roxmltree Add support for custom entity definitions in ParsingOptions

I’m using the roxmltree library version 0.18.0 to parse some epub, but I encountered a problem with entity references. Some of the files do not have standard <!ENTITY ...> settings, such as the common ones

.add_entity("nbsp", " ")
.add_entity("copy", "©")
.add_entity("reg", "®")

This causes the parsing to fail. I think it would be more convenient and flexible if the library could provide a way to inject additional entity definitions in the ParsingOptions, instead of modifying the file content.

Is this possible to implement? Thank you for your consideration.

Sep 11 '23 10:09 nihilox

EPUB is HTML, not XML. And HTML parsing is out of scope.

Sep 11 '23 13:09 RazrFalcon

XHTML to be precise, at least for the task I am working on. It works well except for the entity definition part.

Sep 11 '23 16:09 nihilox

I'm not familiar with XHTML, but if it can be handled by generic XML parsers, then sure, we could add user-defined entities support.

I'm not sure what would be the correct way to implement this. Everything in XML is very complicated. Will take a look.

Sep 11 '23 17:09 RazrFalcon