Confluence XML has an option to store details about the imported confluence pages: Store Confluence Details - storeConfluenceDetailsEnabled
It is disabled by default. The reason is likely that this hasn’t been generally useful, and probably useful only for niche reasons, so these objects would clog the wiki for no obvious reasons (I can see one though: having traces of a migration can be useful for various reasons - this is my archivist spirit speaking)
However, we now have a Confluence URL Mapping extension that allows administrators to setup redirection from old Confluence URLs to the imported content. It makes use of the ConfluencePageClass object.
I believe it would be useful to enable Store Confluence Details by default, so people importing Confluence content can then easily make use of the Confluence URL Mapping extension without having to think about enabling the option, or even know it is needed/useful.
I guess this is theoretically, technically a breaking change, hence me asking the community what its things about all this. I don’t think it will break anymore in practice: it will change things for new imports, people doing imports likely do them in one shot and don’t keep doing this ad vitam eternam, so there should not be observable behavior change for the vast majority of people.
I don’t personally need this change but I think it is a nice thing to bring.
I’d prefer to have the information present on the imported pages, too.
Another use case of the URL Mapping is with links in migrated diagrams; if these are linked to Confluence pages, the migrated diagram will keep these links as they are (as they are embedded in some XML content). If the information which XWiki page corresponds to that page is conserved, the links can be repaired e.g. on the first save after the import. Otherwise this might not be possible. I think that there are more cases where Confluence extensions have links in custom data structures that are not converted by the migration process.
