For LITE messages we append the type name to the data, avoiding an
indirection and a relocation.
For descriptor messages we mark them as such and have a secondary vtable. This
way we avoid having to put those pointers on each message ClassData. We save in
data size and relocations. If in the future we determine that the extra jump is
costly we can inline the relevant descriptor method pointers into `ClassData`.
PiperOrigin-RevId: 565775188
Unfortunately in OSS with Bazel it appears that we must include this empty file, otherwise we get compile errors. More info in: b/300637173
PiperOrigin-RevId: 565699708
We do not want to repeatedly check `using_sso` when iterating over an array. Hence instead of using `Get(i) / Mutable(i)` from base class, we request a pointer to the underlying array (thus, reducing number of checks from N to 1).
Additionally, there is a loop with a `Delete` call - updated it to use `CommonHandler`, so that it reduces code bloat.
PiperOrigin-RevId: 565665669
The new rules are:
- `upb_minitable_proto_library()`: contains the MiniTables only
- `upb_c_proto_library()`: Contains the C API. Depends on the MiniTables
This involved splitting upb code generation into two separate aspects, one for MiniTables and one for the C API.
PiperOrigin-RevId: 565518070
Change PYUPB_DESCRIPTOR_MODULE to "google.protobuf.descriptor_pb2" because protobuf python is using google.protobuf.descriptor_pb2 everywhere
Also check import result of PYUPB_DESCRIPTOR_MODULE
PiperOrigin-RevId: 565409795
This restores a CI run that used to exist in the upb repo. I still need to add
test runs to exercise the wheels, but this change adds the functionality for
building and uploading them.
PiperOrigin-RevId: 564896637
Rust does not need (and should not need) these to be propagated up the tree. The Rust aspect just needs to use the thunks for a single proto_library() rule.
PiperOrigin-RevId: 564827273
A couple weeks ago we moved upb into the protobuf Git repo, and this change
continues the merger of the two repos by making them into a single Bazel repo.
This was mostly a matter of deleting upb's WORKSPACE file and fixing up a bunch
of references to reflect the new structure.
Most of the changes are pretty mechanical, but one thing that needed more
invasive changes was the Python script for generating CMakeLists.txt,
make_cmakelists.py. The WORKSPACE file it relied on no longer exists with this
change, so I updated it to hardcode the information it needed from that file.
PiperOrigin-RevId: 564810016
This should ensure that the upb workspace uses the same version of absl that
the protobuf workspace does. Soon this will be irrelevant since we will be
deleting the upb workspace, but this will help keep everything building
successfully until then.
PiperOrigin-RevId: 564405718
This is the first step of a migration to split apart upb_proto_library() into upb_c_proto_library() and upb_minitable_proto_library().
PiperOrigin-RevId: 563837216
I merged a handful of PRs on the upb repo after upb moved into the protobuf repo. This PR cherry-picks them here so that they will not be lost.
```
commit 7afb426a5a
Author: Keith Smiley <keithbsmiley@gmail.com>
Date: Thu Sep 7 11:36:01 2023 -0700
[bazel] Fix disallowing dylibs on darwin (#1180)
Since this bazel commit
ec5553352f
building dylibs like the ones in this rule on darwin platforms has been
unsupported. This feature is a default C++ toolchain feature to indicate
this. In bazel 7.x these dylibs will fail to link if they are still
built. As far as I can tell in the tests even if they are built they are
never used on macOS.
Co-authored-by: Adam Cozzette <acozzette@google.com>
commit 72decab5ec
Author: Keith Smiley <keithbsmiley@gmail.com>
Date: Thu Sep 7 09:42:20 2023 -0700
Add missing darwin_x86_64 CPU (#1181)
This CPU is often used when cross compiling from M1 machines. I'm also
hoping we can remove the legacy 'darwin' CPU.
commit ccadaf3196
Author: messense <messense@icloud.com>
Date: Fri Sep 8 00:28:54 2023 +0800
Fix `PyUpb_Message_MergeInternal` segfault (#1338)
when `PyUpb_Message_MergeFromString` returns `NULL`, currently
`PyUpb_Message_MergeInternal` will call `Py_DECREF` on `NULL`
which results in a segmentation fault.
This patch switches to `Py_XDECREF` to fix the segfault.
commit 2a5724d86e
Author: Kevin Greene <kgreenek@gmail.com>
Date: Wed Sep 6 16:46:35 2023 -0700
Fix lambda capture compiler error with c++20 (#1502)
When compiling with C++20, the following error is produced:
```
upb/mini_table.hpp:63:22: note: add explicit 'this' or '*this' capture
upb/mini_table.hpp: In lambda function:
upb/mini_table.hpp:71:22: error: implicit capture of 'this' via '[=]' is deprecated in C++20 [-Werror=deprecated]
71 | return appender_([=](char* buf) {
```
In C++20, it is no longer allowed to implicitly capture 'this' in a
lambda using [=].
This commit explicitly captures required values in the appropriate
lambdas and removes all uses of [=] with lambdas.
```
Closes#13908
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/13908 from acozzette:upb 7afb426a5a
PiperOrigin-RevId: 563784513
This partially fixes https://github.com/protocolbuffers/protobuf/issues/10088. The test case from that bug significantly improves with this change. However we still have a global map that does not shrink, which can still create the appearance of leaking memory, as it will not be freed until the module is unloaded.
PiperOrigin-RevId: 563124724
Protobuf Team Bot
Protobuf Team Bot
Joshua Haberman
Jie Luo
Hong Shin
Eric Salo
Adam Cozzette
Laramie Leavitt
Mike Kruskal
Richard Levasseur