Change-Id: Ic4f49e94c67ba23f07e5ba7e87ca94e8db5d5906
Tested: on Linux.

Change-Id: I4db7f032440c00d31c7434975b8a3f17c29d74b9
Tested: on Linux.

Change-Id: Ia098126c92ea6f8abe94561bd9c09b29ac8f61f5
Context: https://github.com/google/flatbuffers/issues/4591
Tested: on Linux.

Change-Id: I79724416110a3d2241d7eeecf0e12aa6d5b760a0

Change-Id: I9420bf01d2bdb6ccfe3703c835297be73457c0d7

Removed internal variable with name "data" inside a generated method _nested_root(). Local variable with name "data" violates [-Werror=shadow]. (#4587)

(mirrored from cl/180944741)

Change-Id: Ib9f1613ccc355528c5c495fd953a9b3944ecb4ce

Date:   Mon Jan 15 11:38:20 2018 -0200

Compilation failure with grpc.h

If cmake run with flag FLATBUFFERS_BUILD_GRPCTEST=ON
compilation fails.

Fix :
	-Fix argument list for overriden function in grpc.
	-Fix member function name called by FlatBufferBuilder from
	buf() to scratch_data()

Change-Id: I81cf4d268670bdd11d1d56ca9f2de78c120df842
Tested: on Linux.

- Added Conan recipe to build Flatbuffers
- Added Travis job to create Conan package
Signed-off-by: Uilian Ries <uilianries@gmail.com>

Previously, FlatBufferBuilder used 3 resizable buffers:
- serialization (vector_downward)
- field offsets (std::vector)
- vtable offsets (std::vector)

Since the serialization buffer grows downwards, the bottom part of
it can be used as a "scratchpad" storage for the other two. Since
field offsets are only accumulated during table construction, and
vtable offsets only after table construction, the two can trivially
share the same storage.

Not only does this reduce the amount of allocation, it also removes
the bulk of std::vector usage from FlatBufferBuilder which was
the #1 cause of slow-down in debug mode, see e.g.:
https://stackoverflow.com/questions/36085285/any-way-to-improve-flatbuffer-performance-in-debug-c-msvc

Change-Id: I0224cf2f2a863d2d7ef762bc9163b52fdc149522
Tested: on Linux.

Without this change, the compiler tries to select the following overload
when CreateString is passed a `char *`:

    template<typename T>
    Offset<String> CreateString(const T &str) {
      return CreateString(str.c_str(), str.length());
    }

which is not valid since char pointers don't have methods.

(Fixes #4579)
Signed-off-by: Andrew Gunnerson <chenxiaolong@cxl.epac.to>

Avoids the following compile error when char is unsigned:

error: comparison of unsigned expression >= 0 is always true [-Werror,-Wtautological-unsigned-zero-compare]

GenerateText generate a text string for default identifier of enum and float scalar. An enum identifier printed as the strictly quoted string. (#4527)

* new maven jar called flatbuffers-java-grpc which contains utility class supporting generated code for grpc over flatbuffers; grpc java unit test based on maven and junit (maven is used only for simplicity of testing); removed straneous namespace_test/NamespaceA/TableInC.java which is not longer used in the test and no longer generated but contains complilation errors if java compiler picks it up

* moved java grpc tests files according to review request

* Added missing generated code for Java gRPC.

Change-Id: Iada090fe2e99b80a4b7c2f8e39c838a992a1feae

* added missing name and url

Change-Id: Iec720991e68ad27580537135bfd8eb4159333921
Tested: on Linux.

Change-Id: I6d2a9cc8b32c531d4a1e272ddf29974a0a1a2ee3

* grpc bindings generator for Java and a few minor supporting changes in improvements

* restored formatting before my previous changes for ease of review

* Fixed grpc java code generation bug resulting in duplicate extractor declarations in case the same is used in more than a single RPC method

* fixed previous merge issue

* removed extra space

* restored extra space

* restored extra space

* fixed java codegen bug documented in https://github.com/google/flatbuffers/issues/4563

Also added missing generated files.

Change-Id: Ifd22a643a08e3f2edfce92812ed57b87fc0e1875

* added config files for popular formatting tools

* edited and moved formatting files to root per review suggestion

* grpc bindings generator for Java and a few minor supporting changes in improvements

* restored formatting before my previous changes for ease of review

* Fixed grpc java code generation bug resulting in duplicate extractor declarations in case the same is used in more than a single RPC method

* updateed cpp_generator.cc to be compatible with the latest grpc version

* preserved the original license

* synchronized grpc cpp_generator with latest version as of today: GRPC 1.8.1. Regenerated test/monster_test.grpc.fb.* files and verified that grpctest is nicely passing

* fixed merge glitch

* updateed cpp_generator.cc to be compatible with the latest grpc version

* preserved the original license

Add support for Python lib continuous deployment.

* Eclipse ignore

* TypeScript support

* Prefixing enums

* Test results

* Merged JS and TS generators

* Fixed AppVeyor build problems

* Fixed more AppVeyor build problems

* Fixed more AppVeyor build problems

* Changed TS flag to options struct

* Storing options by value

* Removed unneeded const

* Re-export support for unions

* Uint support

* Casting bools to numbers for mutation

* TS shell tests

* Reverted generates js test file to original version

* Backing up js tests and properly generating test data

* Not importing flatbuffers for TS test generation

* Not overwriting generated js for tests

* AppVeyor test fixes

* Generating the most strict TS code possible

* Not returning null when creating vectors

* Not returning null from struct contructors

* Vector of unions for ts/js

* Sanity check for languages

* Indentation fix + output test files

* Vectors of unions for php

* Fixes to union vector handling + tests

* Fix for strictPropertyInitialization

Change-Id: I16bf46682ad68c62b778ed1514b6b706c8cb8f6f
Tested: on Linux.

added "native_custom_alloc" attribute to tables/structs, eg.

table parent_table( native_custom_alloc:"custom_alloc_name" ) {
...
}

with a custom allocator defined as

template <typename T> class custom_alloc_name : public std::allocator<T> {
public:

 typedef T*       pointer;

 template <class U>
 struct rebind {
  typedef custom_alloc_name<U> other;
 };

 pointer allocate(const std::size_t n) {
      return ....;
 }

 void deallocate(T* ptr, std::size_t n) {
    ...
 }

 custom_alloc_name() throw() {}
 template <class U> custom_alloc_name(const custom_alloc_name<U>&) throw() {}
};
};

Use a combination of travis and twine to publish to PyPI. New
publications will be made:

* When `master` is updated. This will trigger the publication of a
  the Python artifact versioned an iso-compliant build datetime. In this
  way, the cutting edge version will always be available via PyPI.
* When a new git tag is pushed. Tag pushes trigger the publication of
  a python artifact with the same version as the git tag, with the
  leading `v` stripped if present (`v1.2.3` becomes `1.2.3`).

Publications rely on Travis having a PYPI_PASSWORD environment set in
the project settings. See the Travis CI documentation for information on
[setting environment variables which containing sensitive data][1]. Make
extra sure the "Display value in build log" switch is OFF.

In addition to setting the previously mentioned `PYPI_PASSWORD`
environment variable, the owner of the PyPI `flatbuffers` repository
should, after merging this commit into master, add his own commit to
change `mikeholler` in `.travis/deploy-python.sh` to his username. It's
also recommended that the owner of `flatbuffers` use a separate account
in the unlikely event that the environment variable somehow becomes
compromised. Again, this is very unlikely, since the environment
variable is only set for "safe" builds approved by maintainers (not on
random pull requests).

[1]: https://docs.travis-ci.com/user/environment-variables/#Defining-Variables-in-Repository-Settings