How to load DocBook XML files with PHP's DomDocument having named entities defined in other file?

Question

I'm trying to load XML files from the PHP repository of the online documentation. It's by the DocBook standard.

However, when I load an XML file I get errors like:

Warning: DOMDocument::load(): Entity 'reftitle.intro' not defined

I somewhat understand that error, but I'm having problems to fix it. I tried with the external entities loader callback (libxml_set_external_entity_loader) but it's only called once, for the file itself.

So somehow I think I need to tell DOMDocument either a) to ignore those undefined entities or b) to provide these entities.

There is a file named language-defs.ent which contains all these entities like inside a DTD, e.g.:

<!ENTITY reftitle.intro        '<title xmlns="http://docbook.org/ns/docbook">Introduction</title>'>

But I don't know how to make use of that. Maybe it's just the time of the day.

Answer 1

Roughly for the moment as I have no clue why, however (one might be already enough):

$fake = '<?xml version="1.0"?>
<!DOCTYPE book [
<!ENTITY % myent SYSTEM "file:///c:/path%20to/phpdoc-en-svn/extensions.ent">
%myent;
<!ENTITY % myent SYSTEM "file:///c:/path%20to/phpdoc-en-svn/language-defs.ent">
%myent;
<!ENTITY % myent SYSTEM "file:///c:/path%20to/phpdoc-en-svn/language-snippets.ent">
%myent;
]>';

and then (all XML $file come with a XML declaration that needs to be removed):

$doc = new DOMDocument();
$buffer = file_get_contents($file);
libxml_use_internal_errors(true);
$result = $doc->loadXML($fake.strstr($buffer, "\n"));

and it then works. It still has errors, hence the libxml_use_internal_errors, however the XML files are loading.

Not all entities are resolved.