Thanks for the full details -- being able to see exactly how the queries are recieved & parsed is important for rulling out simple things like client side escaping (or lack of) and server side metacharacter handling in the query parser.
: Some things work the way I'd expect, some clearly don't. So my question : was, in the first instance "Is there full regex support?" Clearly, : there's supposed to be, so something is wrong, or I don't know the right : escape syntax. I think it depends on your definition of "full" ? Based on the doc for the supported syntax, it doesn't look to me like there is any direct support for some any of the pre-defined character classes (ie: "\s", "\w", etc..) or boundary matchers (ie: "^", "\b", etc...) https://lucene.apache.org/core/4_4_0/core/org/apache/lucene/util/automaton/RegExp.html It looks like there are hooks in the underlying RegExp and RegexpQuery classes for registering named Automotons so the "<...>" syntax can refer to named character classes that are defined at runtime, but none are registered for you by default, and there is no way to configure that with the QueryParser. I'm sorry, but I don't understand the RegExp automoton stuff enough to understand why the predefined character classes from java.regex.Pattern aren't supported by default, or even how you could conceptually implement the boundary matchers using the underlying java API. I suspect the existing regex query support isn't going to work for what you're trying to do. -Hoss
