Resource Dictionary and Inherited Page Attributes

Q: I’m trying to add a font object to each page’s resource dictionary. My current approach is to check for the resource dictionary inside the page, and if it is not present, create it:

PDF::Page page = doc.GetPage (1);

SDF::Obj dict = page.GetSDFObj ();

SDF::Obj resources = dict.FindObj (“Resources”);

if (!(bool)resources)

resources = dict.PutDict (“Resources”);

However, on certain documents, after adding the resource dictionary, ElementReader throws an exception: “Resource not found”. How is this possible, given that there was no resource dictionary in the page to start with?

Fortunately, PDFNet offers a simple solution to this problem. Rather than using:

SDF::Obj resources = dict.FindObj(“Resources”);

You should use:

SDF::Obj resources = dict. FindInheritedAttribute(“Resources”);

This will resolve page attribute inheritance, and return the correct dictionary for the page. This method should be used for all inheritable attributes, as specified in the PDF specification.
A: The confusion here stems from a corner of the PDF Specification, Inheritance of Page Attributes. According to the specification, “when an attribute is omitted from a page object, its value shall be inherited from an ancestor node in the page tree.” In other words, by adding an empty resource dictionary to the ‘Page’ dictionary, you are overriding the existing, inherited resource dictionary. As a result, the page loses the references to its existing resources, and the document is now corrupt.