parser/serializer for files.xml

[rvanheest]

Write a parser and serializer for files.xml <-> FilesXml.
Don't expose them via the FilesXml (maybe put them in an internal package, such that they're still accessible if needed), but use them in functions like read (parse xml to object) and save (serialize object to xml).
Make sure that files.xml -> FilesXml -> files.xml results in the same xml as with which you started.

[jo-pol]

Does this imply a refactoring of FilesXml in the deposit-api? It currently serializes files in a folder to a scala.xml.Elem written to files.xml by DepositDir.createXMLs.

[rvanheest]

@jo-pol probably after this milestone has been completed, your FilesXml has to be rewritten indeed. As described in #16, files will automatically be added to files.xml when they're added to the bag (and when .save() is called!).

Before .save() is called, they are only stored in the object model for FilesXml and on .save() they're being serialized into XML. On .read() they are parsed again from files.xml to the FilesXml object model.

[rvanheest]

Points for discussion:

• does the parser (files.xml -> FilesXml) have to return an error if files.xml is invalid according to the schema?
• when reading the files.xml in the context of a bag (so, not using FilesXml.read() directly, but via DansBag.read()), should the bag also validate all its payload files against the files listed in files.xml?
  • Note: we don't do this yet with manifest-sha1.txt, fetch.txt or tagmanifest-sha1.txt!