Move away from looking up selectors at proto initialization time to doing optimized string compares at implementation resolving time.
PiperOrigin-RevId: 623183331
Scrub all the objc related sources for references to "syntax"
and update them to no longer tie things accordingly.
The ObjC Protos `GPBFileDescriptor` concept still exposes the
`syntax`, but it was already marked as deprecated. This will be
removed in the future as nothing should need to look at the value.
Validate that all generation calls are on the helpers to get the
editions support from the C++ code.
PiperOrigin-RevId: 555971288
Move Field & Extension support for lack of UsesClassRef into the legacy api
so the only things taking the hit are old generate code.
PiperOrigin-RevId: 508074967
Add a new generator interface to:
- provide the message name directly instead of the overhead needed to try and compute it
- provide the data for the GPBFileDescriptor as a struct.
This allows -[GPBMessage fullName] to much more easily compute the fullname.
It also will allow the creation of all GPBFileDescriptors to be done on demand thus reducing the work in the startup codepaths as well as reducing memory requirements.
The generation changes are in a following commit.
PiperOrigin-RevId: 508063366
There already are explicit checks in debug builds, but this adds an addition
check within all builds to catch when unknown flags are being passed to the
runtime, which is an indication that the source generation is from a newer
version of the library.
PiperOrigin-RevId: 506091257
Looking back through history, figure out each place a generation change was
added and added compile asserts to help leave signals for what can be cleaned up
when if the legacy support is ever moved forward.
PiperOrigin-RevId: 506020387
Since it really just wraps some basic types, there are lots of instances
that are actually the same content wise, so to better match the contract
for NSObject it should compare the fields.
PiperOrigin-RevId: 505167021
The root class only needs to be started up if the message scopes extensions.
This updates the startup flows so the root is only started when the message
scopes extensions and thus the runtime requires that startup.
Bump the generated source version to account for the new initializer that
doesn't take the rootClass argument.
Fix typo while at it.
PiperOrigin-RevId: 504899046
The avoids passing four arguments and instead just passes one, so there tradeoff
in code is likely slightly better and from a perf pov, likely a wash.
PiperOrigin-RevId: 490540510
- FieldDescriptor:
- Add a new flag to mark when the enum on the field is closed (vs open).
- Support computing the state for when generated sources predate the support.
- EnumDescriptor:
- Support passing flags to the descriptor creation, currently closed is the
only new flag.
- Add an isClosed property to expose the state of the enum.
This does NOT update generation yet, allows things to be tested before the
generation support is added.
PiperOrigin-RevId: 488671606
- Streamline the FieldDescriptor loop a little.
- Move some of the init methods from PackagePrivate into the impl file as
nothing outside calls them and it better encapsulates the signature.
PiperOrigin-RevId: 488671041
The static analyzer in some versions of Xcode now flag the enumDescriptor_ as
retained by the generated methods and not being release in dealloc. So avoid the
whole issue by just bringing back the explicit getter.
PiperOrigin-RevId: 488436139
Going back to the first ObjC support commit, there never really was support for
not have the EnumDescriptors, so start removing that partial support.
PiperOrigin-RevId: 488381501
- Add a Descriptor flag to capture if the field should clear on being zeroed.
- Update the runtime to use the new clear on zero flag.
- Add a flag on message initialization to indicate the sources were generated
with/without this support so the runtime can backfill the flag for older
generated sources.
This should reduce binary size slightly, small performance improvement, and improve linkage by forcing references to all used classes.
Note that this maintains backwards compatibility for sources generated by older protoc for the time being. If you want the benefits
you will need to recompile your protos with the newer protoc.
Added new API to GPBEnumDescriptor to enable introspection of enum values.
Refactored implementation so that this contains a minimum of added code.
Clarified comments regarding behavior in the presence of the alias_allowed option.
Added unit tests for new functionality and for the alias case.
The Undefined Behavior sanitizer flags one part of the unittests for this.
For default values for `bytes` we write a length on the front of a c-string
in the static data, apparently the compiler/linker doesn't always make this
4 byte aligned, so it get flagged for undefined/degraded performance. Avoid
this by using memcpy instead.
- Capture the ObjC prefix used when generating the the file.
- Track the containing type on descriptors.
- Mark descriptors where the message class name got a suffix added to it.
- Expose a fullName property on Descriptors.
- Add helpers for packing/unpacking Any messages.
- Bump the ObjC runtime version number. Since we added methods and invoke them
in the generated code, ensure the code is running against a matching version.
Otherwise, someone could compile against headers, but run with a framework
that is older and get unknown selector failures. This should trip clearer
messaging.
Fixes https://github.com/google/protobuf/issues/1674
As bazel folks are looking at getting auto generation of module maps going and
the importing of sources files causes issues there. We were only do it to
hack around some of the apple linker behaviors around objc classes and
categories, but even that isn't complete and CocoaPods was already doing -ObjC,
and developers not using pods could have still needed it to ensure everything
was linked anyways; so drop the hack of importing sources.
Working on https://github.com/google/protobuf/issues/1599, specifically:
- Turn on more warnings that the Xcode UI calls out with individual controls.
- Manually add:
-Wundef
-Wswitch-enum
- Manually add and then diable in the unittests because of XCTest's headers:
-Wreserved-id-macro
-Wdocumentation-unknown-command
- Manually add -Wdirect-ivar-access, but disable it for the unittests and in
the library code (via #pragmas to suppress it). This is done so proto users
can enable the warning.
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).
- Add more to the ObjC dir readme.
- Merge the ExtensionField and ExtensionDescriptor to reduce overhead.
- Fix an initialization race.
- Clean up the Xcode schemes.
- Remove the class/enum filter.
- Remove some forced inline that were bloating things without proof of performance wins.
- Rename some internal types to avoid conflicts with the well know types protos.
- Drop the use of ApplyFunctions to the compiler/optimizer can do what it wants.
- Better document some possible future improvements.
- Add missing support for parsing repeated primitive fields in packed or unpacked forms.
- Improve -hash.
- Add *Count for repeated and map<> fields to avoid auto create when checking for them being set.
- Style fixups in the code.
- map<> serialization fixes and more tests.
- Autocreation of map<> fields (to match repeated fields).
- @@protoc_insertion_point(global_scope|imports).
- Fixup proto2 syntax extension support.
- Move all startup code to +initialize so it happen on class usage and not app startup.
- Have generated headers use forward declarations and move imports into generated code, reduces what is need at compile time to speed up compiled and avoid pointless rippling of rebuilds.
Dave MacLachlan
Joshua Haberman
Thomas Van Lenten
Protobuf Team Bot
Mike Kruskal
listonjesse
Joshua Haberman
Dave MacLachlan
Parveen Bhatia
Sergio Campamá
leovitch
Dimitris Koutsogiorgas