The Bazel action currently isn't setup to regenerate the proto gencode itself, meaning that it won't actually be running on changes in the pending PR. Long-term, we should migrate this back to Bazel after enabling proto codegen properly.
PiperOrigin-RevId: 611238217
This ensures that any dependent descriptors are already initialized before we use them and we're not initializing them lazily as-we-go. This should help ensure that dependent descriptors are fully initialized before we need them.
PiperOrigin-RevId: 610858264
Avoid potential races with other threads reading features that do not share a lock while features are temporarily null.
Special handling for proto1 mutable should not actually be needed, since setProto doesn't update dependency protos.
PiperOrigin-RevId: 610783483
This change keeps the existing naming for properties generated from proto2 files, while improving the experience under editions. This introduces a C# incompatibility when upgrading a proto from proto2 to editions, but we anticipate this being a relatively rare problem.
PiperOrigin-RevId: 610783407
GPB_ENUM_FWD_DECLARE needs `int32_t`, so add the import so GPBBootstrap.h
stands on its own without needing users to import something else.
Add another compile test file to ensure this doesn't break.
PiperOrigin-RevId: 610403337
JUnit4 does not support nested tests so these weren't running. Fixes setup problems and test logic. Oneof fields now inherit from their oneof, and top-level extensions inherit from top-level file when parent descriptor is null.
PiperOrigin-RevId: 609840087
The non-string field setters can bypass the vtables already, the string setters still go through the vtable path here because its more important to let them take an `impl SettableValue` to be able to set either a &str or a &ProtoStr already.
PiperOrigin-RevId: 609705233
The actual construction of the zeroed block is now entirely safe.
This will be accessed in nearly every program using protobuf.
Using a static in .bss has much less overhead than an atomically-constructed
dynamic allocation and is far more predictable for space-constrained systems.
In the future, if dynamic allocation is kept, it should use std::sync::OnceLock
instead of the much less safe Once combined with `static mut`.
PiperOrigin-RevId: 609635555
It now uses the same prefix as other thunks needed for the proxied type,
so the RawMapThunk helper can be used for enums.
Calling it a "iter next" thunk is misleading.
It does not increment the iterator as "next" implies,
it only gets the current key/value the iterator points to.
PiperOrigin-RevId: 609527442
Protobuf Team Bot
Mike Kruskal
Hong Shin
Sandy Zhang
Chris Kennelly
tony
Marcin Kowalczyk
Protobuf Team Bot
Thomas Van Lenten
Alyssa Haroldsen
Ilya Tokar
YoloMao
Anton Grbin