Add `failIfExtra` feature to `CharReaderBuilder`.

1. failing regression tests, from #164 and #107
2. implemented; tests pass
3. allow trailing comments

revert '642befc8',
but keep the *added* methods for `decodedNumber()` and `decodedDouble()`.

re: 28836b8a

A global instance of a Value (viz. 'null') was a mistake,
but dropping it breaks binary-compatibility. So we will keep it
everywhere except the one platform where it was crashing, ARM.

revert ae3c7a7a

This reverts commit 5bf16105.

revert 68db6553
issue #147

for binary-compatibility with 0.6.0
issue #147
was #57

This reverts commit 45cd9490.

Ignored ValueInternal* changes, since those did not produce symbols for
Debian build. (They must not have used the INTERNAL stuff.)

  https://github.com/open-source-parsers/jsoncpp/issues/78

Conflicts:
	include/json/value.h
	src/lib_json/json_internalarray.inl
	src/lib_json/json_internalmap.inl
	src/lib_json/json_value.cpp

By not calling validate(), we can add
non-invasive features which will be simply ignored when user-code
is compiled against an old version. That way, we can often
avoid a minor version-bump.

The user can call validate() himself if he prefers that behavior.

Fixes #88 and #56.

This is an improper solution. If multiple Readers exist,
then the effect stackLimit is reduced because of side-effects.
But our options are limited. We need to address the security
hole without breaking binary-compatibility.

However, this is not likely to cause any practical problems because:

* Anyone using `operator>>(istream, Json::Value)` will be using the
new code already
* Multiple Readers are uncommon.
* The stackLimit is quite high.
* Deeply nested JSON probably would have hit the system limits anyway.

And throw instead of return null for invalid settings.

Remove some experimental classes that are not needed for 1.4.0. This also helps 0.8.0 binary compatibility with 0.6.0-rc2.

Reimplement the new Builders.

Issue #131.