nanopb-0.4.6-bp155.2.6

- Revert back to 0.4.6 since 0.4.7 breaks nanopb_generator.py thus breaking
  some packages (e.g. Criterion: https://github.com/nanopb/nanopb/issues/845)

- Revert fix-missing-file.patch.

- Add fix-missing-file.patch (generator/proto/__init__.py).

- update to 0.4.7:
  * Fix comments generation for submessages
  * Fix handling of spaces in arguments passed through protoc
  * Fix problems with multiple files and mangle_names option
  * Fix generator error when using M_STRIP_PACKAGE without package name
  * Fix compilation error with fixed size array and FT_POINTER
  * Fix wrong format in Python Poetry project file
  * Fix unnecessary generator message when using --quiet
  * Fix enum_to_string with C++
  Fix /* */ inside .proto file comment
  * Workaround python-protobuf version issues
  * Safeguard substraction in pb_read() with custom streams
  * Always include pb_release() as function, instead of macro.
  Allow using = instead of : with generator option -s
  * Allow specifying include format without special characters
  * Allow including headers from inside of extern C
  * Add option NANOPB_PB2_TEMP_DIR to store nanopb_pb2.py in a temporary dir
  * Add compile-time error message for when PB_FIELD_32BIT is needed
  * Add --c-style command line option for naming style
  * Add --protoc-opt to nanopb_generator.py
  * Add ENUMTYPE convenience macros
  * Add Bazel build rules

- Add 'PB_ENABLE_MALLOC=1' so that we can build Criterion.

- Update to release 0.4.6
  * Fix passing of error message from substream callback
  * Fix regression in 0.4.3 where generator did not find all
    dependencies
  * Fix double-definition errors with size_union
  * Fix generator error with same inner message name
  * Fix infinite recursion in generator/protoc script
  * Use C11 static assert mechanism by default
  * Support --proto-path as alias to -I
  * Use memcpy instead of iterating on buf_read/write
  * Add generator support for PlatformIO
  * Improve handling of varint overflows
  * Improve optimization for little-endian platforms

- Update to version 0.4.1:
  * Fix for CVE-2020-5235 - porential exploitable out of memory
    condition.
  * Fixes for AVR builds and tests.

- As we download the tags directly no need for services and we
  can fetch it using URL paths
- Do not build generators as those require python2 and we are
  removing that from the dsitribution
- Update to 0.4.0:
  * Make nanopb_generator automatically compile .proto files (#462)
  * Allow installing as Python package from pip (#460)
  * Use protoc from grpcio-tools Python package if available (#463)
  * Change proto3 message types to be optional (#308, #452)
  * Add pb_decode_ex(), pb_encode_ex() functions.
  * Automatically rebuild nanopb_pb2.py
  * Use plugin.proto that comes with python-protobuf (#234)

- Enable PB_FIELD_16BIT setting to add support for tag numbers >
  255 and fields larger than 255 bytes.

- Add nanopb-source package - some Go programs using gRPC need
  protobuf definitions which are included inside the source code,
  but are not included in the devel package.

- Update to version 0.3.9.2:
  * Update changelog & authors
  * Fix splint test error (#359)
  * Clarify the usage of callback fields inside oneof sections in the docs
  * cmake: Include CMAKE_INSTALL_LIBDIR in CMAKE_INSTALL_CMAKEDIR
  * cmake: Allow to build a shared library
  * Made some changes for Error Output.
  * Update README.md
  * Further test fail fix for Windows

- Initial release

- Install nanopb also under /usr/bin/nanopb_generator.py - this fixes an
  error when using with protoc

- Update to 0.4.5:
  * Fix CVE-2021-21401: Release of Invalid Pointer or Reference
  * Fix unordered field numbers inside oneof causing fields to be ignored
  * Fix pb_decode() not initializing fields inside oneof
  * Fix compiler errors with complex oneof hierarchy and sizeof()
  * Fix descriptor width calculation for 64-bit types
  * Fix compiler error in generated initializer for submessage callback
  * Fix duplicate union definition in generated file
  * Add support for including comments from .proto file
- Update to 0.4.4:
  * Fix CVE-2020-26243, memory leak
  * Fix generator error when oneof contains a single unresolved field size
  * Fix unsigned enums not working correctly inside OneOf
  * Fix recursion depth exceeded error in generator
  * Add support for proto3 optional fields introduced in protoc 3.12
  * Add better error message when enum type is not found
  * Allow specifying descriptorsize on field level also
  * Allow multiple targets to be created by calling NANOPB_GENERATE_CPP()
- Update to 0.4.3
  * Fix issue with unknown message sizes being referred inside oneof
    encoded size calculation
  * Fix problem with [default=xxxx, (nanopb).proto3=true] field
    option combination
  * Better support C++ types in generated structs
  * Use grpcio protoc
  * Make pb_decode_varint32() reject overlong varint encodings.
  * Performance optimizations
  * !! NOTE: version 0.4.3 changes layout of pb_msgdesc_t. It requires
    recompiling .pb.c files and thus breaks ABI compatibility.
- Enabled build of generator (requires Python3)

- update to 0.4.2:
  * Fix buffer overflow when encoding bytes with size set to 65535 (#547, GHSA-3p39-mfxg-hrq4)
  * Fix segfault with pointer fields and proto3_singular_msgs = true. (#504,#505)
  * Fix Windows 10 temp file handling (#486)
  * Fix macro name conflicts (ATMEGA32U4 UENUM and more) (#522)
  * Fix generator error with nested messages and default values (#487)
  * Fix nanopb_generator exception on enums with aliases (#535)
  * Fix compile error when struct members are called X or a (#492)
  * Fix sizeof(union ...) fallback not compiling with C++ (#415, #494)
  * Fix "missing required field" error with submessage callback (#544)
  * Fix field descriptor sizing with submsg_callback option (#545)
  * Fix protoc calling on Python 2 (#503)
  * Fix handling of varying NaN representations in PB_CONVERT_DOUBLE_FLOAT (#543)
  * Fix clang undefined behavior sanitizer errors.
  * Change generator to use Python 3 by default (#441, #509)
  * Binary packages updated to use Python 3 and grpcio-tools
  * Add support for infinity and nan floating-point defaults (#530, #538)
  * Add generator option sort_by_tag (#542)
  * Add type_override option to override type defined in .proto (#497)
  * Set proto2 enum fields to first value when no default is given, even if nonzero (#532,#539)
  * Include protoc-gen-nanopb in path in protoc wrapper script
  * Properly pass error status from protoc wrapper when calling binary protoc
  * Generator: pass --include_imports when calling protoc (#494)
  * Create intermediate directories when writing files to relative path (#512)
  * Add current directory to include path for protoc (#499)
  * Update readme to use nanopb_generator.py directly
  * Regression test for proto3 incorrectly considered empty (#504)
  * CMake: change package name to Nanopb for cmake 3.17 compatibility (#506)
  * CMake: remove find_package(PythonInterp) (#508)
  * CMake: use split --nanopb_opt only on protoc >= 3.6 (#515)
  * CMake: Relax python version spec, allowing Python3. (#534)
  * Swift package manager (#549)
  * Rename BUILD as BUILD.bazel (#537

- Update to version 0.4.1:
  * Fix for CVE-2020-5235 - porential exploitable out of memory
    condition.
  * Fixes for AVR builds and tests.

- As we download the tags directly no need for services and we
  can fetch it using URL paths
- Do not build generators as those require python2 and we are
  removing that from the dsitribution
- Update to 0.4.0:
  * Make nanopb_generator automatically compile .proto files (#462)
  * Allow installing as Python package from pip (#460)
  * Use protoc from grpcio-tools Python package if available (#463)
  * Change proto3 message types to be optional (#308, #452)
  * Add pb_decode_ex(), pb_encode_ex() functions.
  * Automatically rebuild nanopb_pb2.py
  * Use plugin.proto that comes with python-protobuf (#234)

- Update to version 0.3.9.4:
  * Publishing nanopb-0.3.9.4
  * Update changelog
  * Add test for PB_ENCODE_ARRAYS_UNPACKED
  * Fix "unused variable" warning when using PB_ENCODE_ARRAYS_UNPACKED
  * Fix undefined behavior with bool fields (#434)
  * Make fuzztest detect invalid bool values (#434)
  * .gitignore: don't ignore generator-bin files (#419)
  * Don't call stream callback with count=0 (#421)
  * network_server example: Detect count=0 in read_callback() (#421)
  * Add scons package to travis rules, for build automation
  * Introduce new compile time flag: PB_ENCODE_ARRAYS_UNPACKED
  * Add migration note about enum min/max
  * Use min/max enum value for min/max
  * Publishing nanopb-0.3.9.3
  * Fix generator error with mangle_names option (#380)
  * Add testcase for #380
  * Fix incorrect PB_STATIC_ASSERT for bytes inside oneof (#363)
  * Add testcase for #363
  * Generator: Allow comma separated options in plugin mode (#343)
  * Fix encoding of fixed size arrays inside proto3 submessages (#376)
  * Add regression test for issue #376

- Enable PB_FIELD_16BIT setting to add support for tag numbers >
  255 and fields larger than 255 bytes.

- Add nanopb-source package - some Go programs using gRPC need
  protobuf definitions which are included inside the source code,
  but are not included in the devel package.

- Update to version 0.3.9.2:
  * Update changelog & authors
  * Fix splint test error (#359)
  * Clarify the usage of callback fields inside oneof sections in the docs
  * cmake: Include CMAKE_INSTALL_LIBDIR in CMAKE_INSTALL_CMAKEDIR
  * cmake: Allow to build a shared library
  * Made some changes for Error Output.
  * Update README.md
  * Further test fail fix for Windows

- Initial release