You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
Now that we have shim layers we can support multiple versions of spark from a single jar (including unreleased SNAPSHOT versions) Generally it is bad practice to depend on a SNAPSHOT version in a release.
Describe the solution you'd like
Have a maven profile that lets us disable all of the shim layers that depend on a SNAPSHOT release. Ideally it would look at the version numbers to do this, but I don't think maven is capable of that, so we probably need to manually maintain the list of shim projects that should be a part of a release and which should not.
Would also be good to include this as a part of the pre-commit build so we know if someone broke the build by having a a release version of something depend on a non-release version. Sadly this is really only ever going to be the case for data bricks, so until we have databricks as a part of the pre-commit build we might not see a lot of success from this.
Describe alternatives you've considered
capture the spark SNAPSHOT jars this is built against so we could release it, but that might be problematic too.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Now that we have shim layers we can support multiple versions of spark from a single jar (including unreleased SNAPSHOT versions) Generally it is bad practice to depend on a SNAPSHOT version in a release.
Describe the solution you'd like
Have a maven profile that lets us disable all of the shim layers that depend on a SNAPSHOT release. Ideally it would look at the version numbers to do this, but I don't think maven is capable of that, so we probably need to manually maintain the list of shim projects that should be a part of a release and which should not.
Would also be good to include this as a part of the pre-commit build so we know if someone broke the build by having a a release version of something depend on a non-release version. Sadly this is really only ever going to be the case for data bricks, so until we have databricks as a part of the pre-commit build we might not see a lot of success from this.
Describe alternatives you've considered
capture the spark SNAPSHOT jars this is built against so we could release it, but that might be problematic too.
The text was updated successfully, but these errors were encountered: