This past reliance doesn't work well with Swift, and the sources
were trying to over hand optimize to minimize rebuilds, and that
likely isn't worth the trouble, so explicit imports.
The remaining places that still use them are where they are
needed within the header to deal with relationships between
the local definitions or where there is a cycle between the
headers and it allows either one to be imported first and still
get a complete definition in the using context.
PiperOrigin-RevId: 671379912
Move some of the new logic out of `GPBUnknownField` so it will end up as a much
simpler "container", with all the serialization logic inside `GPBUnknownFields`
instead.
Also move some of the internal logic needed into static C functions to reduce the
ObjC class overhead of `GPBUnknownFields`.
This was all inspired by realizing during serialization related apis the `type` of
each field was being examined multiple times and reduces that in addition to reducing
the number of methods being invoked.
PiperOrigin-RevId: 650631975
There is a special case where message factories can be confused: if a module
written in C++ with pybind11 links against a self-recursive message, and that
message is part of another message loaded from Python, then the confusion
will happen.
Example:
# This one is also linked into the C++ module.
message SelfRecursive {
optional SelfRecursive self_recursive = 1;
}
# This one is used only in Python and not linked.
message OnlyUsedInPython {
optional SelfRecursive self_recursive = 2;
}
The caching through message_factory::RegisterMessageClass then happens on one
instance of the factory, but traversal with the lookup in another.
This occurs in the pure Python and upb implementations that have their own
default descriptor pools (and thus message factory).
Fix this by using the already passed message factory to registering the
message class to cache.
A test accounts for this case to avoid regressions.
PiperOrigin-RevId: 642551744
General
* License changed from Apache 2.0 to New BSD.
* It is now possible to define custom "options", which are basically
annotations which may be placed on definitions in a .proto file.
For example, you might define a field option called "foo" like so:
import "google/protobuf/descriptor.proto"
extend google.protobuf.FieldOptions {
optional string foo = 12345;
}
Then you annotate a field using the "foo" option:
message MyMessage {
optional int32 some_field = 1 [(foo) = "bar"]
}
The value of this option is then visible via the message's
Descriptor:
const FieldDescriptor* field =
MyMessage::descriptor()->FindFieldByName("some_field");
assert(field->options().GetExtension(foo) == "bar");
This feature has been implemented and tested in C++ and Java.
Other languages may or may not need to do extra work to support
custom options, depending on how they construct descriptors.
C++
* Fixed some GCC warnings that only occur when using -pedantic.
* Improved static initialization code, making ordering more
predictable among other things.
* TextFormat will no longer accept messages which contain multiple
instances of a singular field. Previously, the latter instance
would overwrite the former.
* Now works on systems that don't have hash_map.
Python
* Strings now use the "unicode" type rather than the "str" type.
String fields may still be assigned ASCII "str" values; they will
automatically be converted.
* Adding a property to an object representing a repeated field now
raises an exception. For example:
# No longer works (and never should have).
message.some_repeated_field.foo = 1
Thomas Van Lenten
Protobuf Team Bot
Joshua Haberman
listonjesse
Adam Cozzette
Feng Xiao
jieluo@google.com
xiaofeng@google.com
kenton@google.com
Jon Skeet
temporal