MartijnVisser commented on code in PR #28995:
URL: https://github.com/apache/flink/pull/28995#discussion_r4185785116
##########
flink-dist/pom.xml:
##########
@@ -593,6 +593,26 @@ under the License.
<skip>true</skip>
</configuration>
</execution>
+ <execution>
+ <!-- Keep OkHttp/okio off the
core Flink distribution classpath. Metrics reporters that
+ bundle OkHttp declare
it as an optional (non-transitive) dependency and ship it shaded
+ inside their plugin
jar, so it never reaches this distribution tree. -->
+
<id>forbid-okhttp-on-core-classpath</id>
+ <goals>
+ <goal>enforce</goal>
+ </goals>
+ <configuration>
+ <rules>
+
<bannedDependencies>
+
<excludes>
Review Comment:
Yes, with `-Didea.version` the datadog okhttp becomes a provided transitive
and the rule fails. Provided scope is now allowed, like test scope in
`forbid-eol-okhttp`. Compile scope still fails.
##########
flink-kubernetes/pom.xml:
##########
@@ -75,10 +80,12 @@ under the License.
</exclusions>
</dependency>
- <!-- Since 7.0.0, Fabric8 uses Vert.x as its HTTP client, but
we want to use OkHttp to keep the same dependencies as used before this
version. -->
+ <!-- Since 7.0.0, Fabric8 uses Vert.x as its default HTTP
client. We use the built-in java.net.http (JDK)
+ client instead (see fabric8.httpclient.impl above),
which keeps OkHttp/okio/Kotlin off the classpath and
+ avoids pulling in Vert.x/Netty. -->
<dependency>
<groupId>io.fabric8</groupId>
- <artifactId>kubernetes-httpclient-okhttp</artifactId>
+
<artifactId>kubernetes-httpclient-${fabric8.httpclient.impl}</artifactId>
Review Comment:
Yes, reproduced for both `proxy-url` and `HTTPS_PROXY`: master uses the
proxy, this branch fails on client creation. The JDK client has no SOCKS
support (JDK-8214516), so I've documented it and spelled it out in the release
note.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
