package google.protobuf

Get desktop application:
View/edit binary Protocol Buffers messages

Describes a message type.

Used in: FileDescriptorProto

• optional string name = 1

• repeated FieldDescriptorProto field = 2

• repeated FieldDescriptorProto extension = 6

• repeated DescriptorProto nested_type = 3

• repeated EnumDescriptorProto enum_type = 4

• repeated DescriptorProto.ExtensionRange extension_range = 5

• repeated OneofDescriptorProto oneof_decl = 8

• optional MessageOptions options = 7

• repeated DescriptorProto.ReservedRange reserved_range = 9

• repeated string reserved_name = 10

  Reserved field names, which may not be used by fields in the same message. A given name may only be reserved once.

Used in: DescriptorProto

• optional int32 start = 1

  Inclusive.

• optional int32 end = 2

  Exclusive.

• optional ExtensionRangeOptions options = 3

Range of reserved tag numbers. Reserved tag numbers may not be used by fields or extension ranges in the same message. Reserved ranges may not overlap.

Used in: DescriptorProto

• optional int32 start = 1

  Inclusive.

• optional int32 end = 2

  Exclusive.

The full set of known editions.

Used in: FeatureSetDefaults, FeatureSetDefaults.FeatureSetEditionDefault, FieldOptions.EditionDefault, FieldOptions.FeatureSupport, FileDescriptorProto

• EDITION_UNKNOWN = 0

  A placeholder for an unknown edition value.

• EDITION_LEGACY = 900

  A placeholder edition for specifying default behaviors *before* a feature was first introduced. This is effectively an "infinite past".

• EDITION_PROTO2 = 998

  Legacy syntax "editions". These pre-date editions, but behave much like distinct editions. These can't be used to specify the edition of proto files, but feature definitions must supply proto2/proto3 defaults for backwards compatibility.

• EDITION_PROTO3 = 999

• EDITION_2023 = 1000

  Editions that have been released. The specific values are arbitrary and should not be depended on, but they will always be time-ordered for easy comparison.

• EDITION_2024 = 1001

• EDITION_1_TEST_ONLY = 1

  Placeholder editions for testing feature resolution. These should not be used or relyed on outside of tests.

• EDITION_2_TEST_ONLY = 2

• EDITION_99997_TEST_ONLY = 99997

• EDITION_99998_TEST_ONLY = 99998

• EDITION_99999_TEST_ONLY = 99999

• EDITION_MAX = 2147483647

  Placeholder for specifying unbounded edition support. This should only ever be used by plugins that can expect to never require any changes to support a new edition.

Describes an enum type.

Used in: DescriptorProto, FileDescriptorProto

• optional string name = 1

• repeated EnumValueDescriptorProto value = 2

• optional EnumOptions options = 3

• repeated EnumDescriptorProto.EnumReservedRange reserved_range = 4

  Range of reserved numeric values. Reserved numeric values may not be used by enum values in the same enum declaration. Reserved ranges may not overlap.

• repeated string reserved_name = 5

  Reserved enum value names, which may not be reused. A given name may only be reserved once.

Range of reserved numeric values. Reserved values may not be used by entries in the same enum. Reserved ranges may not overlap. Note that this is distinct from DescriptorProto.ReservedRange in that it is inclusive such that it can appropriately represent the entire int32 domain.

Used in: EnumDescriptorProto

• optional int32 start = 1

  Inclusive.

• optional int32 end = 2

  Inclusive.

Used in: EnumDescriptorProto

• optional bool allow_alias = 2

  Set this option to true to allow mapping different tag names to the same value.

• optional bool deprecated = 3

  Is this enum deprecated? Depending on the target platform, this can emit Deprecated annotations for the enum, or it will be completely ignored; in the very least, this is a formalization for deprecating enums.

• optional bool deprecated_legacy_json_field_conflicts = 6

  Enable the legacy handling of JSON field name conflicts. This lowercases and strips underscored from the fields before comparison in proto3 only. The new behavior takes `json_name` into account and applies to proto2 as well. TODO Remove this legacy behavior once downstream teams have had time to migrate.

• optional FeatureSet features = 7

  Any features defined in the specific edition.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Describes a value within an enum.

Used in: EnumDescriptorProto

• optional string name = 1

• optional int32 number = 2

• optional EnumValueOptions options = 3

Used in: EnumValueDescriptorProto

• optional bool deprecated = 1

  Is this enum value deprecated? Depending on the target platform, this can emit Deprecated annotations for the enum value, or it will be completely ignored; in the very least, this is a formalization for deprecating enum values.

• optional FeatureSet features = 2

  Any features defined in the specific edition.

• optional bool debug_redact = 3

  Indicate that fields annotated with this enum value should not be printed out when using debug formats, e.g. when the field contains sensitive credentials.

• optional FieldOptions.FeatureSupport feature_support = 4

  Information about the support window of a feature value.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Used in: DescriptorProto.ExtensionRange

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

• repeated ExtensionRangeOptions.Declaration declaration = 2

  For external users: DO NOT USE. We are in the process of open sourcing extension declaration and executing internal cleanups before it can be used externally.

• optional FeatureSet features = 50

  Any features defined in the specific edition.

• optional ExtensionRangeOptions.VerificationState verification = 3

  The verification state of the range. TODO: flip the default to DECLARATION once all empty ranges are marked as UNVERIFIED.

Used in: ExtensionRangeOptions

• optional int32 number = 1

  The extension number declared within the extension range.

• optional string full_name = 2

  The fully-qualified name of the extension field. There must be a leading dot in front of the full name.

• optional string type = 3

  The fully-qualified type name of the extension field. Unlike Metadata.type, Declaration.type must have a leading dot for messages and enums.

• optional bool reserved = 5

  If true, indicates that the number is reserved in the extension range, and any extension field with the number will fail to compile. Set this when a declared extension field is deleted.

• optional bool repeated = 6

  If true, indicates that the extension must be defined as repeated. Otherwise the extension must be defined as optional.

The verification state of the extension range.

Used in: ExtensionRangeOptions

• DECLARATION = 0

  All the extensions of the range must be declared.

• UNVERIFIED = 1

TODO Enums in C++ gencode (and potentially other languages) are not well scoped. This means that each of the feature enums below can clash with each other. The short names we've chosen maximize call-site readability, but leave us very open to this scenario. A future feature will be designed and implemented to handle this, hopefully before we ever hit a conflict here.

Used in: EnumOptions, EnumValueOptions, ExtensionRangeOptions, FeatureSetDefaults.FeatureSetEditionDefault, FieldOptions, FileOptions, MessageOptions, MethodOptions, OneofOptions, ServiceOptions

• optional FeatureSet.FieldPresence field_presence = 1

• optional FeatureSet.EnumType enum_type = 2

• optional FeatureSet.RepeatedFieldEncoding repeated_field_encoding = 3

• optional FeatureSet.Utf8Validation utf8_validation = 4

• optional FeatureSet.MessageEncoding message_encoding = 5

• optional FeatureSet.JsonFormat json_format = 6

Used in: FeatureSet

• ENUM_TYPE_UNKNOWN = 0

• OPEN = 1

• CLOSED = 2

Used in: FeatureSet

• FIELD_PRESENCE_UNKNOWN = 0

• EXPLICIT = 1

• IMPLICIT = 2

• LEGACY_REQUIRED = 3

Used in: FeatureSet

• JSON_FORMAT_UNKNOWN = 0

• ALLOW = 1

• LEGACY_BEST_EFFORT = 2

Used in: FeatureSet

• MESSAGE_ENCODING_UNKNOWN = 0

• LENGTH_PREFIXED = 1

• DELIMITED = 2

Used in: FeatureSet

• REPEATED_FIELD_ENCODING_UNKNOWN = 0

• PACKED = 1

• EXPANDED = 2

Used in: FeatureSet

• UTF8_VALIDATION_UNKNOWN = 0

• VERIFY = 2

• NONE = 3

A compiled specification for the defaults of a set of features. These messages are generated from FeatureSet extensions and can be used to seed feature resolution. The resolution with this object becomes a simple search for the closest matching edition, followed by proto merges.

• repeated FeatureSetDefaults.FeatureSetEditionDefault defaults = 1

• optional Edition minimum_edition = 4

  The minimum supported edition (inclusive) when this was constructed. Editions before this will not have defaults.

• optional Edition maximum_edition = 5

  The maximum known edition (inclusive) when this was constructed. Editions after this will not have reliable defaults.

A map from every known edition with a unique set of defaults to its defaults. Not all editions may be contained here. For a given edition, the defaults at the closest matching edition ordered at or before it should be used. This field must be in strict ascending order by edition.

Used in: FeatureSetDefaults

• optional Edition edition = 3

• optional FeatureSet overridable_features = 4

  Defaults of features that can be overridden in this edition.

• optional FeatureSet fixed_features = 5

  Defaults of features that can't be overridden in this edition.

Describes a field within a message.

Used in: DescriptorProto, FileDescriptorProto

• optional string name = 1

• optional int32 number = 3

• optional FieldDescriptorProto.Label label = 4

• optional FieldDescriptorProto.Type type = 5

  If type_name is set, this need not be set. If both this and type_name are set, this must be one of TYPE_ENUM, TYPE_MESSAGE or TYPE_GROUP.

• optional string type_name = 6

  For message and enum types, this is the name of the type. If the name starts with a '.', it is fully-qualified. Otherwise, C++-like scoping rules are used to find the type (i.e. first the nested types within this message are searched, then within the parent, on up to the root namespace).

• optional string extendee = 2

  For extensions, this is the name of the type being extended. It is resolved in the same manner as type_name.

• optional string default_value = 7

  For numeric types, contains the original text representation of the value. For booleans, "true" or "false". For strings, contains the default text contents (not escaped in any way). For bytes, contains the C escaped value. All bytes >= 128 are escaped.

• optional int32 oneof_index = 9

  If set, gives the index of a oneof in the containing type's oneof_decl list. This field is a member of that oneof.

• optional string json_name = 10

  JSON name of this field. The value is set by protocol compiler. If the user has set a "json_name" option on this field, that option's value will be used. Otherwise, it's deduced from the field's name by converting it to camelCase.

• optional FieldOptions options = 8

• optional bool proto3_optional = 17

  If true, this is a proto3 "optional". When a proto3 field is optional, it tracks presence regardless of field type. When proto3_optional is true, this field must belong to a oneof to signal to old proto3 clients that presence is tracked for this field. This oneof is known as a "synthetic" oneof, and this field must be its sole member (each proto3 optional field gets its own synthetic oneof). Synthetic oneofs exist in the descriptor only, and do not generate any API. Synthetic oneofs must be ordered after all "real" oneofs. For message fields, proto3_optional doesn't create any semantic change, since non-repeated message fields always track presence. However it still indicates the semantic detail of whether the user wrote "optional" or not. This can be useful for round-tripping the .proto file. For consistency we give message fields a synthetic oneof also, even though it is not required to track presence. This is especially important because the parser can't tell if a field is a message or an enum, so it must always create a synthetic oneof. Proto2 optional fields do not set this flag, because they already indicate optional with `LABEL_OPTIONAL`.

Used in: FieldDescriptorProto

• LABEL_OPTIONAL = 1

  0 is reserved for errors

• LABEL_REPEATED = 3

• LABEL_REQUIRED = 2

  The required label is only allowed in google.protobuf. In proto3 and Editions it's explicitly prohibited. In Editions, the `field_presence` feature can be used to get this behavior.

Used in: FieldDescriptorProto

• TYPE_DOUBLE = 1

  0 is reserved for errors. Order is weird for historical reasons.

• TYPE_FLOAT = 2

• TYPE_INT64 = 3

  Not ZigZag encoded. Negative numbers take 10 bytes. Use TYPE_SINT64 if negative values are likely.

• TYPE_UINT64 = 4

• TYPE_INT32 = 5

  Not ZigZag encoded. Negative numbers take 10 bytes. Use TYPE_SINT32 if negative values are likely.

• TYPE_FIXED64 = 6

• TYPE_FIXED32 = 7

• TYPE_BOOL = 8

• TYPE_STRING = 9

• TYPE_GROUP = 10

  Tag-delimited aggregate. Group type is deprecated and not supported after google.protobuf. However, Proto3 implementations should still be able to parse the group wire format and treat group fields as unknown fields. In Editions, the group wire format can be enabled via the `message_encoding` feature.

• TYPE_MESSAGE = 11

  Length-delimited aggregate.

• TYPE_BYTES = 12

  New in version 2.

• TYPE_UINT32 = 13

• TYPE_ENUM = 14

• TYPE_SFIXED32 = 15

• TYPE_SFIXED64 = 16

• TYPE_SINT32 = 17

  Uses ZigZag encoding.

• TYPE_SINT64 = 18

  Uses ZigZag encoding.

Used in: FieldDescriptorProto

• optional FieldOptions.CType ctype = 1

  NOTE: ctype is deprecated. Use `features.(pb.cpp).string_type` instead. The ctype option instructs the C++ code generator to use a different representation of the field than it normally would. See the specific options below. This option is only implemented to support use of [ctype=CORD] and [ctype=STRING] (the default) on non-repeated fields of type "bytes" in the open source release. TODO: make ctype actually deprecated.

• optional bool packed = 2

  The packed option can be enabled for repeated primitive fields to enable a more efficient representation on the wire. Rather than repeatedly writing the tag and type for each element, the entire array is encoded as a single length-delimited blob. In proto3, only explicit setting it to false will avoid using packed encoding. This option is prohibited in Editions, but the `repeated_field_encoding` feature can be used to control the behavior.

• optional FieldOptions.JSType jstype = 6

  The jstype option determines the JavaScript type used for values of the field. The option is permitted only for 64 bit integral and fixed types (int64, uint64, sint64, fixed64, sfixed64). A field with jstype JS_STRING is represented as JavaScript string, which avoids loss of precision that can happen when a large value is converted to a floating point JavaScript. Specifying JS_NUMBER for the jstype causes the generated JavaScript code to use the JavaScript "number" type. The behavior of the default option JS_NORMAL is implementation dependent. This option is an enum to permit additional types to be added, e.g. goog.math.Integer.

• optional bool lazy = 5

  Should this field be parsed lazily? Lazy applies only to message-type fields. It means that when the outer message is initially parsed, the inner message's contents will not be parsed but instead stored in encoded form. The inner message will actually be parsed when it is first accessed. This is only a hint. Implementations are free to choose whether to use eager or lazy parsing regardless of the value of this option. However, setting this option true suggests that the protocol author believes that using lazy parsing on this field is worth the additional bookkeeping overhead typically needed to implement it. This option does not affect the public interface of any generated code; all method signatures remain the same. Furthermore, thread-safety of the interface is not affected by this option; const methods remain safe to call from multiple threads concurrently, while non-const methods continue to require exclusive access. Note that lazy message fields are still eagerly verified to check ill-formed wireformat or missing required fields. Calling IsInitialized() on the outer message would fail if the inner message has missing required fields. Failed verification would result in parsing failure (except when uninitialized messages are acceptable).

• optional bool unverified_lazy = 15

  unverified_lazy does no correctness checks on the byte stream. This should only be used where lazy with verification is prohibitive for performance reasons.

• optional bool deprecated = 3

  Is this field deprecated? Depending on the target platform, this can emit Deprecated annotations for accessors, or it will be completely ignored; in the very least, this is a formalization for deprecating fields.

• optional bool weak = 10

  For Google-internal migration only. Do not use.

• optional bool debug_redact = 16

  Indicate that the field value should not be printed out when using debug formats, e.g. when the field contains sensitive credentials.

• optional FieldOptions.OptionRetention retention = 17

• repeated FieldOptions.OptionTargetType targets = 19

• repeated FieldOptions.EditionDefault edition_defaults = 20

• optional FeatureSet features = 21

  Any features defined in the specific edition.

• optional FieldOptions.FeatureSupport feature_support = 22

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Used in: FieldOptions

• STRING = 0

  Default mode.

• CORD = 1

  The option [ctype=CORD] may be applied to a non-repeated field of type "bytes". It indicates that in C++, the data should be stored in a Cord instead of a string. For very large strings, this may reduce memory fragmentation. It may also allow better performance when parsing from a Cord, or when parsing with aliasing enabled, as the parsed Cord may then alias the original buffer.

• STRING_PIECE = 2

Used in: FieldOptions

• optional Edition edition = 3

• optional string value = 2

  Textproto value.

Information about the support window of a feature.

Used in: EnumValueOptions, FieldOptions

• optional Edition edition_introduced = 1

  The edition that this feature was first available in. In editions earlier than this one, the default assigned to EDITION_LEGACY will be used, and proto files will not be able to override it.

• optional Edition edition_deprecated = 2

  The edition this feature becomes deprecated in. Using this after this edition may trigger warnings.

• optional string deprecation_warning = 3

  The deprecation warning text if this feature is used after the edition it was marked deprecated in.

• optional Edition edition_removed = 4

  The edition this feature is no longer available in. In editions after this one, the last default assigned will be used, and proto files will not be able to override it.

Used in: FieldOptions

• JS_NORMAL = 0

  Use the default type.

• JS_STRING = 1

  Use JavaScript strings.

• JS_NUMBER = 2

  Use JavaScript numbers.

If set to RETENTION_SOURCE, the option will be omitted from the binary. Note: as of January 2023, support for this is in progress and does not yet have an effect (b/264593489).

Used in: FieldOptions

• RETENTION_UNKNOWN = 0

• RETENTION_RUNTIME = 1

• RETENTION_SOURCE = 2

This indicates the types of entities that the field may apply to when used as an option. If it is unset, then the field may be freely used as an option on any kind of entity. Note: as of January 2023, support for this is in progress and does not yet have an effect (b/264593489).

Used in: FieldOptions

• TARGET_TYPE_UNKNOWN = 0

• TARGET_TYPE_FILE = 1

• TARGET_TYPE_EXTENSION_RANGE = 2

• TARGET_TYPE_MESSAGE = 3

• TARGET_TYPE_FIELD = 4

• TARGET_TYPE_ONEOF = 5

• TARGET_TYPE_ENUM = 6

• TARGET_TYPE_ENUM_ENTRY = 7

• TARGET_TYPE_SERVICE = 8

• TARGET_TYPE_METHOD = 9

Describes a complete .proto file.

Used in: FileDescriptorSet

• optional string name = 1

  file name, relative to root of source tree

• optional string package = 2

  e.g. "foo", "foo.bar", etc.

• repeated string dependency = 3

  Names of files imported by this file.

• repeated int32 public_dependency = 10

  Indexes of the public imported files in the dependency list above.

• repeated int32 weak_dependency = 11

  Indexes of the weak imported files in the dependency list. For Google-internal migration only. Do not use.

• repeated DescriptorProto message_type = 4

  All top-level definitions in this file.

• repeated EnumDescriptorProto enum_type = 5

• repeated ServiceDescriptorProto service = 6

• repeated FieldDescriptorProto extension = 7

• optional FileOptions options = 8

• optional SourceCodeInfo source_code_info = 9

  This field contains optional information about the original source code. You may safely remove this entire field without harming runtime functionality of the descriptors -- the information is needed only by development tools.

• optional string syntax = 12

  The syntax of the proto file. The supported values are "proto2", "proto3", and "editions". If `edition` is present, this value must be "editions".

• optional Edition edition = 14

  The edition of the proto file.

The protocol compiler can output a FileDescriptorSet containing the .proto files it parses.

• repeated FileDescriptorProto file = 1

Used in: FileDescriptorProto

• optional string java_package = 1

  Sets the Java package where classes generated from this .proto will be placed. By default, the proto package is used, but this is often inappropriate because proto packages do not normally start with backwards domain names.

• optional string java_outer_classname = 8

  Controls the name of the wrapper Java class generated for the .proto file. That class will always contain the .proto file's getDescriptor() method as well as any top-level extensions defined in the .proto file. If java_multiple_files is disabled, then all the other classes from the .proto file will be nested inside the single wrapper outer class.

• optional bool java_multiple_files = 10

  If enabled, then the Java code generator will generate a separate .java file for each top-level message, enum, and service defined in the .proto file. Thus, these types will *not* be nested inside the wrapper class named by java_outer_classname. However, the wrapper class will still be generated to contain the file's getDescriptor() method as well as any top-level extensions defined in the file.

• optional bool java_generate_equals_and_hash = 20

  This option does nothing.

• optional bool java_string_check_utf8 = 27

  A proto2 file can set this to true to opt in to UTF-8 checking for Java, which will throw an exception if invalid UTF-8 is parsed from the wire or assigned to a string field. TODO: clarify exactly what kinds of field types this option applies to, and update these docs accordingly. Proto3 files already perform these checks. Setting the option explicitly to false has no effect: it cannot be used to opt proto3 files out of UTF-8 checks.

• optional FileOptions.OptimizeMode optimize_for = 9

• optional string go_package = 11

  Sets the Go package where structs generated from this .proto will be placed. If omitted, the Go package will be derived from the following: - The basename of the package import path, if provided. - Otherwise, the package statement in the .proto file, if present. - Otherwise, the basename of the .proto file, without extension.

• optional bool cc_generic_services = 16

  Should generic services be generated in each language? "Generic" services are not specific to any particular RPC system. They are generated by the main code generators in each language (without additional plugins). Generic services were the only kind of service generation supported by early versions of google.protobuf. Generic services are now considered deprecated in favor of using plugins that generate code specific to your particular RPC system. Therefore, these default to false. Old code which depends on generic services should explicitly set them to true.

• optional bool java_generic_services = 17

• optional bool py_generic_services = 18

• optional bool deprecated = 23

  Is this file deprecated? Depending on the target platform, this can emit Deprecated annotations for everything in the file, or it will be completely ignored; in the very least, this is a formalization for deprecating files.

• optional bool cc_enable_arenas = 31

  Enables the use of arenas for the proto messages in this file. This applies only to generated classes for C++.

• optional string objc_class_prefix = 36

  Sets the objective c class prefix which is prepended to all objective c generated classes from this .proto. There is no default.

• optional string csharp_namespace = 37

  Namespace for generated classes; defaults to the package.

• optional string swift_prefix = 39

  By default Swift generators will take the proto package and CamelCase it replacing '.' with underscore and use that to prefix the types/symbols defined. When this options is provided, they will use this value instead to prefix the types/symbols defined.

• optional string php_class_prefix = 40

  Sets the php class prefix which is prepended to all php generated classes from this .proto. Default is empty.

• optional string php_namespace = 41

  Use this option to change the namespace of php generated classes. Default is empty. When this option is empty, the package name will be used for determining the namespace.

• optional string php_metadata_namespace = 44

  Use this option to change the namespace of php generated metadata classes. Default is empty. When this option is empty, the proto file name will be used for determining the namespace.

• optional string ruby_package = 45

  Use this option to change the package of ruby generated classes. Default is empty. When this option is not set, the package name will be used for determining the ruby package.

• optional FeatureSet features = 50

  Any features defined in the specific edition.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See the documentation for the "Options" section above.

Generated classes can be optimized for speed or code size.

Used in: FileOptions

• SPEED = 1

  Generate complete code for parsing, serialization,

• CODE_SIZE = 2

  etc.

  Use ReflectionOps to implement these methods.

• LITE_RUNTIME = 3

  Generate code using MessageLite and the lite runtime.

Describes the relationship between generated code and its original source file. A GeneratedCodeInfo message is associated with only one generated source file, but may contain references to different source .proto files.

• repeated GeneratedCodeInfo.Annotation annotation = 1

  An Annotation connects some span of text in generated code to an element of its generating .proto file.

Used in: GeneratedCodeInfo

• repeated int32 path = 1

  Identifies the element in the original source .proto file. This field is formatted the same as SourceCodeInfo.Location.path.

• optional string source_file = 2

  Identifies the filesystem path to the original source .proto.

• optional int32 begin = 3

  Identifies the starting offset in bytes in the generated code that relates to the identified object.

• optional int32 end = 4

  Identifies the ending offset in bytes in the generated code that relates to the identified object. The end offset should be one past the last relevant byte (so the length of the text = end - begin).

• optional Annotation.Semantic semantic = 5

Represents the identified object's effect on the element in the original .proto file.

Used in: Annotation

• NONE = 0

  There is no effect or the effect is indescribable.

• SET = 1

  The element is set or otherwise mutated.

• ALIAS = 2

  An alias to the element is returned.

Used in: DescriptorProto

• optional bool message_set_wire_format = 1

  Set true to use the old proto1 MessageSet wire format for extensions. This is provided for backwards-compatibility with the MessageSet wire format. You should not use this for any other reason: It's less efficient, has fewer features, and is more complicated. The message must be defined exactly as follows: message Foo { option message_set_wire_format = true; extensions 4 to max; } Note that the message cannot have any defined fields; MessageSets only have extensions. All extensions of your type must be singular messages; e.g. they cannot be int32s, enums, or repeated messages. Because this is an option, the above two restrictions are not enforced by the protocol compiler.

• optional bool no_standard_descriptor_accessor = 2

  Disables the generation of the standard "descriptor()" accessor, which can conflict with a field of the same name. This is meant to make migration from proto1 easier; new code should avoid fields named "descriptor".

• optional bool deprecated = 3

  Is this message deprecated? Depending on the target platform, this can emit Deprecated annotations for the message, or it will be completely ignored; in the very least, this is a formalization for deprecating messages.

• optional bool map_entry = 7

  Whether the message is an automatically generated map entry type for the maps field. For maps fields: map<KeyType, ValueType> map_field = 1; The parsed descriptor looks like: message MapFieldEntry { option map_entry = true; optional KeyType key = 1; optional ValueType value = 2; } repeated MapFieldEntry map_field = 1; Implementations may choose not to generate the map_entry=true message, but use a native map in the target language to hold the keys and values. The reflection APIs in such implementations still need to work as if the field is a repeated message field. NOTE: Do not set the option in .proto files. Always use the maps syntax instead. The option should only be implicitly set by the proto compiler parser.

• optional bool deprecated_legacy_json_field_conflicts = 11

  Enable the legacy handling of JSON field name conflicts. This lowercases and strips underscored from the fields before comparison in proto3 only. The new behavior takes `json_name` into account and applies to proto2 as well. This should only be used as a temporary measure against broken builds due to the change in behavior for JSON field name conflicts. TODO This is legacy behavior we plan to remove once downstream teams have had time to migrate.

• optional FeatureSet features = 12

  Any features defined in the specific edition.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Describes a method of a service.

Used in: ServiceDescriptorProto

• optional string name = 1

• optional string input_type = 2

  Input and output type names. These are resolved in the same way as FieldDescriptorProto.type_name, but must refer to a message type.

• optional string output_type = 3

• optional MethodOptions options = 4

• optional bool client_streaming = 5

  Identifies if client streams multiple client messages

• optional bool server_streaming = 6

  Identifies if server streams multiple server messages

Used in: MethodDescriptorProto

• optional bool deprecated = 33

  Is this method deprecated? Depending on the target platform, this can emit Deprecated annotations for the method, or it will be completely ignored; in the very least, this is a formalization for deprecating methods.

• optional MethodOptions.IdempotencyLevel idempotency_level = 34

• optional FeatureSet features = 35

  Any features defined in the specific edition.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Is this method side-effect-free (or safe in HTTP parlance), or idempotent, or neither? HTTP based RPC implementation may choose GET verb for safe methods, and PUT verb for idempotent methods instead of the default POST.

Used in: MethodOptions

• IDEMPOTENCY_UNKNOWN = 0

• NO_SIDE_EFFECTS = 1

  implies idempotent

• IDEMPOTENT = 2

  idempotent, but may have side effects

Describes a oneof.

Used in: DescriptorProto

• optional string name = 1

• optional OneofOptions options = 2

Used in: OneofDescriptorProto

• optional FeatureSet features = 1

  Any features defined in the specific edition.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Describes a service.

Used in: FileDescriptorProto

• optional string name = 1

• repeated MethodDescriptorProto method = 2

• optional ServiceOptions options = 3

Used in: ServiceDescriptorProto

• optional FeatureSet features = 34

  Any features defined in the specific edition.

• optional bool deprecated = 33

  Is this service deprecated? Depending on the target platform, this can emit Deprecated annotations for the service, or it will be completely ignored; in the very least, this is a formalization for deprecating services.

• repeated UninterpretedOption uninterpreted_option = 999

  The parser stores options it doesn't recognize here. See above.

Encapsulates information about the original source file from which a FileDescriptorProto was generated.

Used in: FileDescriptorProto

• repeated SourceCodeInfo.Location location = 1

  A Location identifies a piece of source code in a .proto file which corresponds to a particular definition. This information is intended to be useful to IDEs, code indexers, documentation generators, and similar tools. For example, say we have a file like: message Foo { optional string foo = 1; } Let's look at just the field definition: optional string foo = 1; ^ ^^ ^^ ^ ^^^ a bc de f ghi We have the following locations: span path represents [a,i) [ 4, 0, 2, 0 ] The whole field definition. [a,b) [ 4, 0, 2, 0, 4 ] The label (optional). [c,d) [ 4, 0, 2, 0, 5 ] The type (string). [e,f) [ 4, 0, 2, 0, 1 ] The name (foo). [g,h) [ 4, 0, 2, 0, 3 ] The number (1). Notes: - A location may refer to a repeated field itself (i.e. not to any particular index within it). This is used whenever a set of elements are logically enclosed in a single code segment. For example, an entire extend block (possibly containing multiple extension definitions) will have an outer location whose path refers to the "extensions" repeated field without an index. - Multiple locations may have the same path. This happens when a single logical declaration is spread out across multiple places. The most obvious example is the "extend" block again -- there may be multiple extend blocks in the same scope, each of which will have the same path. - A location's span is not always a subset of its parent's span. For example, the "extendee" of an extension declaration appears at the beginning of the "extend" block and is shared by all extensions within the block. - Just because a location's span is a subset of some other location's span does not mean that it is a descendant. For example, a "group" defines both a type and a field in a single declaration. Thus, the locations corresponding to the type and field and their components will overlap. - Code which tries to interpret locations should probably be designed to ignore those that it doesn't understand, as more types of locations could be recorded in the future.

Used in: SourceCodeInfo

• repeated int32 path = 1

  Identifies which part of the FileDescriptorProto was defined at this location. Each element is a field number or an index. They form a path from the root FileDescriptorProto to the place where the definition appears. For example, this path: [ 4, 3, 2, 7, 1 ] refers to: file.message_type(3) // 4, 3 .field(7) // 2, 7 .name() // 1 This is because FileDescriptorProto.message_type has field number 4: repeated DescriptorProto message_type = 4; and DescriptorProto.field has field number 2: repeated FieldDescriptorProto field = 2; and FieldDescriptorProto.name has field number 1: optional string name = 1; Thus, the above path gives the location of a field name. If we removed the last element: [ 4, 3, 2, 7 ] this path refers to the whole field declaration (from the beginning of the label to the terminating semicolon).

• repeated int32 span = 2

  Always has exactly three or four elements: start line, start column, end line (optional, otherwise assumed same as start line), end column. These are packed into a single field for efficiency. Note that line and column numbers are zero-based -- typically you will want to add 1 to each before displaying to a user.

• optional string leading_comments = 3

  If this SourceCodeInfo represents a complete declaration, these are any comments appearing before and after the declaration which appear to be attached to the declaration. A series of line comments appearing on consecutive lines, with no other tokens appearing on those lines, will be treated as a single comment. leading_detached_comments will keep paragraphs of comments that appear before (but not connected to) the current element. Each paragraph, separated by empty lines, will be one comment element in the repeated field. Only the comment content is provided; comment markers (e.g. //) are stripped out. For block comments, leading whitespace and an asterisk will be stripped from the beginning of each line other than the first. Newlines are included in the output. Examples: optional int32 foo = 1; // Comment attached to foo. // Comment attached to bar. optional int32 bar = 2; optional string baz = 3; // Comment attached to baz. // Another line attached to baz. // Comment attached to moo. // // Another line attached to moo. optional double moo = 4; // Detached comment for corge. This is not leading or trailing comments // to moo or corge because there are blank lines separating it from // both. // Detached comment for corge paragraph 2. optional string corge = 5; /* Block comment attached * to corge. Leading asterisks * will be removed. */ /* Block comment attached to * grault. */ optional int32 grault = 6; // ignored detached comments.

• optional string trailing_comments = 4

• repeated string leading_detached_comments = 6

A message representing a option the parser does not recognize. This only appears in options protos created by the compiler::Parser class. DescriptorPool resolves these when building Descriptor objects. Therefore, options protos in descriptor objects (e.g. returned by Descriptor::options(), or produced by Descriptor::CopyTo()) will never have UninterpretedOptions in them.

Used in: EnumOptions, EnumValueOptions, ExtensionRangeOptions, FieldOptions, FileOptions, MessageOptions, MethodOptions, OneofOptions, ServiceOptions

• repeated UninterpretedOption.NamePart name = 2

• optional string identifier_value = 3

  The value of the uninterpreted option, in whatever type the tokenizer identified it as during parsing. Exactly one of these should be set.

• optional uint64 positive_int_value = 4

• optional int64 negative_int_value = 5

• optional double double_value = 6

• optional bytes string_value = 7

• optional string aggregate_value = 8

The name of the uninterpreted option. Each string represents a segment in a dot-separated name. is_extension is true iff a segment represents an extension (denoted with parentheses in options specs in .proto files). E.g.,{ ["foo", false], ["bar.baz", true], ["moo", false] } represents "foo.(bar.baz).moo".

Used in: UninterpretedOption

• required string name_part = 1

• required bool is_extension = 2