Protocol Buffers - Google's data interchange format (grpc依赖) https://developers.google.com/protocol-buffers/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

728 lines
26 KiB

// Generated by the protocol buffer compiler. DO NOT EDIT!
// source: google/protobuf/type.proto
// This CPP symbol can be defined to use imports that match up to the framework
// imports needed when using CocoaPods.
#if !defined(GPB_USE_PROTOBUF_FRAMEWORK_IMPORTS)
#define GPB_USE_PROTOBUF_FRAMEWORK_IMPORTS 0
#endif
#if GPB_USE_PROTOBUF_FRAMEWORK_IMPORTS
#import <Protobuf/GPBProtocolBuffers_RuntimeSupport.h>
#else
#import "GPBProtocolBuffers_RuntimeSupport.h"
#endif
#import <stdatomic.h>
#if GPB_USE_PROTOBUF_FRAMEWORK_IMPORTS
#import <Protobuf/GPBType.pbobjc.h>
#import <Protobuf/GPBAny.pbobjc.h>
#import <Protobuf/GPBSourceContext.pbobjc.h>
#else
#import "GPBType.pbobjc.h"
#import "GPBAny.pbobjc.h"
#import "GPBSourceContext.pbobjc.h"
#endif
// @@protoc_insertion_point(imports)
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Wdeprecated-declarations"
#pragma clang diagnostic ignored "-Wdollar-in-identifier-extension"
#pragma mark - Objective C Class declarations
// Forward declarations of Objective C classes that we can use as
// static values in struct initializers.
// We don't use [Foo class] because it is not a static value.
GPBObjCClassDeclaration(GPBAny);
GPBObjCClassDeclaration(GPBEnumValue);
GPBObjCClassDeclaration(GPBField);
GPBObjCClassDeclaration(GPBOption);
GPBObjCClassDeclaration(GPBSourceContext);
#pragma mark - GPBTypeRoot
@implementation GPBTypeRoot
// No extensions in the file and none of the imports (direct or indirect)
// defined extensions, so no need to generate +extensionRegistry.
@end
#pragma mark - GPBTypeRoot_FileDescriptor
static GPBFileDescriptor *GPBTypeRoot_FileDescriptor(void) {
// This is called by +initialize so there is no need to worry
// about thread safety of the singleton.
static GPBFileDescriptor *descriptor = NULL;
if (!descriptor) {
GPB_DEBUG_CHECK_RUNTIME_VERSIONS();
descriptor = [[GPBFileDescriptor alloc] initWithPackage:@"google.protobuf"
objcPrefix:@"GPB"
syntax:GPBFileSyntaxProto3];
}
return descriptor;
}
#pragma mark - Enum GPBSyntax
GPBEnumDescriptor *GPBSyntax_EnumDescriptor(void) {
static _Atomic(GPBEnumDescriptor*) descriptor = nil;
if (!descriptor) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
static const char *valueNames =
"SyntaxProto2\000SyntaxProto3\000";
static const int32_t values[] = {
GPBSyntax_SyntaxProto2,
GPBSyntax_SyntaxProto3,
};
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
GPBEnumDescriptor *worker =
[GPBEnumDescriptor allocDescriptorForName:GPBNSStringifySymbol(GPBSyntax)
valueNames:valueNames
values:values
count:(uint32_t)(sizeof(values) / sizeof(int32_t))
enumVerifier:GPBSyntax_IsValidValue];
GPBEnumDescriptor *expected = nil;
if (!atomic_compare_exchange_strong(&descriptor, &expected, worker)) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
[worker release];
}
}
return descriptor;
}
BOOL GPBSyntax_IsValidValue(int32_t value__) {
switch (value__) {
case GPBSyntax_SyntaxProto2:
case GPBSyntax_SyntaxProto3:
return YES;
default:
return NO;
}
}
#pragma mark - GPBType
@implementation GPBType
@dynamic name;
@dynamic fieldsArray, fieldsArray_Count;
@dynamic oneofsArray, oneofsArray_Count;
@dynamic optionsArray, optionsArray_Count;
@dynamic hasSourceContext, sourceContext;
@dynamic syntax;
typedef struct GPBType__storage_ {
uint32_t _has_storage_[1];
GPBSyntax syntax;
NSString *name;
NSMutableArray *fieldsArray;
NSMutableArray *oneofsArray;
NSMutableArray *optionsArray;
GPBSourceContext *sourceContext;
} GPBType__storage_;
// This method is threadsafe because it is initially called
// in +initialize for each subclass.
+ (GPBDescriptor *)descriptor {
static GPBDescriptor *descriptor = nil;
if (!descriptor) {
static GPBMessageFieldDescription fields[] = {
{
.name = "name",
.dataTypeSpecific.clazz = Nil,
.number = GPBType_FieldNumber_Name,
.hasIndex = 0,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBType__storage_, name),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "fieldsArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBField),
.number = GPBType_FieldNumber_FieldsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBType__storage_, fieldsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
{
.name = "oneofsArray",
.dataTypeSpecific.clazz = Nil,
.number = GPBType_FieldNumber_OneofsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBType__storage_, oneofsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeString,
},
{
.name = "optionsArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBOption),
.number = GPBType_FieldNumber_OptionsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBType__storage_, optionsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
{
.name = "sourceContext",
.dataTypeSpecific.clazz = GPBObjCClass(GPBSourceContext),
.number = GPBType_FieldNumber_SourceContext,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.hasIndex = 1,
.offset = (uint32_t)offsetof(GPBType__storage_, sourceContext),
.flags = GPBFieldOptional,
.dataType = GPBDataTypeMessage,
},
{
.name = "syntax",
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.dataTypeSpecific.enumDescFunc = GPBSyntax_EnumDescriptor,
.number = GPBType_FieldNumber_Syntax,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.hasIndex = 2,
.offset = (uint32_t)offsetof(GPBType__storage_, syntax),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldHasEnumDescriptor | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeEnum,
},
};
GPBDescriptor *localDescriptor =
[GPBDescriptor allocDescriptorForClass:[GPBType class]
rootClass:[GPBTypeRoot class]
file:GPBTypeRoot_FileDescriptor()
fields:fields
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
fieldCount:(uint32_t)(sizeof(fields) / sizeof(GPBMessageFieldDescription))
storageSize:sizeof(GPBType__storage_)
flags:(GPBDescriptorInitializationFlags)(GPBDescriptorInitializationFlag_UsesClassRefs | GPBDescriptorInitializationFlag_Proto3OptionalKnown)];
#if defined(DEBUG) && DEBUG
NSAssert(descriptor == nil, @"Startup recursed!");
#endif // DEBUG
descriptor = localDescriptor;
}
return descriptor;
}
@end
int32_t GPBType_Syntax_RawValue(GPBType *message) {
GPBDescriptor *descriptor = [GPBType descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBType_FieldNumber_Syntax];
return GPBGetMessageRawEnumField(message, field);
}
void SetGPBType_Syntax_RawValue(GPBType *message, int32_t value) {
GPBDescriptor *descriptor = [GPBType descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBType_FieldNumber_Syntax];
GPBSetMessageRawEnumField(message, field, value);
}
#pragma mark - GPBField
@implementation GPBField
@dynamic kind;
@dynamic cardinality;
@dynamic number;
@dynamic name;
@dynamic typeURL;
@dynamic oneofIndex;
@dynamic packed;
@dynamic optionsArray, optionsArray_Count;
@dynamic jsonName;
@dynamic defaultValue;
typedef struct GPBField__storage_ {
uint32_t _has_storage_[1];
GPBField_Kind kind;
GPBField_Cardinality cardinality;
int32_t number;
int32_t oneofIndex;
NSString *name;
NSString *typeURL;
NSMutableArray *optionsArray;
NSString *jsonName;
NSString *defaultValue;
} GPBField__storage_;
// This method is threadsafe because it is initially called
// in +initialize for each subclass.
+ (GPBDescriptor *)descriptor {
static GPBDescriptor *descriptor = nil;
if (!descriptor) {
static GPBMessageFieldDescription fields[] = {
{
.name = "kind",
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.dataTypeSpecific.enumDescFunc = GPBField_Kind_EnumDescriptor,
.number = GPBField_FieldNumber_Kind,
.hasIndex = 0,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, kind),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldHasEnumDescriptor | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeEnum,
},
{
.name = "cardinality",
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.dataTypeSpecific.enumDescFunc = GPBField_Cardinality_EnumDescriptor,
.number = GPBField_FieldNumber_Cardinality,
.hasIndex = 1,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, cardinality),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldHasEnumDescriptor | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeEnum,
},
{
.name = "number",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_Number,
.hasIndex = 2,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, number),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeInt32,
},
{
.name = "name",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_Name,
.hasIndex = 3,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, name),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "typeURL",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_TypeURL,
.hasIndex = 4,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, typeURL),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldTextFormatNameCustom | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "oneofIndex",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_OneofIndex,
.hasIndex = 5,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, oneofIndex),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeInt32,
},
{
.name = "packed",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_Packed,
.hasIndex = 6,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = 7, // Stored in _has_storage_ to save space.
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeBool,
},
{
.name = "optionsArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBOption),
.number = GPBField_FieldNumber_OptionsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, optionsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
{
.name = "jsonName",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_JsonName,
.hasIndex = 8,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, jsonName),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "defaultValue",
.dataTypeSpecific.clazz = Nil,
.number = GPBField_FieldNumber_DefaultValue,
.hasIndex = 9,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBField__storage_, defaultValue),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
};
GPBDescriptor *localDescriptor =
[GPBDescriptor allocDescriptorForClass:[GPBField class]
rootClass:[GPBTypeRoot class]
file:GPBTypeRoot_FileDescriptor()
fields:fields
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
fieldCount:(uint32_t)(sizeof(fields) / sizeof(GPBMessageFieldDescription))
storageSize:sizeof(GPBField__storage_)
flags:(GPBDescriptorInitializationFlags)(GPBDescriptorInitializationFlag_UsesClassRefs | GPBDescriptorInitializationFlag_Proto3OptionalKnown)];
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
#if !GPBOBJC_SKIP_MESSAGE_TEXTFORMAT_EXTRAS
static const char *extraTextFormatInfo =
"\001\006\004\241!!\000";
[localDescriptor setupExtraTextInfo:extraTextFormatInfo];
#endif // !GPBOBJC_SKIP_MESSAGE_TEXTFORMAT_EXTRAS
#if defined(DEBUG) && DEBUG
NSAssert(descriptor == nil, @"Startup recursed!");
#endif // DEBUG
descriptor = localDescriptor;
}
return descriptor;
}
@end
int32_t GPBField_Kind_RawValue(GPBField *message) {
GPBDescriptor *descriptor = [GPBField descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBField_FieldNumber_Kind];
return GPBGetMessageRawEnumField(message, field);
}
void SetGPBField_Kind_RawValue(GPBField *message, int32_t value) {
GPBDescriptor *descriptor = [GPBField descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBField_FieldNumber_Kind];
GPBSetMessageRawEnumField(message, field, value);
}
int32_t GPBField_Cardinality_RawValue(GPBField *message) {
GPBDescriptor *descriptor = [GPBField descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBField_FieldNumber_Cardinality];
return GPBGetMessageRawEnumField(message, field);
}
void SetGPBField_Cardinality_RawValue(GPBField *message, int32_t value) {
GPBDescriptor *descriptor = [GPBField descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBField_FieldNumber_Cardinality];
GPBSetMessageRawEnumField(message, field, value);
}
#pragma mark - Enum GPBField_Kind
GPBEnumDescriptor *GPBField_Kind_EnumDescriptor(void) {
static _Atomic(GPBEnumDescriptor*) descriptor = nil;
if (!descriptor) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
static const char *valueNames =
"TypeUnknown\000TypeDouble\000TypeFloat\000TypeInt"
"64\000TypeUint64\000TypeInt32\000TypeFixed64\000Type"
"Fixed32\000TypeBool\000TypeString\000TypeGroup\000Ty"
"peMessage\000TypeBytes\000TypeUint32\000TypeEnum\000"
"TypeSfixed32\000TypeSfixed64\000TypeSint32\000Typ"
"eSint64\000";
static const int32_t values[] = {
GPBField_Kind_TypeUnknown,
GPBField_Kind_TypeDouble,
GPBField_Kind_TypeFloat,
GPBField_Kind_TypeInt64,
GPBField_Kind_TypeUint64,
GPBField_Kind_TypeInt32,
GPBField_Kind_TypeFixed64,
GPBField_Kind_TypeFixed32,
GPBField_Kind_TypeBool,
GPBField_Kind_TypeString,
GPBField_Kind_TypeGroup,
GPBField_Kind_TypeMessage,
GPBField_Kind_TypeBytes,
GPBField_Kind_TypeUint32,
GPBField_Kind_TypeEnum,
GPBField_Kind_TypeSfixed32,
GPBField_Kind_TypeSfixed64,
GPBField_Kind_TypeSint32,
GPBField_Kind_TypeSint64,
};
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
GPBEnumDescriptor *worker =
[GPBEnumDescriptor allocDescriptorForName:GPBNSStringifySymbol(GPBField_Kind)
valueNames:valueNames
values:values
count:(uint32_t)(sizeof(values) / sizeof(int32_t))
enumVerifier:GPBField_Kind_IsValidValue];
GPBEnumDescriptor *expected = nil;
if (!atomic_compare_exchange_strong(&descriptor, &expected, worker)) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
[worker release];
}
}
return descriptor;
}
BOOL GPBField_Kind_IsValidValue(int32_t value__) {
switch (value__) {
case GPBField_Kind_TypeUnknown:
case GPBField_Kind_TypeDouble:
case GPBField_Kind_TypeFloat:
case GPBField_Kind_TypeInt64:
case GPBField_Kind_TypeUint64:
case GPBField_Kind_TypeInt32:
case GPBField_Kind_TypeFixed64:
case GPBField_Kind_TypeFixed32:
case GPBField_Kind_TypeBool:
case GPBField_Kind_TypeString:
case GPBField_Kind_TypeGroup:
case GPBField_Kind_TypeMessage:
case GPBField_Kind_TypeBytes:
case GPBField_Kind_TypeUint32:
case GPBField_Kind_TypeEnum:
case GPBField_Kind_TypeSfixed32:
case GPBField_Kind_TypeSfixed64:
case GPBField_Kind_TypeSint32:
case GPBField_Kind_TypeSint64:
return YES;
default:
return NO;
}
}
#pragma mark - Enum GPBField_Cardinality
GPBEnumDescriptor *GPBField_Cardinality_EnumDescriptor(void) {
static _Atomic(GPBEnumDescriptor*) descriptor = nil;
if (!descriptor) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
static const char *valueNames =
"CardinalityUnknown\000CardinalityOptional\000C"
"ardinalityRequired\000CardinalityRepeated\000";
static const int32_t values[] = {
GPBField_Cardinality_CardinalityUnknown,
GPBField_Cardinality_CardinalityOptional,
GPBField_Cardinality_CardinalityRequired,
GPBField_Cardinality_CardinalityRepeated,
};
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
GPBEnumDescriptor *worker =
[GPBEnumDescriptor allocDescriptorForName:GPBNSStringifySymbol(GPBField_Cardinality)
valueNames:valueNames
values:values
count:(uint32_t)(sizeof(values) / sizeof(int32_t))
enumVerifier:GPBField_Cardinality_IsValidValue];
GPBEnumDescriptor *expected = nil;
if (!atomic_compare_exchange_strong(&descriptor, &expected, worker)) {
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
[worker release];
}
}
return descriptor;
}
BOOL GPBField_Cardinality_IsValidValue(int32_t value__) {
switch (value__) {
case GPBField_Cardinality_CardinalityUnknown:
case GPBField_Cardinality_CardinalityOptional:
case GPBField_Cardinality_CardinalityRequired:
case GPBField_Cardinality_CardinalityRepeated:
return YES;
default:
return NO;
}
}
#pragma mark - GPBEnum
@implementation GPBEnum
@dynamic name;
@dynamic enumvalueArray, enumvalueArray_Count;
@dynamic optionsArray, optionsArray_Count;
@dynamic hasSourceContext, sourceContext;
@dynamic syntax;
typedef struct GPBEnum__storage_ {
uint32_t _has_storage_[1];
GPBSyntax syntax;
NSString *name;
NSMutableArray *enumvalueArray;
NSMutableArray *optionsArray;
GPBSourceContext *sourceContext;
} GPBEnum__storage_;
// This method is threadsafe because it is initially called
// in +initialize for each subclass.
+ (GPBDescriptor *)descriptor {
static GPBDescriptor *descriptor = nil;
if (!descriptor) {
static GPBMessageFieldDescription fields[] = {
{
.name = "name",
.dataTypeSpecific.clazz = Nil,
.number = GPBEnum_FieldNumber_Name,
.hasIndex = 0,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnum__storage_, name),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "enumvalueArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBEnumValue),
.number = GPBEnum_FieldNumber_EnumvalueArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnum__storage_, enumvalueArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
{
.name = "optionsArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBOption),
.number = GPBEnum_FieldNumber_OptionsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnum__storage_, optionsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
{
.name = "sourceContext",
.dataTypeSpecific.clazz = GPBObjCClass(GPBSourceContext),
.number = GPBEnum_FieldNumber_SourceContext,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.hasIndex = 1,
.offset = (uint32_t)offsetof(GPBEnum__storage_, sourceContext),
.flags = GPBFieldOptional,
.dataType = GPBDataTypeMessage,
},
{
.name = "syntax",
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.dataTypeSpecific.enumDescFunc = GPBSyntax_EnumDescriptor,
.number = GPBEnum_FieldNumber_Syntax,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.hasIndex = 2,
.offset = (uint32_t)offsetof(GPBEnum__storage_, syntax),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldHasEnumDescriptor | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeEnum,
},
};
GPBDescriptor *localDescriptor =
[GPBDescriptor allocDescriptorForClass:[GPBEnum class]
rootClass:[GPBTypeRoot class]
file:GPBTypeRoot_FileDescriptor()
fields:fields
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
fieldCount:(uint32_t)(sizeof(fields) / sizeof(GPBMessageFieldDescription))
storageSize:sizeof(GPBEnum__storage_)
flags:(GPBDescriptorInitializationFlags)(GPBDescriptorInitializationFlag_UsesClassRefs | GPBDescriptorInitializationFlag_Proto3OptionalKnown)];
#if defined(DEBUG) && DEBUG
NSAssert(descriptor == nil, @"Startup recursed!");
#endif // DEBUG
descriptor = localDescriptor;
}
return descriptor;
}
@end
int32_t GPBEnum_Syntax_RawValue(GPBEnum *message) {
GPBDescriptor *descriptor = [GPBEnum descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBEnum_FieldNumber_Syntax];
return GPBGetMessageRawEnumField(message, field);
}
void SetGPBEnum_Syntax_RawValue(GPBEnum *message, int32_t value) {
GPBDescriptor *descriptor = [GPBEnum descriptor];
GPBFieldDescriptor *field = [descriptor fieldWithNumber:GPBEnum_FieldNumber_Syntax];
GPBSetMessageRawEnumField(message, field, value);
}
#pragma mark - GPBEnumValue
@implementation GPBEnumValue
@dynamic name;
@dynamic number;
@dynamic optionsArray, optionsArray_Count;
typedef struct GPBEnumValue__storage_ {
uint32_t _has_storage_[1];
int32_t number;
NSString *name;
NSMutableArray *optionsArray;
} GPBEnumValue__storage_;
// This method is threadsafe because it is initially called
// in +initialize for each subclass.
+ (GPBDescriptor *)descriptor {
static GPBDescriptor *descriptor = nil;
if (!descriptor) {
static GPBMessageFieldDescription fields[] = {
{
.name = "name",
.dataTypeSpecific.clazz = Nil,
.number = GPBEnumValue_FieldNumber_Name,
.hasIndex = 0,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnumValue__storage_, name),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "number",
.dataTypeSpecific.clazz = Nil,
.number = GPBEnumValue_FieldNumber_Number,
.hasIndex = 1,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnumValue__storage_, number),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeInt32,
},
{
.name = "optionsArray",
.dataTypeSpecific.clazz = GPBObjCClass(GPBOption),
.number = GPBEnumValue_FieldNumber_OptionsArray,
.hasIndex = GPBNoHasBit,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBEnumValue__storage_, optionsArray),
.flags = GPBFieldRepeated,
.dataType = GPBDataTypeMessage,
},
};
GPBDescriptor *localDescriptor =
[GPBDescriptor allocDescriptorForClass:[GPBEnumValue class]
rootClass:[GPBTypeRoot class]
file:GPBTypeRoot_FileDescriptor()
fields:fields
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
fieldCount:(uint32_t)(sizeof(fields) / sizeof(GPBMessageFieldDescription))
storageSize:sizeof(GPBEnumValue__storage_)
flags:(GPBDescriptorInitializationFlags)(GPBDescriptorInitializationFlag_UsesClassRefs | GPBDescriptorInitializationFlag_Proto3OptionalKnown)];
#if defined(DEBUG) && DEBUG
NSAssert(descriptor == nil, @"Startup recursed!");
#endif // DEBUG
descriptor = localDescriptor;
}
return descriptor;
}
@end
#pragma mark - GPBOption
@implementation GPBOption
@dynamic name;
@dynamic hasValue, value;
typedef struct GPBOption__storage_ {
uint32_t _has_storage_[1];
NSString *name;
GPBAny *value;
} GPBOption__storage_;
// This method is threadsafe because it is initially called
// in +initialize for each subclass.
+ (GPBDescriptor *)descriptor {
static GPBDescriptor *descriptor = nil;
if (!descriptor) {
static GPBMessageFieldDescription fields[] = {
{
.name = "name",
.dataTypeSpecific.clazz = Nil,
.number = GPBOption_FieldNumber_Name,
.hasIndex = 0,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBOption__storage_, name),
.flags = (GPBFieldFlags)(GPBFieldOptional | GPBFieldClearHasIvarOnZero),
.dataType = GPBDataTypeString,
},
{
.name = "value",
.dataTypeSpecific.clazz = GPBObjCClass(GPBAny),
.number = GPBOption_FieldNumber_Value,
.hasIndex = 1,
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
.offset = (uint32_t)offsetof(GPBOption__storage_, value),
.flags = GPBFieldOptional,
.dataType = GPBDataTypeMessage,
},
};
GPBDescriptor *localDescriptor =
[GPBDescriptor allocDescriptorForClass:[GPBOption class]
rootClass:[GPBTypeRoot class]
file:GPBTypeRoot_FileDescriptor()
fields:fields
Shrink ObjC overhead (generated size and some runtime sizes) NOTE: This is a binary breaking change as structure sizes have changed size and/or order. - Drop capturing field options, no other options were captured and other mobile targeted languages don't try to capture this sort information (saved 8 bytes for every field defined (in static data and again in field descriptor instance size data). - No longer generate/compile in the messages/enums in descriptor.proto. If developers need it, they should generate it and compile it in. Reduced the overhead of the core library. - Compute the number of has_bits actually needs to avoid over reserving. - Let the boolean single fields store via a has_bit to avoid storage, makes the common cases of the instance size smaller. - Reorder some flags and down size the enums to contain the bits needed. - Reorder the items in the structures to manually ensure they are are packed better (especially when generating 64bit code - 8 bytes for every field, 16 bytes for every extension, instance sizes 8 bytes also). - Split off the structure initialization so when the default is zero, the generated static storage doesn't need to reserve the space. This is batched at the message level, so all the fields for the message have to have zero defaults to get the saves. By definition all proto3 syntax files fall into this case but it also saves space for the proto2 that use the standard defaults. (saves 8 bytes of static data for every field that had a zero default) - Don't track the enums defined by a message. Nothing in the runtime needs it and it was just generation and runtime overhead. (saves 8 bytes per enum) - Ensure EnumDescriptors are started up threadsafe in all cases. - Split some of the Descriptor initialization into multiple methods so the generated code isn't padded with lots of zero/nil args. - Change how oneof info is feed to the runtime enabling us to generate less static data (8 bytes saved per oneof for 64bit). - Change how enum value informat is capture to pack the data and only decode it if it ends up being needed. Avoids padding issues causing bloat of 64bit, and removes the needs for extra pointers in addition to the data (just the data and one pointer now).
9 years ago
fieldCount:(uint32_t)(sizeof(fields) / sizeof(GPBMessageFieldDescription))
storageSize:sizeof(GPBOption__storage_)
flags:(GPBDescriptorInitializationFlags)(GPBDescriptorInitializationFlag_UsesClassRefs | GPBDescriptorInitializationFlag_Proto3OptionalKnown)];
#if defined(DEBUG) && DEBUG
NSAssert(descriptor == nil, @"Startup recursed!");
#endif // DEBUG
descriptor = localDescriptor;
}
return descriptor;
}
@end
#pragma clang diagnostic pop
// @@protoc_insertion_point(global_scope)