protobuf-net

Provides support for common .NET types that do not have a direct representation in protobuf, using the definitions from bcl.proto

Creates a new instance of the specified type, bypassing the constructor.

The type to create The new instance If the platform does not support constructor-skipping

Writes a TimeSpan to a protobuf stream

Parses a TimeSpan from a protobuf stream

Parses a DateTime from a protobuf stream

Writes a DateTime to a protobuf stream

Parses a decimal from a protobuf stream

Writes a decimal to a protobuf stream

Writes a Guid to a protobuf stream

Parses a Guid from a protobuf stream

Reads an *implementation specific* bundled .NET object, including (as options) type-metadata, identity/re-use, etc.

Writes an *implementation specific* bundled .NET object, including (as options) type-metadata, identity/re-use, etc.

Optional behaviours that introduce .NET-specific functionality

No special behaviour

Enables full object-tracking/full-graph support.

Embeds the type information into the stream, allowing usage with types not known in advance.

If false, the constructor for the type is bypassed during deserialization, meaning any field initializers or other initialization code is skipped.

Should the object index be reserved, rather than creating an object promptly

Provides a simple buffer-based implementation of an extension object.

Provides addition capability for supporting unexpected fields during protocol-buffer serialization/deserialization. This allows for loss-less round-trip/merge, even when the data is not fully understood.

Requests a stream into which any unexpected fields can be persisted.

A new stream suitable for storing data.

Indicates that all unexpected fields have now been stored. The implementing class is responsible for closing the stream. If "commit" is not true the data may be discarded.

The stream originally obtained by BeginAppend. True if the append operation completed successfully.

Requests a stream of the unexpected fields previously stored.

A prepared stream of the unexpected fields.

Indicates that all unexpected fields have now been read. The implementing class is responsible for closing the stream.

The stream originally obtained by BeginQuery.

Requests the length of the raw binary stream; this is used when serializing sub-entities to indicate the expected size.

The length of the binary stream representing unexpected data.

Specifies a method on the root-contract in an hierarchy to be invoked before serialization.

Specifies a method on the root-contract in an hierarchy to be invoked after serialization.

Specifies a method on the root-contract in an hierarchy to be invoked before deserialization.

Specifies a method on the root-contract in an hierarchy to be invoked after deserialization.

Pushes a null reference onto the stack. Note that this should only be used to return a null (or set a variable to null); for null-tests use BranchIfTrue / BranchIfFalse.

Creates a new "using" block (equivalent) around a variable; the variable must exist, and note that (unlike in C#) it is the variables *final* value that gets disposed. If you need *original* disposal, copy your variable first. It is the callers responsibility to ensure that the variable's scope fully-encapsulates the "using"; if not, the variable may be re-used (and thus re-assigned) unexpectedly.

Sub-format to use when serializing/deserializing data

Uses the default encoding for the data-type.

When applied to signed integer-based data (including Decimal), this indicates that zigzag variant encoding will be used. This means that values with small magnitude (regardless of sign) take a small amount of space to encode.

When applied to signed integer-based data (including Decimal), this indicates that two's-complement variant encoding will be used. This means that any -ve number will take 10 bytes (even for 32-bit), so should only be used for compatibility.

When applied to signed integer-based data (including Decimal), this indicates that a fixed amount of space will be used.

When applied to a sub-message, indicates that the value should be treated as group-delimited.

Simple base class for supporting unexpected fields allowing for loss-less round-tips/merge, even if the data is not understod. The additional fields are (by default) stored in-memory in a buffer.

As an example of an alternative implementation, you might choose to use the file system (temporary files) as the back-end, tracking only the paths [such an object would ideally be IDisposable and use a finalizer to ensure that the files are removed].

Indicates that the implementing type has support for protocol-buffer extensions.

Can be implemented by deriving from Extensible.

Retrieves the extension object for the current instance, optionally creating it if it does not already exist.

Should a new extension object be created if it does not already exist? The extension object if it exists (or was created), or null if the extension object does not exist or is not available. The createIfMissing argument is false during serialization, and true during deserialization upon encountering unexpected fields.

Retrieves the extension object for the current instance, optionally creating it if it does not already exist.

Provides a simple, default implementation for extension support, optionally creating it if it does not already exist. Designed to be called by classes implementing .

Should a new extension object be created if it does not already exist? The extension field to check (and possibly update). The extension object if it exists (or was created), or null if the extension object does not exist or is not available. The createIfMissing argument is false during serialization, and true during deserialization upon encountering unexpected fields.

Appends the value as an additional (unexpected) data-field for the instance. Note that for non-repeated sub-objects, this equates to a merge operation; for repeated sub-objects this adds a new instance to the set; for simple values the new value supercedes the old value.

Note that appending a value does not remove the old value from the stream; avoid repeatedly appending values for the same field. The type of the value to append. The extensible object to append the value to. The field identifier; the tag should not be defined as a known data-field for the instance. The value to append.

Note that appending a value does not remove the old value from the stream; avoid repeatedly appending values for the same field. The data-type of the field. The data-format to use when encoding the value. The extensible object to append the value to. The field identifier; the tag should not be defined as a known data-field for the instance. The value to append.

Queries an extensible object for an additional (unexpected) data-field for the instance. The value returned is the composed value after merging any duplicated content; if the value is "repeated" (a list), then use GetValues instead.

The data-type of the field. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The effective value of the field, or the default value if not found.

The data-type of the field. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. The effective value of the field, or the default value if not found.

Queries an extensible object for an additional (unexpected) data-field for the instance. The value returned (in "value") is the composed value after merging any duplicated content; if the value is "repeated" (a list), then use GetValues instead.

The data-type of the field. The effective value of the field, or the default value if not found. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. True if data for the field was present, false otherwise.

The data-type of the field. The effective value of the field, or the default value if not found. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. Allow tags that are present as part of the definition; for example, to query unknown enum values. True if data for the field was present, false otherwise.

Queries an extensible object for an additional (unexpected) data-field for the instance. Each occurrence of the field is yielded separately, making this usage suitable for "repeated" (list) fields.

The extended data is processed lazily as the enumerator is iterated. The data-type of the field. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. An enumerator that yields each occurrence of the field.

Queries an extensible object for an additional (unexpected) data-field for the instance. Each occurrence of the field is yielded separately, making this usage suitable for "repeated" (list) fields.

The extended data is processed lazily as the enumerator is iterated. The data-type of the field. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. An enumerator that yields each occurrence of the field.

The data-type of the field. The model to use for configuration. The effective value of the field, or the default value if not found. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. Allow tags that are present as part of the definition; for example, to query unknown enum values. True if data for the field was present, false otherwise.

Queries an extensible object for an additional (unexpected) data-field for the instance. Each occurrence of the field is yielded separately, making this usage suitable for "repeated" (list) fields.

The extended data is processed lazily as the enumerator is iterated. The model to use for configuration. The data-type of the field. The extensible object to obtain the value from. The field identifier; the tag should not be defined as a known data-field for the instance. The data-format to use when decoding the value. An enumerator that yields each occurrence of the field.

Note that appending a value does not remove the old value from the stream; avoid repeatedly appending values for the same field. The model to use for configuration. The data-format to use when encoding the value. The extensible object to append the value to. The field identifier; the tag should not be defined as a known data-field for the instance. The value to append.

This class acts as an internal wrapper allowing us to do a dynamic methodinfo invoke; an't put into Serializer as don't want on public API; can't put into Serializer<T> since we need to invoke accross classes, which isn't allowed in Silverlight)

All this does is call GetExtendedValuesTyped with the correct type for "instance"; this ensures that we don't get issues with subclasses declaring conflicting types - the caller must respect the fields defined for the type they pass in.

Not all frameworks are created equal (fx1.1 vs fx2.0, micro-framework, compact-framework, silverlight, etc). This class simply wraps up a few things that would otherwise make the real code unnecessarily messy, providing fallback implementations if necessary.

Intended to be a direct map to regular TypeCode, but: - with missing types - existing on WinRT

Specifies the method used to infer field tags for members of the type under consideration. Tags are deduced using the invariant alphabetic sequence of the members' names; this makes implicit field tags very brittle, and susceptible to changes such as field names (normally an isolated change).

No members are serialized implicitly; all members require a suitable attribute such as [ProtoMember]. This is the recmomended mode for most scenarios.

Public properties and fields are eligible for implicit serialization; this treats the public API as a contract. Ordering beings from ImplicitFirstTag.

Public and non-public fields are eligible for implicit serialization; this acts as a state/implementation serializer. Ordering beings from ImplicitFirstTag.

Represents the set of serialization callbacks to be used when serializing/deserializing a type.

Called before serializing an instance

Called before deserializing an instance

Called after serializing an instance

Called after deserializing an instance

True if any callback is set, else False

Represents a type at runtime for use with protobuf, allowing the field mappings (etc) to be defined

Get the name of the type being represented

Adds a known sub-type to the inheritance model

Assigns the callbacks to use during serialiation/deserialization.

The method (or null) called before serialization begins. The method (or null) called when serialization is complete. The method (or null) called before deserialization begins (or when a new instance is created during deserialization). The method (or null) called when deserialization is complete. The set of callbacks.

Assigns the callbacks to use during serialiation/deserialization.

The name of the method (or null) called before serialization begins. The name of the method (or null) called when serialization is complete. The name of the method (or null) called before deserialization begins (or when a new instance is created during deserialization). The name of the method (or null) called when deserialization is complete. The set of callbacks.

Designate a factory-method to use to create instances of this type

Throws an exception if the type has been made immutable

Adds a member (by name) to the MetaType

Adds a member (by name) to the MetaType, returning the ValueMember rather than the fluent API. This is otherwise identical to Add.

Adds a member (by name) to the MetaType

Performs serialization of this type via a surrogate; all other serialization options are ignored and handled by the surrogate's configuration.

Adds a set of members (by name) to the MetaType

Adds a member (by name) to the MetaType

Adds a member (by name) to the MetaType, including an itemType and defaultType for representing lists

Adds a member (by name) to the MetaType, including an itemType and defaultType for representing lists, returning the ValueMember rather than the fluent API. This is otherwise identical to Add.

Returns the ValueMember instances associated with this type

Returns the SubType instances associated with this type

Compiles the serializer for this type; this is *not* a full standalone compile, but can significantly boost performance while allowing additional types to be added.

An in-place compile can access non-public types / members

Gets the base-type for this type

When used to compile a model, should public serialization/deserialzation methods be included for this type?

Should this type be treated as a reference by default?

Indicates whether the current type has defined callbacks

Indicates whether the current type has defined subtypes

Returns the set of callbacks defined for this type

Gets or sets the name of this contract.

The runtime type that the meta-type represents

Gets or sets whether the type should use a parameterless constructor (the default), or whether the type should skip the constructor completely. This option is not supported on compact-framework.

The concrete type to create when a new instance of this type is needed; this may be useful when dealing with dynamic proxies, or with interface-based APIs

Returns the ValueMember that matchs a given field number, or null if not found

Returns the ValueMember that matchs a given member (property/field), or null if not found

Gets or sets a value indicating that an enum should be treated directly as an int/short/etc, rather than enforcing .proto enum rules. This is useful *in particul* for [Flags] enums.

Gets or sets a value indicating that this type should NOT be treated as a list, even if it has familiar list-like characteristics (enumerable, add, etc)

Provides protobuf serialization support for a number of types that can be defined at runtime

Provides protobuf serialization support for a number of types

Resolve a System.Type to the compiler-specific type

This is the more "complete" version of Serialize, which handles single instances of mapped types. The value is written as a complete field, including field-header and (for sub-objects) a length-prefix In addition to that, this provides support for: - basic values; individual int / string / Guid / etc - IEnumerable sequences of any type handled by TrySerializeAuxiliaryType

Writes a protocol-buffer representation of the given instance to the supplied stream.

The existing instance to be serialized (cannot be null). The destination stream to write to.

Writes a protocol-buffer representation of the given instance to the supplied stream.

The existing instance to be serialized (cannot be null). The destination stream to write to. Additional information about this serialization operation.

Writes a protocol-buffer representation of the given instance to the supplied writer.

The existing instance to be serialized (cannot be null). The destination writer to write to.

Applies a protocol-buffer stream to an existing instance (or null), using length-prefixed data - useful with network IO.

The type being merged. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). How to encode the length prefix. The tag used as a prefix to each record (only used with base-128 style prefixes). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Applies a protocol-buffer stream to an existing instance (or null), using length-prefixed data - useful with network IO.

The type being merged. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). How to encode the length prefix. The tag used as a prefix to each record (only used with base-128 style prefixes). Used to resolve types on a per-field basis. The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Applies a protocol-buffer stream to an existing instance (or null), using length-prefixed data - useful with network IO.

The type being merged. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). How to encode the length prefix. The tag used as a prefix to each record (only used with base-128 style prefixes). Used to resolve types on a per-field basis. Returns the number of bytes consumed by this operation (includes length-prefix overheads and any skipped data). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Reads a sequence of consecutive length-prefixed items from a stream, using either base-128 or fixed-length prefixes. Base-128 prefixes with a tag are directly comparable to serializing multiple items in succession (use the tag to emulate the implicit behavior when serializing a list/array). When a tag is specified, any records with different tags are silently omitted. The tag is ignored. The tag is ignores for fixed-length prefixes.

The binary stream containing the serialized records. The prefix style used in the data. The tag of records to return (if non-positive, then no tag is expected and all records are returned). On a field-by-field basis, the type of object to deserialize (can be null if "type" is specified). The type of object to deserialize (can be null if "resolver" is specified). The sequence of deserialized objects.

The type of object to deserialize. The binary stream containing the serialized records. The prefix style used in the data. The tag of records to return (if non-positive, then no tag is expected and all records are returned). The sequence of deserialized objects.

Writes a protocol-buffer representation of the given instance to the supplied stream, with a length-prefix. This is useful for socket programming, as DeserializeWithLengthPrefix can be used to read the single object back from an ongoing stream.

The type being serialized. The existing instance to be serialized (cannot be null). How to encode the length prefix. The destination stream to write to. The tag used as a prefix to each record (only used with base-128 style prefixes).

Applies a protocol-buffer stream to an existing instance (which may be null).

The type (including inheritance) to consider. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Applies a protocol-buffer stream to an existing instance (which may be null).

The type (including inheritance) to consider. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). The number of bytes to consume. The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Applies a protocol-buffer stream to an existing instance (which may be null).

The type (including inheritance) to consider. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). The number of bytes to consume (or -1 to read to the end of the stream). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance. Additional information about this serialization operation.

Applies a protocol-buffer reader to an existing instance (which may be null).

The type (including inheritance) to consider. The existing instance to be modified (can be null). The reader to apply to the instance (cannot be null). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

This is the more "complete" version of Deserialize, which handles single instances of mapped types. The value is read as a complete field, including field-header and (for sub-objects) a length-prefix..kmc In addition to that, this provides support for: - basic values; individual int / string / Guid / etc - IList sets of any type handled by TryDeserializeAuxiliaryType

Creates a new runtime model, to which the caller can add support for a range of types. A model can be used "as is", or can be compiled for optimal performance.

Applies common proxy scenarios, resolving the actual type to consider

Indicates whether the supplied type is explicitly modelled by the model

Provides the key that represents a given type in the current model. The type is also normalized for proxies at the same time.

Provides the key that represents a given type in the current model.

Writes a protocol-buffer representation of the given instance to the supplied stream.

Represents the type (including inheritance) to consider. The existing instance to be serialized (cannot be null). The destination stream to write to.

Applies a protocol-buffer stream to an existing instance (which may be null).

Represents the type (including inheritance) to consider. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Create a deep clone of the supplied instance; any sub-items are also cloned.

Indicates that while an inheritance tree exists, the exact type encountered was not specified in that hierarchy and cannot be processed.

Indicates that the given type was not expected, and cannot be processed.

Indicates that the given type cannot be constructed; it may still be possible to deserialize into existing instances.

Returns true if the type supplied is either a recognised contract type, or a *list* of a recognised contract type.

Note that primitives always return false, even though the engine will, if forced, try to serialize such True if this type is recognised as a serializable entity, else false

Returns true if the type supplied is a basic type with inbuilt handling, a recognised contract type, or a *list* of a basic / contract type.

Returns true if the type supplied is a basic type with inbuilt handling, or a *list* of a basic type with inbuilt handling

Suggest a .proto definition for the given type

The type to generate a .proto definition for, or null to generate a .proto that represents the entire model The .proto definition as a string

Creates a new IFormatter that uses protocol-buffer [de]serialization.

A new IFormatter to be used during [de]serialization. The type of object to be [de]deserialized by the formatter.

Used to provide custom services for writing and parsing type names when using dynamic types. Both parsing and formatting are provided on a single API as it is essential that both are mapped identically at all times.

Indicates the type of callback to be used

Invoked before an object is serialized

Invoked after an object is serialized

Invoked before an object is deserialized (or when a new instance is created)

Invoked after an object is deserialized

Returns a sequence of the Type instances that can be processed by this model.

Suggest a .proto definition for the given type

The type to generate a .proto definition for, or null to generate a .proto that represents the entire model The .proto definition as a string

Adds support for an additional type in this model, optionally appplying inbuilt patterns. If the type is already known to the model, the existing type is returned **without** applying any additional behaviour.

Inbuilt patterns include: [ProtoContract]/[ProtoMember(n)] [DataContract]/[DataMember(Order=n)] [XmlType]/[XmlElement(Order=n)] [On{Des|S}erializ{ing|ed}] ShouldSerialize*/*Specified The type to be supported Whether to apply the inbuilt configuration patterns (via attributes etc), or just add the type with no additional configuration (the type must then be manually configured). The MetaType representing this type, allowing further configuration.

Verifies that the model is still open to changes; if not, an exception is thrown

Prevents further changes to this model

Provides the key that represents a given type in the current model.

Writes a protocol-buffer representation of the given instance to the supplied stream.

Represents the type (including inheritance) to consider. The existing instance to be serialized (cannot be null). The destination stream to write to.

Applies a protocol-buffer stream to an existing instance (which may be null).

Compiles the serializers individually; this is *not* a full standalone compile, but can significantly boost performance while allowing additional types to be added.

An in-place compile can access non-public types / members

Fully compiles the current model into a static-compiled model instance

A full compilation is restricted to accessing public types / members An instance of the newly created compiled type-model

Fully compiles the current model into a static-compiled serialization dll (the serialization dll still requires protobuf-net for support services).

A full compilation is restricted to accessing public types / members The name of the TypeModel class to create The path for the new dll An instance of the newly created compiled type-model

Fully compiles the current model into a static-compiled serialization dll (the serialization dll still requires protobuf-net for support services).

A full compilation is restricted to accessing public types / members An instance of the newly created compiled type-model

Designate a factory-method to use to create instances of any type; note that this only affect types seen by the serializer *after* setting the factory.

Global default that enables/disables automatic tag generation based on the existing name / order of the defined members. See for usage and important warning / explanation. You must set the global default before attempting to serialize/deserialize any impacted type.

Global default that determines whether types are considered serializable if they have [DataContract] / [XmlType]. With this enabled, ONLY types marked as [ProtoContract] are added automatically.

Global switch that enables or disables the implicit handling of "zero defaults"; meanning: if no other default is specified, it assumes bools always default to false, integers to zero, etc. If this is disabled, no such assumptions are made and only *explicit* default values are processed. This is enabled by default to preserve similar logic to v1.

Global switch that determines whether types with a .ToString() and a Parse(string) should be serialized as strings.

The default model, used to support ProtoBuf.Serializer

Obtains the MetaType associated with a given Type for the current model, allowing additional configuration.

Should serializers be compiled on demand? It may be useful to disable this for debugging purposes.

Should support for unexpected types be added automatically? If false, an exception is thrown when unexpected types are encountered.

The amount of time to wait if there are concurrent metadata access operations

If a lock-contention is detected, this event signals the *owner* of the lock responsible for the blockage, indicating what caused the problem; this is only raised if the lock-owning code successfully completes.

Represents configuration options for compiling a model to a standalone assembly.

Import framework options from an existing type

The TargetFrameworkAttribute FrameworkName value to burn into the generated assembly

The TargetFrameworkAttribute FrameworkDisplayName value to burn into the generated assembly

The name of the TypeModel class to create

The path for the new dll

The runtime version for the generated assembly

The acecssibility of the generated serializer

Type accessibility

Available to all callers

Available to all callers in the same assembly, or assemblies specified via [InternalsVisibleTo(...)]

Contains the stack-trace of the owning code when a lock-contention scenario is detected

The stack-trace of the code that owned the lock when a lock-contention scenario occurred

Event-type that is raised when a lock-contention scenario is detected

Represents an inherited type in a type hierarchy.

Creates a new SubType instance.

The field-number that is used to encapsulate the data (as a nested message) for the derived dype. The sub-type to be considered. Specific encoding style to use; in particular, Grouped can be used to avoid buffering, but is not the default.

The field-number that is used to encapsulate the data (as a nested message) for the derived dype.

The sub-type to be considered.

Event arguments needed to perform type-formatting functions; this could be resolving a Type to a string suitable for serialization, or could be requesting a Type from a string. If no changes are made, a default implementation will be used (from the assembly-qualified names).

The type involved in this map; if this is initially null, a Type is expected to be provided for the string in FormattedName.

The formatted-name involved in this map; if this is initially null, a formatted-name is expected from the type in Type.

Delegate type used to perform type-formatting functions; the sender originates as the type-model.

Represents a member (property/field) that is mapped to a protobuf field

Creates a new ValueMember instance

Specifies methods for working with optional data members.

Provides a method (null for none) to query whether this member should be serialized; it must be of the form "bool {Method}()". The member is only serialized if the method returns true. Provides a method (null for none) to indicate that a member was deserialized; it must be of the form "void {Method}(bool)", and will be called with "true" when data is found.

The number that identifies this member in a protobuf stream

Gets the member (field/property) which this member relates to.

Within a list / array / etc, the type of object for each item in the list (especially useful with ArrayList)

The underlying type of the member

For abstract types (IList etc), the type of concrete object to create (if required)

The type the defines the member

The default value of the item (members with this value will not be serialized)

Specifies the rules used to process the field; this is used to determine the most appropriate wite-type, but also to describe subtypes within that wire-type (such as SignedVariant)

Indicates whether this field should follow strict encoding rules; this means (for example) that if a "fixed32" is encountered when "variant" is defined, then it will fail (throw an exception) when parsing. Note that when serializing the defined type is always used.

Indicates whether this field should use packed encoding (which can save lots of space for repeated primitive values). This option only applies to list/array data of primitive types (int, double, etc).

Indicates whether this field should *repace* existing values (the default is false, meaning *append*). This option only applies to list/array data.

Indicates whether this field is mandatory.

Enables full object-tracking/full-graph support.

Embeds the type information into the stream, allowing usage with types not known in advance.

Gets the logical name for this member in the schema (this is not critical for binary serialization, but may be used when inferring a schema).

Should lists have extended support for null values? Note this makes the serialization less efficient.

Specifies the type of prefix that should be applied to messages.

No length prefix is applied to the data; the data is terminated only be the end of the stream.

A base-128 length prefix is applied to the data (efficient for short messages).

A fixed-length (little-endian) length prefix is applied to the data (useful for compatibility).

A fixed-length (big-endian) length prefix is applied to the data (useful for compatibility).

Indicates that a type is defined for protocol-buffer serialization.

Gets or sets the defined name of the type.

Gets or sets the fist offset to use with implicit field tags; only uesd if ImplicitFields is set.

If specified, alternative contract markers (such as markers for XmlSerailizer or DataContractSerializer) are ignored.

If specified, do NOT treat this type as a list, even if it looks like one.

Gets or sets the mechanism used to automatically infer field tags for members. This option should be used in advanced scenarios only. Please review the important notes against the ImplicitFields enumeration.

Enables/disables automatic tag generation based on the existing name / order of the defined members. This option is not used for members marked with ProtoMemberAttribute, as intended to provide compatibility with WCF serialization. WARNING: when adding new fields you must take care to increase the Order for new elements, otherwise data corruption may occur.

If not explicitly specified, the default is assumed from Serializer.GlobalOptions.InferTagFromName.

Has a InferTagFromName value been explicitly set? if not, the default from the type-model is assumed.

Specifies an offset to apply to [DataMember(Order=...)] markers; this is useful when working with mex-generated classes that have a different origin (usually 1 vs 0) than the original data-contract. This value is added to the Order of each member.

If true, the constructor for the type is bypassed during deserialization, meaning any field initializers or other initialization code is skipped.

Should this type be treated as a reference by default? Please also see the implications of this, as recorded on ProtoMemberAttribute.AsReference

Applies only to enums (not to DTO classes themselves); gets or sets a value indicating that an enum should be treated directly as an int/short/etc, rather than enforcing .proto enum rules. This is useful *in particul* for [Flags] enums.

Has a EnumPassthru value been explicitly set?

Used to define protocol-buffer specific behavior for enumerated values.

Indicates whether this instance has a customised value mapping

true if a specific value is set

Gets or sets the specific value to use for this enum during serialization.

Gets or sets the defined name of the enum, as used in .proto (this name is not used during serialization).

Indicates an error during serialization/deserialization of a proto stream.

Creates a new ProtoException instance.

Indicates that a member should be excluded from serialization; this is only normally used when using implict fields.

Indicates that a member should be excluded from serialization; this is only normally used when using implict fields. This allows ProtoIgnoreAttribute usage even for partial classes where the individual members are not under direct control.

Creates a new ProtoPartialIgnoreAttribute instance.

Specifies the member to be ignored.

The name of the member to be ignored.

Indicates the known-types to support for an individual message. This serializes each level in the hierarchy as a nested message to retain wire-compatibility with other protocol-buffer implementations.

Creates a new instance of the ProtoIncludeAttribute.

The unique index (within the type) that will identify this data. The additional type to serialize/deserialize.

Creates a new instance of the ProtoIncludeAttribute.

The unique index (within the type) that will identify this data. The additional type to serialize/deserialize.

Gets the unique index (within the type) that will identify this data.

Gets the additional type to serialize/deserialize.

Specifies whether the inherited sype's sub-message should be written with a length-prefix (default), or with group markers.

Declares a member to be used in protocol-buffer serialization, using the given Tag. A DataFormat may be used to optimise the serialization format (for instance, using zigzag encoding for negative numbers, or fixed-length encoding for large values.

Compare with another ProtoMemberAttribute for sorting purposes

Creates a new ProtoMemberAttribute instance.

Specifies the unique tag used to identify this member within the type.

Gets or sets the original name defined in the .proto; not used during serialization.

Gets or sets the data-format to be used when encoding this value.

Gets the unique tag used to identify this member within the type.

Gets or sets a value indicating whether this member is mandatory.

Gets a value indicating whether this member is packed. This option only applies to list/array data of primitive types (int, double, etc).

Indicates whether this field should *repace* existing values (the default is false, meaning *append*). This option only applies to list/array data.

Enables full object-tracking/full-graph support.

Embeds the type information into the stream, allowing usage with types not known in advance.

Gets or sets a value indicating whether this member is packed (lists/arrays).

Additional (optional) settings that control serialization of members

Default; no additional options

Indicates that repeated elements should use packed (length-prefixed) encoding

Indicates that the given item is required

Enables full object-tracking/full-graph support

Embeds the type information into the stream, allowing usage with types not known in advance

Indicates whether this field should *repace* existing values (the default is false, meaning *append*). This option only applies to list/array data.

Determines whether the types AsReferenceDefault value is used, or whether this member's AsReference should be used

Declares a member to be used in protocol-buffer serialization, using the given Tag and MemberName. This allows ProtoMemberAttribute usage even for partial classes where the individual members are not under direct control. A DataFormat may be used to optimise the serialization format (for instance, using zigzag encoding for negative numbers, or fixed-length encoding for large values.

Creates a new ProtoMemberAttribute instance.

Specifies the unique tag used to identify this member within the type. Specifies the member to be serialized.

The name of the member to be serialized.

A stateful reader, used to read a protobuf stream. Typical usage would be (sequentially) to call ReadFieldHeader and (after matching the field) an appropriate Read* method.

Creates a new reader against a stream

The source stream The model to use for serialization; this can be null, but this will impair the ability to deserialize sub-objects Additional context about this serialization operation

Creates a new reader against a stream

The source stream The model to use for serialization; this can be null, but this will impair the ability to deserialize sub-objects Additional context about this serialization operation The number of bytes to read, or -1 to read until the end of the stream

Releases resources used by the reader, but importantly does not Dispose the underlying stream; in many typical use-cases the stream is used for different processes, so it is assumed that the consumer will Dispose their stream separately.

Reads an unsigned 32-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64

Reads a signed 16-bit integer from the stream: Variant, Fixed32, Fixed64, SignedVariant

Reads an unsigned 16-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64

Reads an unsigned 8-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64

Reads a signed 8-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Reads a signed 32-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Reads a signed 64-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Reads a string from the stream (using UTF8); supported wire-types: String

Throws an exception indication that the given value cannot be mapped to an enum.

Reads a double-precision number from the stream; supported wire-types: Fixed32, Fixed64

Reads (merges) a sub-message from the stream, internally calling StartSubItem and EndSubItem, and (in between) parsing the message in accordance with the model associated with the reader

Makes the end of consuming a nested message in the stream; the stream must be either at the correct EndGroup marker, or all fields of the sub-message must have been consumed (in either case, this means ReadFieldHeader should return zero)

Begins consuming a nested message in the stream; supported wire-types: StartGroup, String

The token returned must be help and used when callining EndSubItem

Reads a field header from the stream, setting the wire-type and retuning the field number. If no more fields are available, then 0 is returned. This methods respects sub-messages.

Looks ahead to see whether the next field in the stream is what we expect (typically; what we've just finished reading - for example ot read successive list items)

Compares the streams current wire-type to the hinted wire-type, updating the reader if necessary; for example, a Variant may be updated to SignedVariant. If the hinted wire-type is unrelated then no change is made.

Verifies that the stream's current wire-type is as expected, or a specialized sub-type (for example, SignedVariant) - in which case the current wire-type is updated. Otherwise an exception is thrown.

Discards the data for the current field.

Reads an unsigned 64-bit integer from the stream; supported wire-types: Variant, Fixed32, Fixed64

Reads a single-precision number from the stream; supported wire-types: Fixed32, Fixed64

Reads a boolean value from the stream; supported wire-types: Variant, Fixed32, Fixed64

Reads a byte-sequence from the stream, appending them to an existing byte-sequence (which can be null); supported wire-types: String

Reads the length-prefix of a message from a stream without buffering additional data, allowing a fixed-length reader to be created.

Reads a little-endian encoded integer. An exception is thrown if the data is not all available.

Reads a big-endian encoded integer. An exception is thrown if the data is not all available.

Reads a varint encoded integer. An exception is thrown if the data is not all available.

Reads a string (of a given lenth, in bytes) directly from the source into a pre-existing buffer. An exception is thrown if the data is not all available.

Reads a given number of bytes directly from the source. An exception is thrown if the data is not all available.

Reads a string (of a given lenth, in bytes) directly from the source. An exception is thrown if the data is not all available.

Reads the length-prefix of a message from a stream without buffering additional data, allowing a fixed-length reader to be created.

The number of bytes consumed; 0 if no data available

Copies the current field into the instance as extension data

Indicates whether the reader still has data remaining in the current sub-item, additionally setting the wire-type for the next field if there is more data. This is used when decoding packed data.

Utility method, not intended for public use; this helps maintain the root object is complex scenarios

Reads a Type from the stream, using the model's DynamicTypeFormatting if appropriate; supported wire-types: String

Merge two objects using the details from the current reader; this is used to change the type of objects when an inheritance relationship is discovered later than usual during deserilazation.

Gets the number of the field being processed.

Indicates the underlying proto serialization format on the wire.

Gets / sets a flag indicating whether strings should be checked for repetition; if true, any repeated UTF-8 byte sequence will result in the same String instance, rather than a second instance of the same string. Enabled by default. Note that this uses a custom interner - the system-wide string interner is not used.

Addition information about this deserialization operation.

Returns the position of the current reader (note that this is not necessarily the same as the position in the underlying stream, if multiple readers are used on the same stream)

Get the TypeModel associated with this reader

Represents an output stream for writing protobuf data. Why is the API backwards (static methods with writer arguments)? See: http://marcgravell.blogspot.com/2010/03/last-will-be-first-and-first-will-be.html

Write an encapsulated sub-object, using the supplied unique key (reprasenting a type).

The object to write. The key that uniquely identifies the type within the model. The destination.

Write an encapsulated sub-object, using the supplied unique key (reprasenting a type) - but the caller is asserting that this relationship is non-recursive; no recursion check will be performed.

The object to write. The key that uniquely identifies the type within the model. The destination.

Writes a field-header, indicating the format of the next data we plan to write.

Writes a byte-array to the stream; supported wire-types: String

Indicates the start of a nested record.

The instance to write. The destination. A token representing the state of the stream; this token is given to EndSubItem.

Indicates the end of a nested record.

The token obtained from StartubItem. The destination.

Creates a new writer against a stream

The destination stream The model to use for serialization; this can be null, but this will impair the ability to serialize sub-objects Additional context about this serialization operation

Flushes data to the underlying stream, and releases any resources. The underlying stream is *not* disposed by this operation.

Writes any buffered data (if possible) to the underlying stream.

The writer to flush It is not always possible to fully flush, since some sequences may require values to be back-filled into the byte-stream.

Writes an unsigned 32-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64

Writes a string to the stream; supported wire-types: String

Writes an unsigned 64-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64

Writes a signed 64-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Writes an unsigned 16-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64

Writes a signed 16-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Writes an unsigned 16-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64

Writes an unsigned 8-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64

Writes a signed 8-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Writes a signed 32-bit integer to the stream; supported wire-types: Variant, Fixed32, Fixed64, SignedVariant

Writes a double-precision number to the stream; supported wire-types: Fixed32, Fixed64

Writes a single-precision number to the stream; supported wire-types: Fixed32, Fixed64

Throws an exception indicating that the given enum cannot be mapped to a serialized value.

Writes a boolean to the stream; supported wire-types: Variant, Fixed32, Fixed64

Copies any extension data stored for the instance to the underlying stream

Used for packed encoding; indicates that the next field should be skipped rather than a field header written. Note that the field number must match, else an exception is thrown when the attempt is made to write the (incorrect) field. The wire-type is taken from the subsequent call to WriteFieldHeader. Only primitive types can be packed.

Specifies a known root object to use during reference-tracked serialization

Writes a Type to the stream, using the model's DynamicTypeFormatting if appropriate; supported wire-types: String

Addition information about this serialization operation.

Get the TypeModel associated with this writer

Additional information about a serialization operation

Convert a SerializationContext to a StreamingContext

Convert a StreamingContext to a SerializationContext

Gets or sets a user-defined object containing additional information about this serialization/deserialization operation.

A default SerializationContext, with minimal information.

Gets or sets the source or destination of the transmitted data.

Provides protocol-buffer serialization capability for concrete, attributed types. This is a *default* model, but custom serializer models are also supported.

Protocol-buffer serialization is a compact binary format, designed to take advantage of sparse data and knowledge of specific data types; it is also extensible, allowing a type to be deserialized / merged even if some data is not recognised.

The field number that is used as a default when serializing/deserializing a list of objects. The data is treated as repeated message with field number 1.

Suggest a .proto definition for the given type

The type to generate a .proto definition for The .proto definition as a string

Create a deep clone of the supplied instance; any sub-items are also cloned.

Applies a protocol-buffer stream to an existing instance.

The type being merged. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Creates a new instance from a protocol-buffer stream

The type to be created. The binary stream to apply to the new instance (cannot be null). A new, initialized instance.

Writes a protocol-buffer representation of the given instance to the supplied stream.

The existing instance to be serialized (cannot be null). The destination stream to write to.

Serializes a given instance and deserializes it as a different type; this can be used to translate between wire-compatible objects (where two .NET types represent the same data), or to promote/demote a type through an inheritance hierarchy.

No assumption of compatibility is made between the types. The type of the object being copied. The type of the new object to be created. The existing instance to use as a template. A new instane of type TNewType, with the data from TOldType.

Writes a protocol-buffer representation of the given instance to the supplied SerializationInfo.

The type being serialized. The existing instance to be serialized (cannot be null). The destination SerializationInfo to write to.

Writes a protocol-buffer representation of the given instance to the supplied SerializationInfo.

The type being serialized. The existing instance to be serialized (cannot be null). The destination SerializationInfo to write to. Additional information about this serialization operation.

Writes a protocol-buffer representation of the given instance to the supplied XmlWriter.

The type being serialized. The existing instance to be serialized (cannot be null). The destination XmlWriter to write to.

Applies a protocol-buffer from an XmlReader to an existing instance.

The type being merged. The existing instance to be modified (cannot be null). The XmlReader containing the data to apply to the instance (cannot be null).

Applies a protocol-buffer from a SerializationInfo to an existing instance.

The type being merged. The existing instance to be modified (cannot be null). The SerializationInfo containing the data to apply to the instance (cannot be null).

Applies a protocol-buffer from a SerializationInfo to an existing instance.

The type being merged. The existing instance to be modified (cannot be null). The SerializationInfo containing the data to apply to the instance (cannot be null). Additional information about this serialization operation.

Precompiles the serializer for a given type.

Creates a new IFormatter that uses protocol-buffer [de]serialization.

The type of object to be [de]deserialized by the formatter. A new IFormatter to be used during [de]serialization.

Creates a new instance from a protocol-buffer stream that has a length-prefix on data (to assist with network IO).

The type to be created. The binary stream to apply to the new instance (cannot be null). How to encode the length prefix. A new, initialized instance.

Creates a new instance from a protocol-buffer stream that has a length-prefix on data (to assist with network IO).

The type to be created. The binary stream to apply to the new instance (cannot be null). How to encode the length prefix. The expected tag of the item (only used with base-128 prefix style). A new, initialized instance.

Applies a protocol-buffer stream to an existing instance, using length-prefixed data - useful with network IO.

The type being merged. The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). How to encode the length prefix. The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Writes a protocol-buffer representation of the given instance to the supplied stream, with a length-prefix. This is useful for socket programming, as DeserializeWithLengthPrefix/MergeWithLengthPrefix can be used to read the single object back from an ongoing stream.

The type being serialized. The existing instance to be serialized (cannot be null). How to encode the length prefix. The destination stream to write to.

Writes a protocol-buffer representation of the given instance to the supplied stream, with a length-prefix. This is useful for socket programming, as DeserializeWithLengthPrefix/MergeWithLengthPrefix can be used to read the single object back from an ongoing stream.

Indicates the number of bytes expected for the next message.

The stream containing the data to investigate for a length. The algorithm used to encode the length. The length of the message, if it could be identified. True if a length could be obtained, false otherwise.

Indicates the number of bytes expected for the next message.

The buffer containing the data to investigate for a length. The offset of the first byte to read from the buffer. The number of bytes to read from the buffer. The algorithm used to encode the length. The length of the message, if it could be identified. True if a length could be obtained, false otherwise.

Releases any internal buffers that have been reserved for efficiency; this does not affect any serialization operations; simply: it can be used (optionally) to release the buffers for garbage collection (at the expense of having to re-allocate a new buffer for the next operation, rather than re-use prior buffers).

Provides non-generic access to the default serializer.

Create a deep clone of the supplied instance; any sub-items are also cloned.

Writes a protocol-buffer representation of the given instance to the supplied stream.

The existing instance to be serialized (cannot be null). The destination stream to write to.

Creates a new instance from a protocol-buffer stream

The type to be created. The binary stream to apply to the new instance (cannot be null). A new, initialized instance.

Applies a protocol-buffer stream to an existing instance.

The existing instance to be modified (cannot be null). The binary stream to apply to the instance (cannot be null). The updated instance

Writes a protocol-buffer representation of the given instance to the supplied stream, with a length-prefix. This is useful for socket programming, as DeserializeWithLengthPrefix/MergeWithLengthPrefix can be used to read the single object back from an ongoing stream.

The existing instance to be serialized (cannot be null). How to encode the length prefix. The destination stream to write to. The tag used as a prefix to each record (only used with base-128 style prefixes).

Applies a protocol-buffer stream to an existing instance (or null), using length-prefixed data - useful with network IO.

The existing instance to be modified (can be null). The binary stream to apply to the instance (cannot be null). How to encode the length prefix. Used to resolve types on a per-field basis. The updated instance; this may be different to the instance argument if either the original instance was null, or the stream defines a known sub-type of the original instance.

Indicates whether the supplied type is explicitly modelled by the model

Global switches that change the behavior of protobuf-net

Maps a field-number to a type

Perform the steps necessary to serialize this data.

The value to be serialized. The writer entity that is accumulating the output data.

Perform the steps necessary to deserialize this data.

The current value, if appropriate. The reader providing the input data. The updated / replacement value.

Emit the IL necessary to perform the given actions to serialize this data.

Details and utilities for the method being generated. The source of the data to work against; If the value is only needed once, then LoadValue is sufficient. If the value is needed multiple times, then note that a "null" means "the top of the stack", in which case you should create your own copy - GetLocalWithValue.

Emit the IL necessary to perform the given actions to deserialize this data.

Details and utilities for the method being generated. For nested values, the instance holding the values; note that this is not always provided - a null means not supplied. Since this is always a variable or argument, it is not necessary to consume this value.

The type that this serializer is intended to work for.

Indicates whether a Read operation replaces the existing value, or extends the value. If false, the "value" parameter to Read is discarded, and should be passed in as null.

Now all Read operations return a value (although most do); if false no value should be expected.

Uses protocol buffer serialization on the specified operation; note that this must be enabled on both the client and server.

Configuration element to swap out DatatContractSerilaizer with the XmlProtoSerializer for a given endpoint.

Creates a new ProtoBehaviorExtension instance.

Creates a behavior extension based on the current configuration settings.

The behavior extension.

Gets the type of behavior.

Behavior to swap out DatatContractSerilaizer with the XmlProtoSerializer for a given endpoint. Add the following to the server and client app.config in the system.serviceModel section: Configure your endpoints to have a behaviorConfiguration as follows:

Describes a WCF operation behaviour that can perform protobuf serialization

Create a new ProtoOperationBehavior instance

Creates a protobuf serializer if possible (falling back to the default WCF serializer)

The type-model that should be used with this behaviour

An xml object serializer that can embed protobuf data in a base-64 hunk (looking like a byte[])

Attempt to create a new serializer for the given model and type

A new serializer instance if the type is recognised by the model; null otherwise

Creates a new serializer for the given model and type

Ends an object in the output

Begins an object in the output

Writes the body of an object in the output

Indicates whether this is the start of an object we are prepared to handle

Reads the body of an object

Used to hold particulars relating to nested objects. This is opaque to the caller - simply give back the token you are given at the end of an object.

Indicates the encoding used to represent an individual value in a protobuf stream

Represents an error condition

Base-128 variant-length encoding

Fixed-length 8-byte encoding

Length-variant-prefixed encoding

Indicates the start of a group

Indicates the end of a group

Fixed-length 4-byte encoding

This is not a formal wire-type in the "protocol buffers" spec, but denotes a variant integer that should be interpreted using zig-zag semantics (so -ve numbers aren't a significant overhead)