I could not let it rest, so I googled somewhat and seemed to have found at least one interesting files and file contents indexer:
docfetcher
Its open source, has an active development, java based and even a portable version is available. Best of all: it knows EPUB's. First impressions are not bad, reasonable fast, and it can find strings in epubs.
Perhaps worth to try.
Now, if only the file contents indexer could be incorporated in calibre (plug-in?), the match would be perfect
One likes to dream
