Add include_file_descriptor_set to prost-build
#326
Conversation
jen20
force-pushed
the
jen20/file-descriptor-set
branch
from
aecd357 to
ed927ba
Compare
|
I've force pushed here to rebase this on top of the current |
|
The new CI failures seem unrelated to this change, however. |
|
It seems like my change in 24d7877 would fix the build errors which presumably will hit any PR to this repo right now, but GHA is still using the version of the workflow from the main repo rather than the fork. |
danburkert
force-pushed
the
jen20/file-descriptor-set
branch
2 times, most recently
from
86a1761 to
c4aeb27
Compare
|
It looks like this is passing. Is this PR ready to be merged? It's on the must-have list for the next release, which I know folks are keen to start using (including myself). |
|
Thanks for the work here, this has been very useful as a branch -- is there anything I can do to help get this merged/released? |
Builds are failing on unrelated pull requests (e.g., tokio-rs#326) because the `curl` crate has pushed a new version that depends on Rust 1.40, while our CI tests against 1.39. We could probably bump to require 1.40 as well, especially since we want to use the same `#[non_exhaustive]` attribute, per tokio-rs#43. But a minimal change to unblock development and keep the current minimum stable Rust version is to just pin `curl < 0.4.34`. Test Plan: Running `cargo +1.39.0 test --workspace --all-targets` passes on my Linux machine. wchargin-branch: deps-curl-pre-0.4.34 wchargin-source: c2910530bd49190605176c9f69f3aff1863443c2
Builds are failing on unrelated pull requests (e.g., #326) because the `curl` crate has pushed a new version that depends on Rust 1.40, while our CI tests against 1.39. We could probably bump to require 1.40 as well, especially since we want to use the same `#[non_exhaustive]` attribute, per #43. But a minimal change to unblock development and keep the current minimum stable Rust version is to just pin `curl < 0.4.34`. Test Plan: Running `cargo +1.39.0 test --workspace --all-targets` passes on my Linux machine. wchargin-branch: deps-curl-pre-0.4.34 wchargin-source: c2910530bd49190605176c9f69f3aff1863443c2
jen20
force-pushed
the
jen20/file-descriptor-set
branch
from
9e06601 to
286c0cc
Compare
|
I've rebased this on top of the latest Update: actually the work to fix this seems minimal: #397 should do so, after which this can again be rebased. |
jen20
force-pushed
the
jen20/file-descriptor-set
branch
from
286c0cc to
bd8da93
Compare
|
This is rebased on top of #397, so the build should now succeed. |
jen20
force-pushed
the
jen20/file-descriptor-set
branch
from
bd8da93 to
b36dfcd
Compare
|
Rebased again on top of 278e8ab. |
This commit adds support for writing an encoded version of the appropriate FileDescriptorSet to a file named `file_descriptor_set.bin` in the output directory. This can be used with the `include_bytes` macro by consuming applications or libraries. The rationale for this pull request is to support the gRPC Server Reflection Protocol, which operates in terms of FileDescriptorProto, in Tonic. FileDescriptorProto instances are contained inside a FileDescriptorSet.
jen20
force-pushed
the
jen20/file-descriptor-set
branch
from
b36dfcd to
6e5e9ad
Compare
This pull request adds support for writing an encoded version of the appropriate FileDescriptorSet to a file named
file_descriptor_set.binin the output directory. This can be used with theinclude_bytesmacro by consuming applications or libraries.The rationale for this pull request is to support the gRPC Server Reflection Protocol, which operates in terms of
FileDescriptorProto, in Tonic.FileDescriptorProtoinstances are contained inside aFileDescriptorSet.This supersedes #311 following discussion on that issue.