Since maps have a different in-memory representation than repeated messages, it is not safe to tree shake map entry messages.
Specifically, map fields use `kUpb_FieldMode_Map` instead of `kUpb_FieldMode_Array`, and there is no easy way to get the linker to rewrite this when the map entry is dropped.
However it is safe to tree shake message typed map values, since these use `kUpb_FieldMode_Scalar`, which remains valid even after the type is tree shaken.
PiperOrigin-RevId: 646199050
Python dict is now able to be assigned (by create and copy, not reference) and compared with the Protobuf Struct field.
Python list is now able to be assigned (by create and copy, not reference) and compared with the Protobuf ListValue field.
example usage:
dictionary = {'key1': 5.0, 'key2': {'subkey': 11.0, 'k': False},}
list_value = [6, 'seven', True, False, None, dictionary]
msg = more_messages_pb2.WKTMessage(
optional_struct=dictionary, optional_list_value=list_value
)
self.assertEqual(msg.optional_struct, dictionary)
self.assertEqual(msg.optional_list_value, list_value)
PiperOrigin-RevId: 646099987
We change InferLegacyProtoFeatures to set ctype based on string_type when string_type is set. We need to update CppGenerator::ValidateFeatures to allow both ctype and string_type to be set because it runs after InferLegacyProtoFeatures.
PiperOrigin-RevId: 645480157
When compiling the upb codebase, I randomly noticed there was one usage of directly calling `longjmp` though it seems the convention is to use the `UPB_LONGJMP` macro. AFAICT, the macro is mostly a wrapper to improve behavior on MacOS, so this may improve behavior there, otherwise it should be just making the code more consistent.
Closes#17201
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/17201 from anuraaga:patch-1 14cc027ef2
PiperOrigin-RevId: 645420330
This class is preferred over using `typeid` for a few reasons:
- It works with RTTI disabled.
- It works for `DynamicMessage` types.
- It works in custom vtable mode.
PiperOrigin-RevId: 645145751
`Vec<u8>` is a more idiomatic Rust type to return for serialization.
For the C++ kernel, we are able to return this type with no extra copying. We
still use `SerializedData` type for FFI, but convert the result into a
`Vec<u8>` using a new `into_vec` method.
The upb kernel serializes onto an arena, so for upb we do need to copy the data
to get it into a `Vec<u8>`.
PiperOrigin-RevId: 644444571
Hi,
We get a compiler warning in MSVC that a signed/unsigned comparison is executed.
The warning is at around line 1073 in src/google/protobuf/map.h
```cpp
ABSL_DCHECK_GE(new_num_buckets, kMinTableSize); // <-- signed/unsigned comparison
const auto old_table = table_;
const map_index_t old_table_size = num_buckets_;
num_buckets_ = new_num_buckets;
```
This is due to `kMinTableSize` is a enum value. Probably a C++98 way to enforce constexpr.
This PR changes `kMinTableSize` to be a constexpr unsigned `map_index_t` and removes the casts instead of fixing the warning with a cast at ne 1073.
It seems like `kMinTableSize` is correlated with the `TableSize` of this class, which is map_index_t. Therefore the min value can at maximum the max value of map_index_t. And the min value of `kMinTableSize` can also not be negative, as this makes no sense.
It is also nicer to have fewer casts.
Closes#17013
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/17013 from TinyTinni:removes-msvc-warning-map 4f954e8c9b
PiperOrigin-RevId: 644207177
I've made a new name for the interface, ExtensionSerializer, so we can keep ExtensionWriter in the same place for backwards-compatibility for a little while.
PiperOrigin-RevId: 644172922
These classes are deprecated and will be removed in the next breaking change. Users should update gencode to >= 4.26.x which uses GeneratedMessage instead of GeneratedMessageV3.
Tested with //compatibility:java_conformance_v3.25.0 which builds the runtime against 3.25.0 gencode.
PiperOrigin-RevId: 644136172
If we have a `oneof Foo` in a message `Bar`, we should access the discriminator as `Bar.Builder.fooCase` rather than `Bar.Builder.getFooCase()`.
PiperOrigin-RevId: 644107242
Fixes#16894
It is tested on command execution only, where it fixes the issues of reported memory leaks, heap corruption and segmentation fault.
I could not run tests on my Mac because of #16944Closes#16993
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/16993 from simonberger:bugfix/php-ext-persistent-global-corruption 62b5529b6a
PiperOrigin-RevId: 644034181
This extends previous workaround for java features in unknown fields, to include features extensions that are known but use a mismatched descriptor.
This can happen when users bring their own descriptors via buildFrom.
PiperOrigin-RevId: 644013578
This was introduced by the previous fix for delimited inheritance, and was never released. This fix removes all getType() calls from crosslink, where it's not safe to inspect the message type, which is still a placeholder, until after crosslinking. Using the inferred type is not necessary since we treat messages and groups the same during crosslink.
PiperOrigin-RevId: 643394981
Since statically tree shaken messages can never later become linked, we should not need to use any of the special code in the decoder. By using a distinct "empty" message type, we avoid triggering any of this special behavior. This avoids bugs around hazzers and other presence checks.
Also fixed a bug in the cmake staleness test that was causing test failures.
PiperOrigin-RevId: 643036818
New code should prefer `features.(pb.cpp).string_type`. We can't use the `deprecated=true` option for now due to failing tests.
PiperOrigin-RevId: 643036543
Derek Benson
Joshua Haberman
Protobuf Team Bot
Hong Shin
Jie Luo
Protobuf Team Bot
Anuraag (Rag) Agrawal
Sandy Zhang
Mark Hansen
Marcel Hlopko
Adam Cozzette
Jason Lunn
Matthias Möller
Congcong Cai
Éamonn McManus
Benoît Quenaudon
Simon Berger
Mike Kruskal