diff options
| author | Andres Rey <[email protected]> | 2017-11-12 01:32:03 +0000 |
|---|---|---|
| committer | Andres Rey <[email protected]> | 2017-11-12 01:32:03 +0000 |
| commit | fcf8ba9de0a532433178686bde2f78afc6e063c2 (patch) | |
| tree | 4c550f7a85efe12d5b02316b3408c2b366777f8e /src/Readability.php | |
| parent | 00ab1e503a7d804a19365ffe3ee19cd9ca9f8643 (diff) | |
Add new regexp to check for whitespace include unicode version of
Diffstat (limited to 'src/Readability.php')
| -rw-r--r-- | src/Readability.php | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/Readability.php b/src/Readability.php index b5bc723..8d856d8 100644 --- a/src/Readability.php +++ b/src/Readability.php @@ -528,7 +528,8 @@ class Readability extends Element implements ReadabilityInterface public function isElementWithoutContent() { return ($this->node instanceof \DOMElement && - mb_strlen(trim($this->node->textContent)) === 0 && + // /\x{00A0}|\s+/u TODO to be replaced with regexps array + mb_strlen(preg_replace('/\x{00A0}|\s+/u','',$this->node->textContent)) === 0 && ($this->node->childNodes->length === 0 || $this->node->childNodes->length === $this->node->getElementsByTagName('br')->length + $this->node->getElementsByTagName('hr')->length || /* |