[Bug] Possible collision of index names

[moonshiner-dev]

When working with a Generic Data Index on one of our existing projects, I encountered a problem with Open Search index names created for the individual element types.

There are 3 types of elements -- asset, document and data object -- each with their corresponding ElementTypeAdapter. Every adapter then create an index name for a Element.

And here comes the catch. If you have a Data object class with the name document (or few other keywords) it will collide with the index name created by for example the DocumentTypeAdapter, which obtains the index name from the enum ElementType.

That results in this:

• Index name for Data objects class with name Document --> pimcore_document
• Index name for documents ––> pimcore_document

My approach to this would be to prefix the "dynamically generated" data object index names with some unique string, so the index name for data objects will be something like pimcore_dataobject_prefix_docuement instead of the current pimcore_document.

But I am unfortunately not 100% sure if that would work.