Bug #78129 [Fbk->Asn]: DOMDocument removes whitespace between nodes

Edit report at https://bugs.php.net/bug.php?id=78129&edit=1 ID: 78129 User updated by: nospam at unclassified dot de Reported by: nospam at unclassified dot de Summary: DOMDocument removes whitespace between nodes -Status: Feedback +Status: Assigned Type: Bug Package: DOM XML related Operating System: Linux PHP Version: 7.3.6 Assigned To: cmb Block user comment: N Private report: N New Comment: I'm not sure which version I have. The server is running Ubuntu 16.04. There are two packages with that name, they have the versions "2.9.3+dfsg1-1ubuntu0.7" and "2.9.3+dfsg1-1". I can't try with another version at the moment. Will try to build up another environment for that with Ubuntu 20.04, I have nothing newer. Previous Comments: ------------------------------------------------------------------------ [2020-06-08 09:09:08] cmb@php.net I don't think this related to the OS, but rather to the libxml2 version. Which one do you use? Could you try with latest? ------------------------------------------------------------------------ [2020-06-07 15:33:25] nospam at unclassified dot de Is this issue tracker still in use and is anybody here? It feels a little lonely in here. ------------------------------------------------------------------------ [2019-11-02 19:20:57] nospam at unclassified dot de Can somebody please have a quick look at this? To me this looks like a simple fix, just edit that one line that sets the option for the Linux build. ------------------------------------------------------------------------ [2019-06-08 14:41:14] nospam at unclassified dot de It seems Linux always adds the LIBXML_NOBLANKS option whereas Windows does not. I can get the same result on Windows as on Linux when I use this code: $dom->loadHTML($html, LIBXML_NOBLANKS); But there is not option that inverts that option. To me it looks like the Linux build of PHP has the bug that the LIBXML_NOBLANKS option is always added while it should not. I couldn't find information about that option because the libxml documentation has no description of this option. It's just empty. ------------------------------------------------------------------------ [2019-06-08 14:31:05] nospam at unclassified dot de Description: ------------ The DOMDocument removes whitespace between HTML elements when loading and saving HTML content. This leads to wrong display of content. The bug only affects Linux, it works correctly in Windows (same version). Test script: --------------- <?php $html = '<html> <body> <p> Text </p> <lang>Logged in as</lang> <strong>user</strong> <script src="jquery-3.4.1.min.js"></script> <script src="site.js"></script> </body> </html>'; libxml_use_internal_errors(true); $dom = new DOMDocument(); $dom->loadHTML($html); header('Content-Type: text/plain'); echo $dom->saveHTML(); Expected result: ---------------- This is the output on Windows: <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd"> <html> <body> <p> Text </p> <lang>Logged in as</lang> <strong>user</strong> <script src="jquery-3.4.1.min.js"></script> <script src="site.js"></script> </body> </html> Actual result: -------------- This is the output on Linux: <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd"> <html><body> <p> Text </p> <lang>Logged in as</lang><strong>user</strong> <script src="jquery-3.4.1.min.js"></script><script src="site.js"></script></body></html> There is no space between "Logged in as" and "user". The <lang> element is a custom element that my PHP script will resolve into plain HTML. ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=78129&edit=1