Update data-plane-api#2430
Conversation
Signed-off-by: Kuat Yessenov <kuat@google.com>
|
@kyessenov not sure, have you experimented with adding a message to and instantiating the messages in |
Signed-off-by: Kuat Yessenov <kuat@google.com>
Signed-off-by: Kuat Yessenov <kuat@google.com>
Signed-off-by: Kuat Yessenov <kuat@google.com>
|
Resolved the issue by an explicit use of a proto dummy message ( |
Signed-off-by: Kuat Yessenov <kuat@google.com>
Signed-off-by: Kuat Yessenov <kuat@google.com>
bazel/repositories.bzl
Outdated
| name = "envoy_accesslog", | ||
| actual = "@envoy_api//envoy/service/accesslog/v2:als_cc", | ||
| ) | ||
| native.bind( |
There was a problem hiding this comment.
I'd be happy to clean this up separately to avoid binding so many names...
There was a problem hiding this comment.
I would prefer not using binds (I think they might be an antipattern as they make working with multiple workspaces difficult). We should use "actual" names, since the path is more or less stable.
There was a problem hiding this comment.
So, yeah, some history here. When we first started to Bazelify Envoy, it was unclear to me how best to handle external deps from a few perspectives:
- How would we be able to map to prebuilts at Lyft as well as use native Bazel repositories?
- How would we redirect dependencies when we imported into Google's monorepo.
Bind made sense at that time as an indirection mechanism to provide flexibility.
The prebuilt situation has stabilized and doesn't apply here (we don't prebuild @envoy_api). The import for Google is also now moot. While it's a little neater to rebind, TBH we're sed-ing a ton during the import, so rewriting @envoy_api is trivial.
So, definitely for this dependency, which is now sprawling with binds, let's just switch to direct reference to @envoy_api everywhere.
There was a problem hiding this comment.
+1 if we can just do this now since this is already a giant change and mostly find/replace. I wondered why we did this in the first place, thanks for the explanation.
There was a problem hiding this comment.
Done, with a sed swoosh
Signed-off-by: Kuat Yessenov <kuat@google.com>
mattklein123
left a comment
mattklein123
left a comment
There was a problem hiding this comment.
LGTM, thanks! Will defer to @htuch and @jmillikin-stripe on Bazel. Per your comment I would love to cleanup how we do the binds, seems extremely verbose, but can do that in a follow up.
Signed-off-by: Kuat Yessenov <kuat@google.com>
This reverts commit d6b630a.
This reverts commit d6b630a. Signed-off-by: Matt Klein <mklein@lyft.com>
Description: Adds support for registering a platform KV store implementation via EngineBuilder. Risk Level: Low Testing: CI Signed-off-by: Mike Schore <mike.schore@gmail.com> Signed-off-by: JP Simard <jp@jpsim.com>
Description: Adds support for registering a platform KV store implementation via EngineBuilder. Risk Level: Low Testing: CI Signed-off-by: Mike Schore <mike.schore@gmail.com> Signed-off-by: JP Simard <jp@jpsim.com>
3 participants
build: Apply data-plane-api protobuf refactoring to Envoy
Description:
Mechanical rewrite of bazel dependencies and imports to reflect the movement of sources.
Risk Level: Medium
Any indirect reference to protobuf type (
@type:) may be affected.Testing: No extra testing, but current tests are updated.
Docs Changes: envoyproxy/data-plane-api#421
Release Notes: Updated the protobuf references to reflect the structured packages in data-plane-api.
[Optional API Changes:] envoyproxy/data-plane-api#421