log("Exception caught for {$page->Title}, skipping. Message: " . $ex->getMessage());
                continue;
            }

            // Once all links have been created for this page update HasBrokenLinks
            $count = $pageTrack->BrokenLinks()->count();
            $this->log("Found {$count} broken links");
            if ($count) {
                $siteTreeTable = DataObject::getSchema()->tableName(SiteTree::class);
                // Bypass the ORM as syncLinkTracking does not allow you to update HasBrokenLink to true
                DB::query(sprintf(
                    'UPDATE "%s" SET "HasBrokenLink" = 1 WHERE "ID" = \'%d\'',
                    $siteTreeTable,
                    intval($pageTrack->ID)
                ));
            }
        }

        $status->updateJobInfo('Updating completed pages');
        $status->updateStatus();
        return $status;
    }

    private function updateCompletedPages($trackID = 0)

    {
        $noPages = BrokenExternalPageTrack::get()
            ->filter(array(
                'TrackID' => $trackID,
                'Processed' => 1
            ))
            ->count();
        $track = BrokenExternalPageTrackStatus::get_latest();
        $track->CompletedPages = $noPages;
        $track->write();
        return $noPages;
    }

    private function updateJobInfo($message)
    {
        $track = BrokenExternalPageTrackStatus::get_latest();
        if ($track) {

            $track->JobInfo = $message;

            $track->write();
        }
    }
}


1			<?php
2
3			namespace SilverStripe\ExternalLinks\Tasks;
4
5			use DOMNode;
6			use SilverStripe\CMS\Model\SiteTree;
7			use SilverStripe\Core\Config\Config;
8			use SilverStripe\Core\Injector\Injector;
9			use SilverStripe\Dev\BuildTask;
10			use SilverStripe\Dev\Debug;
11			use SilverStripe\ExternalLinks\Model\BrokenExternalLink;
12			use SilverStripe\ExternalLinks\Model\BrokenExternalPageTrack;
13			use SilverStripe\ExternalLinks\Model\BrokenExternalPageTrackStatus;
14			use SilverStripe\ExternalLinks\Tasks\LinkChecker;
15			use SilverStripe\ORM\DataObject;
16			use SilverStripe\ORM\DB;
17			use SilverStripe\ORM\ValidationException;
18
19			class CheckExternalLinksTask extends BuildTask
20			{
21			private static $dependencies = [
			0 ignored issues – show introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The private property `$dependencies` is not used, and could be removed. Loading history...
22			'LinkChecker' => '%$' . LinkChecker::class
23			];
24
25			private static $segment = 'CheckExternalLinksTask';
			0 ignored issues – show introduced 2017-11-27 20:48 UTC by Report Bug Copy Issue Report The private property `$segment` is not used, and could be removed. Loading history...
26
27			/**
28			* Define a list of HTTP response codes that should not be treated as "broken", where they usually
29			* might be.
30			*
31			* @config
32			* @var array
33			*/
34			private static $ignore_codes = [];
			0 ignored issues – show introduced 2017-11-27 20:48 UTC by Report Bug Copy Issue Report The private property `$ignore_codes` is not used, and could be removed. Loading history...
35
36			/**
37			* @var bool
38			*/
39			protected $silent = false;
40
41			/**
42			* @var LinkChecker
43			*/
44			protected $linkChecker;
45
46			protected $title = 'Checking broken External links in the SiteTree';
47
48			protected $description = 'A task that records external broken links in the SiteTree';
49
50			protected $enabled = true;
51
52			/**
53			* Log a message
54			*
55			* @param string $message
56			*/
57			protected function log($message)
58			{
59			if (!$this->silent) {
60			Debug::message($message);
61			}
62			}
63
64			public function run($request)
65			{
66			$this->runLinksCheck();
67			}
68			/**
69			* Turn on or off message output
70			*
71			* @param bool $silent
72			*/
73			public function setSilent($silent)
74			{
75			$this->silent = $silent;
76			}
77
78			/**
79			* @param LinkChecker $linkChecker
80			*/
81			public function setLinkChecker(LinkChecker $linkChecker)
82			{
83			$this->linkChecker = $linkChecker;
84			}
85
86			/**
87			* @return LinkChecker
88			*/
89			public function getLinkChecker()
90			{
91			return $this->linkChecker;
92			}
93
94			/**
95			* Check the status of a single link on a page
96			*
97			* @param BrokenExternalPageTrack $pageTrack
98			* @param DOMNode $link
99			*/
100			protected function checkPageLink(BrokenExternalPageTrack $pageTrack, DOMNode $link)
101			{
102			$class = $link->getAttribute('class');
103			$href = $link->getAttribute('href');
104			$markedBroken = preg_match('/\b(ss-broken)\b/', $class);
105
106			// Check link
107			$httpCode = $this->linkChecker->checkLink($href);
108			if ($httpCode === null) {
			0 ignored issues – show introduced 2019-03-11 02:44 UTC by Report Bug Copy Issue Report The condition `$httpCode === null` is always `false`. Loading history...
109			return; // Null link means uncheckable, such as an internal link
110			}
111
112			// If this code is broken then mark as such
113			if ($foundBroken = $this->isCodeBroken($httpCode)) {
114			// Create broken record
115			$brokenLink = new BrokenExternalLink();
116			$brokenLink->Link = $href;
117			$brokenLink->HTTPCode = $httpCode;
118			$brokenLink->TrackID = $pageTrack->ID;
			0 ignored issues – show Bug Best Practice introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The property `TrackID` does not exist on `SilverStripe\ExternalLin...odel\BrokenExternalLink`. Since you implemented `__set`, consider adding a @property annotation. Loading history...
119			$brokenLink->StatusID = $pageTrack->StatusID; // Slight denormalisation here for performance reasons
			0 ignored issues – show Bug Best Practice introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The property `StatusID` does not exist on `SilverStripe\ExternalLin...odel\BrokenExternalLink`. Since you implemented `__set`, consider adding a @property annotation. Loading history... Bug Best Practice introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The property `StatusID` does not exist on `SilverStripe\ExternalLin...BrokenExternalPageTrack`. Since you implemented `__get`, consider adding a @property annotation. Loading history...
120			$brokenLink->write();
121			}
122
123			// Check if we need to update CSS class, otherwise return
124			if ($markedBroken == $foundBroken) {
125			return;
126			}
127			if ($foundBroken) {
128			$class .= ' ss-broken';
129			} else {
130			$class = preg_replace('/\s\b(ss-broken)\b\s/', ' ', $class);
131			}
132			$link->setAttribute('class', trim($class));
133			}
134
135			/**
136			* Determine if the given HTTP code is "broken"
137			*
138			* @param int $httpCode
139			* @return bool True if this is a broken code
140			*/
141			protected function isCodeBroken($httpCode)
142			{
143			// Null represents no request attempted
144			if ($httpCode === null) {
			0 ignored issues – show introduced 2019-03-11 02:44 UTC by Report Bug Copy Issue Report The condition `$httpCode === null` is always `false`. Loading history...
145			return false;
146			}
147
148			// do we have any whitelisted codes
149			$ignoreCodes = $this->config()->get('ignore_codes');
150			if (is_array($ignoreCodes) && in_array($httpCode, $ignoreCodes)) {
151			return false;
152			}
153
154			// Check if code is outside valid range
155			return $httpCode < 200 \|\| $httpCode > 302;
156			}
157
158			/**
159			* Runs the links checker and returns the track used
160			*
161			* @param int $limit Limit to number of pages to run, or null to run all
162			* @return BrokenExternalPageTrackStatus
163			*/
164			public function runLinksCheck($limit = null)
165			{
166			// Check the current status
167			$status = BrokenExternalPageTrackStatus::get_or_create();
168
169			// Calculate pages to run
170			$pageTracks = $status->getIncompleteTracks();
171			if ($limit) {
			0 ignored issues – show Bug Best Practice introduced 2019-03-11 02:44 UTC by Report Bug Copy Issue Report The expression `$limit` of type `integer\|null` is loosely compared to `true`; this is ambiguous if the integer can be 0. You might want to explicitly use `!== null` instead. In PHP, under loose comparison (like `==`, or `!=`, or `switch` conditions), values of different types might be equal. For `integer` values, zero is a special case, in particular the following results might be unexpected: 0 == false // true 0 == null // true 123 == false // false 123 == null // false // It is often better to use strict comparison 0 === false // false 0 === null // false Loading history...
172			$pageTracks = $pageTracks->limit($limit);
173			}
174
175			// Check each page
176			foreach ($pageTracks as $pageTrack) {
177			// Flag as complete
178			$pageTrack->Processed = 1;
179			$pageTrack->write();
180
181			// Check value of html area
182			$page = $pageTrack->Page();
183			$this->log("Checking {$page->Title}");
184			$htmlValue = Injector::inst()->create('HTMLValue', $page->Content);
185			if (!$htmlValue->isValid()) {
186			continue;
187			}
188
189			// Check each link
190			$links = $htmlValue->getElementsByTagName('a');
191			foreach ($links as $link) {
192			$this->checkPageLink($pageTrack, $link);
193			}
194
195			// Update content of page based on link fixes / breakages
196			$htmlValue->saveHTML();
197			$page->Content = $htmlValue->getContent();
198			try {
199			$page->write();
200			} catch (ValidationException $ex) {
201			$this->log("Exception caught for {$page->Title}, skipping. Message: " . $ex->getMessage());
202			continue;
203			}
204
205			// Once all links have been created for this page update HasBrokenLinks
206			$count = $pageTrack->BrokenLinks()->count();
207			$this->log("Found {$count} broken links");
208			if ($count) {
209			$siteTreeTable = DataObject::getSchema()->tableName(SiteTree::class);
210			// Bypass the ORM as syncLinkTracking does not allow you to update HasBrokenLink to true
211			DB::query(sprintf(
212			'UPDATE "%s" SET "HasBrokenLink" = 1 WHERE "ID" = \'%d\'',
213			$siteTreeTable,
214			intval($pageTrack->ID)
215			));
216			}
217			}
218
219			$status->updateJobInfo('Updating completed pages');
220			$status->updateStatus();
221			return $status;
222			}
223
224			private function updateCompletedPages($trackID = 0)
			0 ignored issues – show Unused Code introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The method `updateCompletedPages()` is not used, and could be removed. This check looks for private methods that have been defined, but are not used inside the class. Loading history...
225			{
226			$noPages = BrokenExternalPageTrack::get()
227			->filter(array(
228			'TrackID' => $trackID,
229			'Processed' => 1
230			))
231			->count();
232			$track = BrokenExternalPageTrackStatus::get_latest();
233			$track->CompletedPages = $noPages;
234			$track->write();
235			return $noPages;
236			}
237
238			private function updateJobInfo($message)
239			{
240			$track = BrokenExternalPageTrackStatus::get_latest();
241			if ($track) {
			0 ignored issues – show introduced 2019-03-11 02:44 UTC by Report Bug Copy Issue Report `$track` is of type `SilverStripe\ExternalLin...ExternalPageTrackStatus`, thus it always evaluated to `true`. Loading history...
242			$track->JobInfo = $message;
			0 ignored issues – show Bug Best Practice introduced 2017-11-23 02:03 UTC by Report Bug Copy Issue Report The property `JobInfo` does not exist. Although not strictly required by PHP, it is generally a best practice to declare properties explicitly. Loading history...
243			$track->write();
244			}
245			}
246			}
247

silverstripe / silverstripe-externallinks

CheckExternalLinksTask::runLinksCheck() B last analyzed 2019-11-05 01:30 UTC

Complexity

Size

Duplication

Importance

How to fix Long Method

Long Method

Duplication Side-by-Side

Filter issues like

CheckExternalLinksTask::runLinksCheck() B
last analyzed 2019-11-05 01:30 UTC