-
Notifications
You must be signed in to change notification settings - Fork 4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
update to jdk 11 [AS-594] #1372
Changes from 8 commits
df335dd
fd4f4a8
237ebe5
4eb13a8
b5befd2
7c9df33
66f3133
e96c5ec
46cee51
d70439b
3f3d4ad
2f04caf
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1 +1 @@ | ||
1.8 | ||
11 |
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1 +1 @@ | ||
sbt.version = 1.4.6 | ||
sbt.version = 1.4.7 |
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -9,6 +9,7 @@ object Merging { | |
case PathList("com", "typesafe", _ @ _*) => MergeStrategy.last | ||
case PathList("com", "google", "auto", "value", _ @ _*) => MergeStrategy.last | ||
case PathList("io", "sundr", _ @ _*) => MergeStrategy.last | ||
case PathList("javax", "activation", _ @ _*) => MergeStrategy.first | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. This comes in through clients built with swagger codegen. I'm not positive that this is the most correct merge strategy, but I don't think it matters hugely. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Interesting. These generated client libraries are including the classes, not referring to them as a dependency? I'm surprised OpenAPI doesn't have support for transitive dependencies. Is there a link to an issue that can be included as a comment? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. There's a series of issues referenced here that lead to Jackson and then to swagger-core, tbh I don't understand this all well enough to know what would be a useful comment. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. There's a resolution to that last one, which will require updates to the version of swagger used to generate the clients, but I'm not exactly sure I understand what will need to happen. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I much prefer using exclude rules instead of the blunt-intstrument merge strategy. Do you know where the conflicts are, and could you add excludes for those instead? fwiw, we recently also had assembly issues in orch on javax.activation, though that was via javax.mail: broadinstitute/firecloud-orchestration#838 There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I didn't experiment with that, I think I'd need some help to try it. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. +1 for using exclusion rules |
||
// For the following error: | ||
//[error] java.lang.RuntimeException: deduplicate: different file contents found in the following: | ||
//[error] /root/.cache/coursier/v1/https/repo1.maven.org/maven2/com/google/protobuf/protobuf-java/3.11.4/protobuf-java-3.11.4.jar:google/protobuf/field_mask.proto | ||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,3 +1,4 @@ | ||
import MinnieKenny.testSettings | ||
import sbt.Keys._ | ||
import sbt._ | ||
|
||
|
@@ -22,6 +23,9 @@ object Testing { | |
|
||
val commonTestSettings: Seq[Setting[_]] = List( | ||
|
||
testOptions in Test += Tests.Setup(() => | ||
sys.props += "mockserver.logLevel" -> "WARN" | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. The default log verbosity for mockserver logs every mock defined, for every test that defines it. Since this goes to the console for jenkins jobs, it makes the logs huge, to the point that the browser has trouble searching through them. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Awesome!! |
||
), | ||
// SLF4J initializes itself upon the first logging call. Because sbt | ||
// runs tests in parallel it is likely that a second thread will | ||
// invoke a second logging call before SLF4J has completed | ||
|
@@ -55,7 +59,7 @@ object Testing { | |
(testOnly in Test) := ((testOnly in Test) dependsOn validMySqlHost).evaluated, | ||
|
||
parallelExecution in Test := false | ||
) ++ MinnieKenny.testSettings | ||
) ++ (if (sys.props.getOrElse("secrets.skip", false) != "true") MinnieKenny.testSettings else List()) | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I chose this option name somewhat arbitrarily, if it doesn't make sense it can be changed. Alternatively, we could remove Minnie-Kenny being run by sbt; Orch doesn't have it, and it's run manually in the github CI. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. To confirm: Orch doesn't check for the secrets leaks commited in There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Orch doesn't have any minnie-kenny anything of any kind; I'm not aware of any discussion around it, @davidangb do you know of anything? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I don't see any tickets in the backlog. It is used by most other services (like rawls and sam) so maybe we should add it There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. for my clarification: this discussion of orch is tangential to this PR, right? yeah, if orch does not have secrets-detection, it absolutely needs it. I'd prefer to check with appsec first on the progress of their centralized/cross-repo secrets detection before committing to any work, if appsec is already handling it let's just let them. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I think it's tangential. The question came from the switch to skip testing for git-secrets via the line above that says There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. @zarsky-broad please confirm, but my understanding is this toggle is specific to Rawls, and is called from the change in this PR in docker/install.sh. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Yeah. The old build image for Rawls didn't include git at all, so the secrets check was skipped silently. The new image has git, so the secrets check was failing here because it didn't have git-secrets. I added an optional flag to skip it that's only used in the The alternative is to not have sbt run the check at all, since it's explicitly run in the Github CI workflow. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Ah, I see the problem:
Orch has no minnie-kenny at all. Separately, the rawls github workflow runs minnie-kenny. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I think there's a scenario here where you're comparing a Boolean to a String. My quick test in the repl shows that this can cause problems. Recommend coercing everything the same type. |
||
|
||
implicit class ProjectTestSettings(val project: Project) extends AnyVal { | ||
def withTestSettings: Project = project | ||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1 +1 @@ | ||
sbt.version=1.4.6 | ||
sbt.version=1.4.7 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Assuming container-based compilation is absolutely required, someday I hope this custom image gets pushed to a public GCR or anywhere else besides docker.io. We have so many tests in the terra multi-repo that are failing due to pull rate limits from docker.io these days. 😢
Or for easier patching of individual components, just use a more flexible base container like focal and then download java and the 0.016GB sbt package. While Java gets patches much slower, Scala and SBT get patches once a month. For example: sbt 1.4.8 is already out.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So:
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
my take is that even though sbt will automatically download whatever it needs, the entire point of adding sbt to a baseimage is to preload as many dependencies as possible at docker-build time, and avoid re-downloading those dependencies on every build of every service that uses this image. So, counting on sbt to download what it needs defeats the purpose - if we're relying on that, I could see omitting it entirely and downloading it at runtime, avoiding the maintenance of the baseimage.