Appendix B. Format of the revision history

Menu item Tools → Revisions → Store All Revisions in the Document in XMLmind XML Editor - Online Help (and also this option in XMLmind XML Editor - Online Help) allows to store the entire revision history of a document in the XML file containing this document.

The revision history is stored in a single <?xxe-revisions> processing-instruction found at the end of the document. Short example:

  ambiguë d&#8217;un c&#339;ur qui au zéphyr préfère les jattes de
  kiwi</foreignphrase>".</para>
</section>
<?xxe-revisions
#3 2015-09-04T14:35:48Z charles
#2 2015-09-04T14:29:20Z bart
1sPEAAABiXgAEocmAAAHBXOHExMIQwsAhRaEVw==

#1 2015-09-04T14:26:28Z john
1sPEAAABhyYAgQmHJQA/LBhmIjxmb3JlaWduOj0iZnJlPz5Qb3J0ZXogY2UgdmlldXggd2hpc2t5
IGF1IGp1Z2UgYmxvbmQgZnVtZTwvPiJzhAwTB0OBKlMx3zMGEwqTBxMEAQljBvczBAEEUyQBJzME
AQcjDgECkwpjCwCEPBREMoEthAI8g3QhFg+CIoETcJSCBw==
?>

The format of the <?xxe-revisions> processing-instruction is described as follows:

This processing-instruction contains one or more revisions separated by newlines.
A revision consists in a header followed by a binary delta encoded in base 64.
A header, for example:
```
#1 2015-09-04T14:26:28Z john
```
contains the revision number (1), the date this revision was saved to disk (2015-09-04T14:26:28Z) and the author of this revision (john).
First header is about the most recent revision. Hence it is not followed by a binary delta.
For example, let's suppose the above snippet comes from document doc.xml. Most recent revision is #3 and corresponds to the contents of doc.xml. It has been created by charles on 2015-09-04T14:35:48Z.
The binary delta of revision #N is the result of applying the VCDIFF encoding algorithm between a normalized XML form of the content of revision #N+1 and a normalized XML form of the content of revision #N.
For example, the normalized XML form of doc.xml is simply the UTF-8 bytes of doc.xml saved unindented (of course, without the <?xxe-revisions> processing-instruction). The result of applying the VCDIFF decoding algorithm to the UTF-8 bytes of doc.xml, that is, revision #3, and the binary delta "1sPEAAABiXgAEocmAAAHBXOHExMIQwsAhRaEVw==" results in the UTF-8 bytes of revision #2 of doc.xml.


3. A simple XML differencing algorithm