[TYPO3-core] Problem regarding crawler, indexed_search and fe_groups access to search results

[Mathias Schreiber]

Hi Luc,

basically this part of the indexing thing is what can be considered "legacy" - like in "the bad stuff".

It is indeed true that right now you need to add all usergroup combinations (!) to the crawler in order to make them searchable.
This is due to the fact that normally (when not using the crawler) TYPO3 would index the current page based on what usergroup combination you'd have.

Background:
Imagine you have a page with 3 content elements on them.
One is provided to usergroup 20, one to usergroup 10 and the other one is free for all.
The TYPO3 indexed_search will always index the current state of a page for that current usergroup combination.
Because within the page a lot more than the example above can go on (like an extension providing record based on weird ACL rules and such) this is like indexed_search does things.

This works out totally fine if not using crawler, because the users with said usergroup combination will just visit the page and the content will get indexed then.
But when using crawler there is no way to build these usergroup combinations except doing it manually.

One word of advise though... 
Be careful with the amount of usergroup combinations you provide.
These will add up to the amount of data in your index_* tables, thus making them grow and inevitably slowing down the search speed.
I have a site running with about 450 different usergroup combinations and roughly 3GB of index_* tables - searching is a pain which made us rebuild the entire search SQL to fit our needs (but also remove about 80% of indexed_search's functionality in the process :)).

I hope this helped you get a better understanding of the idea *why* indexed_search works like it does now.
cheers
Mathias