Replace rust-protobuf with prost#201
Conversation
See pingcap/kvproto#349 Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Conflicts: Cargo.toml generate-proto.sh src/lib.rs src/storage.rs src/util.rs
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
…ecause their constructors are private Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
|
The rest of the refactoring depends on tikv/protobuf-build#2 |
For at least the proto which is exported via kvproto, the |
|
That sounds reasonable. I'll add the protobuf stuff back. |
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
nrc
left a comment
There was a problem hiding this comment.
Looks great, just two very minor things remaining
|
Test failing because I have no idea about this. |
Signed-off-by: ice1000 <ice1000kotlin@foxmail.com>
|
Travis is failing on Windows only, @Hoverbear can we ignore that? |
Hoverbear
left a comment
There was a problem hiding this comment.
Seems the vast majority of these changes are fairly mechanical. After chatting a bit and investigating the situation around new_() (which I do not like!) and how we'll fix the situation, I feel this is a good step of progress to merge.
I would prefer if we rapidly move to a state where we have functions without this naming convention. Preferably before 0.6.0
|
@nrc We cannot! But this failure is related to travis, not you. We can force a few rebuilds and it should be green. I'll shepherd it. |
This PR:
build.rsto (re)generate prost structs and their corresponding protobuf wrappers (thanks to @nrc!)I suppose test-passing means ready-for-review, please leave your comments!