Last modified: 2012-04-23 15:22:11 UTC
The XML dump files released by Wikimedia contain a <namespaces> section which declares namespace names and numbers for the wiki it was dumped from. But it does not tell you which of those namespaces are actually covered by the dump files. For instance *-*-pages-articles.xml dumps do not contain any "Talk", "* talk", or "User" entries. Not even page title and redirect information. This is fine but with wiki dumps now being produced in the same format also outside Wikimedia with different subsets of namespaces covered, such as http://devtionary.org/w/dump/xmlu/ the dump format is now an interchange format of sorts. So it would be nice if such information which is currently metadata external to the dump files could be made internal and self-contained. This could be quite useful to tools designed to process dump files. Perhaps a new section of the dump files named <dumpinfo> could be added to complement the <siteinfo> section.
Similar to bug 34218 and bug 31955.
Similar to bug 36178.