Instead of shifting/xoring the hash at each field, use the built-in ruby
apis for generating a hash from multiple input values.

Now returns a Fixnum.

Ruby: make sure map parsing frames are GC-rooted.

For JSON encoding we provide a new option to decide at
encode time whether to use camelCase or original proto field
names:

  json = MapMessage.encode_json(m, :preserve_proto_fieldnames => true)

The flags are:

  UPB_JSON_ACCEPT_LEGACY_FIELD_NAMES
  UPB_JSON_WRITE_LEGACY_FIELD_NAMES

The first just allows the parser to accept the old field names.
The second makes the printer print the old field names.

These flags are intended to be temporary, as a migration aid
for users.

More informative and more ruby-like

Also updated the Gemfile.lock since alpha-4 has been
pushed to RubyGems.

Change-Id: I8ddc5f125f28aa9a33c88dfe48251a75a877e1d3

Change-Id: I6cf77f01370204ad4bc7b345a040a9a3de1706a0

Change-Id: Ief77de7134e05e07b1a7e3970d49880c2d5e6fe9

This involved fixing a few important bugs in the
Ruby implementation -- mostly cases of mixing
upb field types and descriptor types (upb field
types do not distinguish between int/sint/fixed/sfixed
like descriptor types do).

Also added protobuf-specific exceptions so parse
errors can be caught specifically.

Change-Id: Ib49d3db976900b2c6f3455c8b88af52cfb86e036

While we are C99 in general, the Ruby build system
for building C extensions enables several flags that
throw warnings for C89/C90 variable ordering rules.
To avoid spewing a million warnings (or trying to
specifically override these warnings with command-line
flags, which would be tricky and possibly fragile)
we conform to Ruby's world of C89/C90.

Change-Id: I0e03e62d95068dfdfde112df0fb16a248a2f32a0

Change-Id: I8e2b425f9008e6b82d41d59783bb8b04af1f886f
Fixes: https://github.com/google/protobuf/issues/474.

Fixes:
  https://github.com/google/protobuf/issues/502
  https://github.com/google/protobuf/issues/425

upb no longer requires -std=c99 but the Ruby/C
code still uses C99 features.

Since this version of upb supports C89, all of the
extra compiler flags are no longer required.

Also fixed lines with > 80 char length.

- Alter encode/decode paths to use the `upb_env` (environment)
  abstraction.
- Update upb amalgamation to upstream `93791bfe`.
- Fix a compilation warning (void*->char* cast).
- Modify build flags so that upb doesn't produce warnings -- the Travis
  build logs were pretty cluttered previously.

* make consistent between mri and jruby
* create a #to_h and have it use symbols for keys
* add #to_json and #to_proto helpers on the Google::Protobuf message classes

ruby arrays don't throw an exception; they return nil.  Lets do the
same!

this fix also includes the ability to use negative array indicies

Includes repro test case from @wfarr.

Previously, we supported map fields in the Ruby DSL. However, we never
connected the final link in the chain and generated `map` DSL commands
for map fields in `.proto` files. My apologies -- I had been testing
with the DSL directly so I missed this.

Also fixed a handlerdata-setup-infinite-loop when a map value field's
type is its containing message.