Convert integer to string if field is string field in json

Use 5.6.8 instead

json_decode automatically convert numeric string to integer, so that
we need to convert it back. However, this will suceed to parse invalid
json data with string field set as integer even though it should have failed.
Because, the failure case is less often than the succeeding case, we decided
to make this change. Users should make sure their data don't use integer for
string fields by themselves.

Default values should also be serialized for json map key/value (#5643)

cherrypick changes from master

* Use set -ex
* Use -std=c++14 so -latomic can be found

* Fix python release build.
* Separate config.sh by platform
* Use KOKORO_JOR_NAME to determine the BUILD_COMMIT.
* Use -std=c++14 on MacOS so -latomic can be used.

Previously, internalAddGeneratedFile has to depend on old non-nested name for sub-messages.
This creates a hard dependency on old generated code for compatibility usage.
If user's code has custome error handler, the deprecation warning in the old generated code
will also be thrown (even though users haven't explicitly depended on the old message name).

To fix this problem, this change added an additional flag in the generated code to tell run
time that it's safe to use new message name. In this way, internalAddGeneratedFile can safely
depend on new name and don't trigger unnecessary deprecation warning.

Update CHANGES.txt

Update version to 3.7.0

Merge new changes into 3.7.x

* Fix more issues for reference values

* Revert change in gdb test

* Add more tests

* Fix protoc release

* Update Dockerfile to include libatomic

* Add TC malloc kokoro test

* Update tcmalloc test

* Fix empty FieldMask json encoding/decoding

* Add failed test to python's conformance failure list

* Fix c extension doesn' allow message reference in array

* Fix array constructor handling reference of array.

* Change test name

Add proto3 to solidity link in third_party.md

* Fixed compile warnings for PHP extension.

* Fixed some compile errors on PHP <7 and with TSR enabled.

* One more fix for TSRM builds.

Revert "doc: apply consistent indentation to Any examples"

This reverts commit b3367aee.

* Fix Any json encoding/decoding for ruby.

* Revert unnecessary changes

* Fix field mask for php in json

* Update byte size

pb3-gen-sol is a proto3 to solidity library generator that supports proto3 native types and uses field option for solidity native types. Both the message and generated code are more efficient than other solutions. It also includes a library to decode protobuf wireformat.

* Ruby nested msg caps

* nested types tests added

* test cases added for RubifyConstant

* extracted NestedMessage to TestNestedMessage

* removed nested tests with leading underscore

* extracted testLowercase to toplevel testLowercaseNested

ObjC doesn't support them, but this ensure when tests are added
they will be handled correctly.

The conformance.proto was updated in PR #5566

Varadic assignment in javascript where the right hand expression is an
object yields a single set assignment with the right most lefthand variable,
while leaving the other left side expressions undefined.

For example:
var a,b = {}

will only initialize a to undefined. But will set b to the {} value.

--

For this code since f is never used, it is a safe operation to remove
it.

* Down-integrate internal changes to github.

* Fix BUILD file and regenerate csharp descritpor

Added a few \ on the comments to match the exceptions scope when is referenced while in a namespace. So the IDEs can be happy.

* Down-integrate internal changes to github.

* fix csharp conformance test

* add comments in conformance.proto for text format

* fix comments

* fix comments, re-generated csharp file

* fix comments, re-generated csharp file

Even though the comments were indented to appear to go with the jspb
case/field, protoc doesn't collect comments like that, so these "hanging"
comments actually "attach" to the next thing added to each.  Looking at
https://github.com/protocolbuffers/protobuf/pull/5566 you see where
the generated code picked up the comment on the wrong field.

* Down-integrate internal changes to github.

* fix python conformance test

* fix csharp conformance test

* add back java map_lite_test.proto's optimize for option

* fix php conformance test