vmVersionRanges might be a single version like "9", a list of versions
like "1.2,1.3,1.4,1.5", an equivalent range of "1.2-1.5", an open range
like "-1.8", "9-" (equivalent to "9+") or a more complex list of ranges
like "-1.6,9-11,13-14,17-" or "8,11,16+".

Empty ranges like in "", " ",  "8,,14", ",5", "6-," will be ignored.
I.e., they will not yield a positive match.

Bogus ranges like "9-11-14" will be ignored, too.

Existing XML test specs using '<line text="..." vm="v,v2,...,vn"/>
have been adjusted to use version ranges rather than long lists of
consecutive versions. Furthermore, ranges with a trailing '+' like '14+'
were replaced by using the new canonical format '14-', even though the
parser still correctly recognises '14+'.

Signed-off-by: Alexander Kriegisch <Alexander@Kriegisch.name>

This reverts commit 8c15d83a46.

Signed-off-by: Alexander Kriegisch <Alexander@Kriegisch.name>

Now there should be no more inspection warnings when working with XML
test definitions. Only the strangely looking XML files used by
PureJavaTests and KnownLimitationsTests are left to be analysed and
fixed.

Signed-off-by: Alexander Kriegisch <Alexander@Kriegisch.name>

The version handling in LangUtil has been overhauled
to cope better with post 1.8 releases (JDK9 and JDK10 or 18.3
or whatever they call it). As part of this moved
to treating JDK9 as '9' rather than '1.9'. Also removed
duplicate version processing logic and had that defer to
the one place in LangUtil where we now deal with it.

Includes some generics tidyup in ajdoc. More ajdoc work
is necessary for Java10 because it removes the standard doclet
(old style). However trying to invoke the internal Javadoc
handler in Java10 is failing due to module visibility rules.

Modified test expectation system so it is possible to say
the test cares about one particular message and the rest
do not matter (prefix message string with '*') - crude but
quick.

Polished many places to exploit generics

Upgraded all the tests to work on Java8 - some serious changes
regarding ajdoc on Java8. Hopefully it has stayed backwards
compatible with earlier JDK versions (e.g. if using AspectJ 1.8.3+
with a JDK less than 8) but no explicit testing done for this.

Use resolved type pattern rather than re-resolving