Consider dropping libxml2 usage

[dsd]

GetBooks uses libxml2-python, and this is an external library.

Could it be switched to using the xml classes that are included in Python? This way we could consider dropping the libxml2-python usage from the sugar platform and OLPC builds.

[godiard]

We need do the following tests:

* Navigate the catalog: Click on "public domain books",and select different categories, up and down. The list of books should be updated.

* When click in any book, the image and data in the bottom side of the screen should be updated.

* Download the book. Try different formats if available.

* Search a title or author.

* Change the source combo to "Internet Archive" and do a search.

* Download a book from Internet Archive.

[humitos]

All of this test that you mentioned here passed without any problem after applying this patch.

[humitos]

New version - v2

[godiard]

I can't see any change in your second patch.

I applied a version with the changes we talked (move _xmldoc and _eroot to get_full_language_name method, and make them method variables)