ComplexPhraseQueryParser with AsciiFoldingFilterFactory (SOLR) [LUCENE-7687]

I modified generic \*_txt-Field type to use AsciiFoldingFilterFactory on query & index.

When quering with

\{!complexphrase}text_txt:"König\*" – there are 0 results

\{!complexphrase}text_txt:"Konig\*" – there are >0 results

\{!complexphrase}text_txt:"König" – there are >0 results (but less than the line above)

and without \{!complexphrase} everything works o.k.

---
Migrated from [LUCENE-7687](https://issues.apache.org/jira/browse/LUCENE-7687) by Jochen Barth, updated Oct 12 2017
Environment:
```
solr-6.4.1 (yes, solr, but I don't know where the bug exactly is)
```

Linked issues:
- [SOLR-11450](https://issues.apache.org/jira/browse/SOLR-11450)

Start by reproducing the three complexphrase queries from the issue in Solr 6.4.1, then trace ComplexPhraseQueryParser together with AsciiFoldingFilterFactory to identify where accented wildcard terms diverge. Done means the accented and unaccented wildcard phrase queries return equivalent expected matches while preserving the existing non-complexphrase behavior.

Written by the indexing model from the issue text.