SEZ9 commented on PR #12293:
URL: https://github.com/apache/seatunnel/pull/12293#issuecomment-6392496104

   Thanks for the quick follow-up, @sudeephazra. Since my last pass at 
`6080bcde6123`, the only new commit is `80b1635222` (head `80b16352229d`), and 
it touches only `ADLSConfigValidator.java`, which matches what you described.
   
   On the `hadoop_adls_properties` denylist in `ADLSConfigValidator.java`: the 
seven new prefixes cover the class-loading and token-provider ABFS keys that 
`rejectsCredentialAndClassLoadingAdvancedProperties` expects, and the trimmed, 
lower-cased `startsWith` handles the per-account suffix and padded/upper-case 
variants, so that part looks right. One thing to settle before approval: the 
same hunk dropped the `fs.s3` prefix from the reject list. Nothing in the 
commit message or this thread indicates that was intentional. Could you either 
restore `fs.s3` (ideally with a test assertion) or explain why removing it is 
correct?
   
   Noted on `unit-test` passing on Ubuntu Java 11/17 and the Windows leg 
failing for an unrelated reason; I will not block on Windows.
   
   Remaining items from my earlier review, as a checklist:
   
   1. Restore (or justify removing) the `fs.s3` prefix in 
`ADLSConfigValidator.java`, with test coverage.
   2. The validator trims option values but `ADLSHadoopConf.java` consumes raw 
values, so whitespace from env-var substitution passes validation and then 
fails later; please apply the normalized values in `ADLSHadoopConf.java` or 
reject untrimmed values up front.
   3. The OAuth token endpoint in `ADLSRuntimeCompatibility.java` is built from 
unvalidated `authority_host`/`tenant_id`; please require a secure scheme and a 
sane host/tenant shape so the client secret cannot be POSTed over plain HTTP or 
to an arbitrary URL.
   4. `docs/en/connectors/sink/ADLSFile.md`: add the HNS (hierarchical 
namespace) requirement behind the transactional `tmp_path` -> `path` rename, 
document the validation rules users will hit (account/container naming, auth 
mutual exclusion, blocked `hadoop_adls_properties` keys), fix the shipped 
template placeholders so they pass those rules, and bring the en/zh option 
tables in line with what the factories expose.
   5. `docs/en/connectors/source/ADLSFile.md`: correct the `discovery_mode` and 
`start_mode` defaults to `ONCE` and `EARLIEST` to match the code and the 
Chinese doc.
   6. Mention the `ConfigShadeUtils.java` change adding `account_key` to the 
default log-mask list in the docs and PR description, since it affects all 
connectors.
   
   If you would prefer to split any of the docs items into a follow-up PR, just 
say so and I am happy to track them separately; items 1-3 I would like in this 
PR.
   
   <!-- streview-comment:2013 -->


-- 
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]

Reply via email to