SEZ9 commented on issue #13226:
URL: https://github.com/apache/seatunnel/issues/13226#issuecomment-6843532638

   Thanks for the detailed follow-ups, @zhangshenghang — the staged split and 
the compatibility table are what I was hoping for, and keeping `env` out of the 
credential path (with `ConfigShadeUtils` only handling 
`source`/`sink`/`transform`) is the right call.
   
   A few remaining asks:
   
   1. **Full `env` vs. `custom_parameters`** — passing the whole `env` block 
through `getEnvOptions()` works for me, with `custom_parameters` kept as an 
optional grouping namespace. Please make sure the EN/ZH docs describe the 
top-level-key form (e.g. `env.my_key`) as the primary usage and state clearly 
that env is never merged into plugin options.
   
   2. **`--check` as an open map** — no longer rejecting unknown top-level 
`env` keys is a behavior change for users who relied on typo detection. Please 
call this out explicitly in the docs, and confirm that declared env options 
(required and type checks) are still validated as before; a test covering one 
unknown key alongside one mistyped declared option would make that contract 
clear.
   
   3. **Stage 2 (`298c6970eb`)** — the local verification list looks thorough. 
Agreed on opening it only once the stage 1 factory API is settled; when you do, 
please include the legacy-stream fallback test in the PR itself so the 
empty-env behavior on restore is covered in CI rather than only locally.
   
   Once the `--check` contract is documented, I'm happy to move forward with 
stage 1.
   
   <!-- streview-comment:2319 -->


-- 
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