Protobuf

C++
* Introduced workaround for Windows issue with std::atomic and std::once_flag initialization (4777, 4773)

PHP
* Added compatibility with PHP 7.3 (4898)

Ruby
* Fixed Ruby crash involving Any encoding (4718)

General
* We are moving protobuf repository to its own github organization (see https://github.com/google/protobuf/issues/4796). Please let us know what you think about the move by taking this survey: https://docs.google.com/forms/d/e/1FAIpQLSeH1ckwm6ZrSfmtrOjRwmF3yCSWQbbO5pTPqPb6_rUppgvBqA/viewform

C++
* Starting from this release, we now require C++11. For those we cannot yet upgrade to C++11, we will try to keep the 3.5.x branch updated with critical bug fixes only. If you have any concerns about this, please comment on issue 2780.
* Moved to C++11 types like std::atomic and std::unique_ptr and away from our old custom-built equivalents.
* Added support for repeated message fields in lite protos using implicit weak fields. This is an experimental feature that allows the linker to strip out more unused messages than previously was possible.
* Fixed SourceCodeInfo for interpreted options and extension range options.
* Fixed always_print_enums_as_ints option for JSON serialization.
* Added support for ignoring unknown enum values when parsing JSON.
* Create std::string in Arena memory.
* Fixed ValidateDateTime to correctly check the day.
* Fixed bug in ZeroCopyStreamByteSink.
* Various other cleanups and fixes.

Java
* Dropped support for Java 6.
* Added a UTF-8 decoder that uses Unsafe to directly decode a byte buffer.
* Added deprecation annotations to generated code for deprecated oneof fields.
* Fixed map field serialization in DynamicMessage.
* Cleanup and documentation for Java Lite runtime.
* Various other fixes and cleanups
* Fixed unboxed arraylists to handle an edge case
* Improved performance for copying between unboxed arraylists
* Fixed lite protobuf to avoid Java compiler warnings
* Improved test coverage for lite runtime
* Performance improvements for lite runtime

Python
* Fixed bytes/string map key incompatibility between C++ and pure-Python implementations (issue 4029)
* Added `__init__.py` files to compiler and util subpackages
* Use /MT for all Windows versions
* Fixed an issue affecting the Python-C++ implementation when used with Cython (issue 2896)
* Various text format fixes
* Various fixes to resolve behavior differences between the pure-Python and Python-C++ implementations

PHP
* Added php_metadata_namespace to control the file path of generated metadata file.
* Changed generated classes of nested message/enum. E.g., Foo.Bar, which previously generates Foo_Bar, now generates Foo/Bar
* Added array constructor. When creating a message, users can pass a php array whose content is field name to value pairs into constructor. The created message will be initialized according to the array. Note that message field should use a message value instead of a sub-array.
* Various bug fixes.

Objective-C
* We removed some helper class methods from GPBDictionary to shrink the size of the library, the functionary is still there, but you may need to do some specific +alloc / -init… methods instead.
* Minor improvements in the performance of object field getters/setters by avoiding some memory management overhead.
* Fix a memory leak during the raising of some errors.
* Make header importing completely order independent.
* Small code improvements for things the undefined behaviors compiler option was flagging.

Ruby
* Added ruby_package file option to control the module of generated class.
* Various bug fixes.

Javascript
* Allow setting string to int64 field.

Csharp
* Unknown fields are now parsed and then sent back on the wire. They can be discarded at parse time via a CodedInputStream option.
* Movement towards working with .NET 3.5 and Unity
* Expression trees are no longer used
* AOT generics issues in Unity/il2cpp have a workaround (see commit 1b219a174c413af3b18a082a4295ce47932314c4 for details)
* Floating point values are now compared bitwise (affects NaN value comparisons)
* The default size limit when parsing is now 2GB rather than 64MB
* MessageParser now supports parsing from a slice of a byte array
* JSON list parsing now accepts null values where the underlying proto representation does

Planned Future Changes
* Make C++ implementation C++11 only: we plan to require C++11 to build protobuf code starting from 3.6.0 release, after unknown fields semantic changes are finished. Please join this [github issue](https://github.com/google/protobuf/issues/2780) to provide your feedback.

protoc
* Fixed a bug introduced in 3.5.0 and protoc in Windows now accepts non-ascii characters in paths again.

C++
* Removed several usages of C++11 features in the code base.
* Fixed some compiler warnings.

PHP
* Fixed memory leak in C-extension implementation.
* Added `discardUnknokwnFields` API.
* Removed duplicatd typedef in C-extension headers.
* Avoided calling private php methods (`timelib_update_ts`).
* Fixed `Any.php` to use fully-qualified name for `DescriptorPool`.

Ruby
* Added `Google_Protobuf_discard_unknown` for discarding unknown fields in
messages.

C
* Unknown fields are now preserved by default.
* Floating point values are now bitwise compared, affecting message equality check and `Contains()` API in map and repeated fields.

Planned Future Changes
* Make C++ implementation C++11 only: we plan to require C++11 to build protobuf code starting from 3.5.0 or 3.6.0 release, after unknown fields semantic changes are finished. Please join this [github issue](https://github.com/google/protobuf/issues/2780) to provide your feedback.

General
* Unknown fields are now preserved in proto3 for most of the language implementations for proto3 by default. See the per-language section for details.
* reserve keyword are now supported in enums

C++
* Proto3 messages are now preserving unknown fields by default. If you rely on unknowns fields being dropped. Please use DiscardUnknownFields() explicitly.
* Deprecated the `unsafe_arena_release_*` and `unsafe_arena_add_allocated_*` methods for string fields.
* Added move constructor and move assignment to RepeatedField, RepeatedPtrField and google::protobuf::Any.
* Added perfect forwarding in Arena::CreateMessage
* In-progress experimental support for implicit weak fields with lite protos. This feature allows the linker to strip out more unused messages and reduce binary size.
* Various performance optimizations.

Java
* Proto3 messages are now preserving unknown fields by default. If you’d like to drop unknown fields, please use the DiscardUnknownFieldsParser API. For example:
java
Parser<Foo> parser = DiscardUnknownFieldsParser.wrap(Foo.parser());
Foo foo = parser.parseFrom(input);

* Added a new `CodedInputStream` decoder for `Iterable<ByteBuffer>` with direct ByteBuffers.
* `TextFormat` now prints unknown length-delimited fields as messages if possible.
* `FieldMaskUtil.merge()` no longer creates unnecessary empty messages when a message field is unset in both source message and destination message.
* Various performance optimizations.

Python
* Proto3 messages are now preserving unknown fields by default. Use `message.DiscardUnknownFields()` to drop unknown fields.
* Add FieldDescriptor.file in generated code.
* Add descriptor pool `FindOneofByName` in pure python.
* Change unknown enum values into unknown field set .
* Add more Python dict/list compatibility for `Struct`/`ListValue`.
* Add utf-8 support for `text_format.Merge()/Parse()`.
* Support numeric unknown enum values for proto3 JSON format.
* Add warning for Unexpected end-group tag in cpp extension.

PHP
* Proto3 messages are now preserving unknown fields.
* Provide well known type messages in runtime.
* Add prefix ‘PB’ to generated class of reserved names.
* Fixed all conformance tests for encode/decode json in php runtime. C extension needs more work.

Objective-C
* Fixed some issues around copying of messages with unknown fields and then mutating the unknown fields in the copy.

C
* Added unknown field support in JsonParser.
* Fixed oneof message field merge.
* Simplify parsing messages from array slices.

Ruby
* Unknown fields are now preserved by default.
* Fixed several bugs for segment fault.

Javascript
* Decoder can handle both paced and unpacked data no matter how the proto is defined.
* Decoder now accept long varint for 32 bit integers.

This is mostly a bug fix release on runtime packages. It is safe to use 3.4.0 protoc packages for this release.
* Fixed the missing files in 3.4.0 tarballs, affecting windows and cmake users.
* C: Fixed dotnet target platform to be net45 again.
* Ruby: Fixed a segmentation error when using maps in multi-threaded cases.
* PHP: php_generic_service file level option tag number (in descriptor.proto) has been reassigned to avoid conflicts.

Planned Future Changes
* Preserve unknown fields in proto3: We are going to bring unknown fields back into proto3. In this release, some languages start to support preserving unknown fields in proto3, controlled by flags/options. Some languages also introduce explicit APIs to drop unknown fields for migration. Please read the change log sections by languages for details. See [general timeline and plan](https://docs.google.com/document/d/1KMRX-G91Aa-Y2FkEaHeeviLRRNblgIahbsk4wA14gRk/view) and [issues and discussions](https://github.com/google/protobuf/issues/272)

* Make C++ implementation C++11 only: we plan to require C++11 to build protobuf code starting from 3.5.0 or 3.6.0 release, after unknown fields semantic changes are finished. Please join this [github issue](https://github.com/google/protobuf/issues/2780) to provide your feedback.

General
* Extension ranges now accept options and are customizable.
* reserve keyword now supports max in field number ranges, e.g. reserve 1000 to max;

C++
* Proto3 messages are now able to preserve unknown fields. The default behavior is still to drop unknowns, which will be flipped in a future release. If you rely on unknowns fields being dropped. Please use Message::DiscardUnknownFields() explicitly.
* Packable proto3 fields are now packed by default in serialization.
* Following C++11 features are introduced when C++11 is available:
- move-constructor and move-assignment are introduced to messages
- Repeated fields constructor now takes std::initializer_list
- rvalue setters are introduced for string fields
* Experimental Table-Driven parsing and serialization available to test. To enable it, pass in table_driven_parsing table_driven_serialization protoc generator flags for C++

$ protoc --cpp_out=table_driven_parsing,table_driven_serialization:./ test.proto

* lite generator parameter supported by the generator. Once set, all generated files, use lite runtime regardless of the optimizer_for setting in the .proto file.
* Various optimizations to make C++ code more performant on PowerPC platform
* Fixed maps data corruption when the maps are modified by both reflection API and generated API.
* Deterministic serialization on maps reflection now uses stable sort.
* file() accessors are introduced to various *Descriptor classes to make writing template function easier.
* ByteSize() and SpaceUsed() are deprecated.Use ByteSizeLong() and SpaceUsedLong() instead
* Consistent hash function is used for maps in DEBUG and NDEBUG build.
* "using namespace std" is removed from stubs/common.h
* Various performance optimizations and bug fixes

Java
* Introduced new parser API DiscardUnknownFieldsParser in preparation of proto3 unknown fields preservation change. Users who want to drop unknown fields should migrate to use this new parser API.
For example:

java
Parser<Foo> parser = DiscardUnknownFieldsParser.wrap(Foo.parser());
Foo foo = parser.parseFrom(input);


* Introduced new TextFormat API printUnicodeFieldValue() that prints field value without escaping unicode characters.
* Added Durations.compare(Duration, Duration) and Timestamps.compare(Timestamp, Timestamp).
* JsonFormat now accepts base64url encoded bytes fields.
* Optimized CodedInputStream to do less copies when parsing large bytes fields.
* Optimized TextFormat to allocate less memory when printing.

Python
* SerializeToString API is changed to SerializeToString(self, **kwargs), deterministic parameter is accepted for deterministic serialization.
* Added sort_keys parameter in json format to make the output deterministic.
* Added indent parameter in json format.
* Added extension support in json format.
* Added __repr__ support for repeated field in cpp implementation.
* Added file in FieldDescriptor.
* Added pretty-print filter to text format.
* Services and method descriptors are always printed even if generic_service option is turned off.
* Note: AppEngine 2.5 is deprecated on June 2017 that AppEngine 2.5 will never update protobuf runtime. Users who depend on AppEngine 2.5 should use old protoc.

PHP
* Support PHP generic services. Specify file option php_generic_service=true to enable generating service interface.
* Message, repeated and map fields setters take value instead of reference.
* Added map iterator in c extension.
* Support json  encode/decode.
* Added more type info in getter/setter phpdoc
* Fixed the problem that c extension and php implementation cannot be used together.
* Added file option php_namespace to use custom php namespace instead of package.
* Added fluent setter.
* Added descriptor API in runtime for custom encode/decode.
* Various bug fixes.

Objective-C
* Fix for GPBExtensionRegistry copying and add tests.
* Optimize GPBDictionary.m codegen to reduce size of overall library by 46K per architecture.
* Fix some cases of reading of 64bit map values.
* Properly error on a tag with field number zero.
* Preserve unknown fields in proto3 syntax files.
* Document the exceptions on some of the writing apis.

C
* Implemented IReadOnlyDictionary<K,V> in MapField<K,V>
* Added TryUnpack method for Any message in addition to Unpack.
* Converted C projects to MSBuild (csproj) format.

Ruby
* Several bug fixes.

Javascript
* Added support of field option js_type. Now one can specify the JS type of a 64-bit integer field to be string in the generated code by adding option [jstype = JS_STRING] on the field.