steveloughran commented on pull request #2971: URL: https://github.com/apache/hadoop/pull/2971#issuecomment-1068012871
@mukund-thakur, I have addressed all the little nits in the code -thank you for reviewing the text and Java docs so thoroughly. I have also improved that rejection of filesystem by schema, adding wasb to the set of unsupported stores, and a test for it. Once we add a better rename api to the filesystem/filecontext, we can add path capabilities for renames to probe for. I have also tried to clarify in the Java docs and comments places where there was ambiguity. This includes comments in the hadoop-azure pom where the references to the mapreduce jars are imported with scopes of `provided` and `test`. Everyone reviewing this patch needs to understand what these scopes mean, so they will understand why I have no intention of changing those declarations or adding anything to hadoop-common. | scope name | classpath of | transitive? | |------------|--------------|-------------| | `provided` | src/main and src/test builds | *no* | | `test` | src/test builds | *no* | It is not a requirement to use the file system, nor is it exported as a new dependency. I am 100% confident of this because these same dependencies were added to hadoop-aws in HADOOP-13786, _Add S3A committers for zero-rename commits to S3 endpoints_ -and nobody has ever reported the filesystem not instantiating. note, cloudstore storediag doesn't put these dependencies on the cp when invoked via `hadoop jar cloudstore.jar storediag` so verifying my statements is straightforward. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
