DanielLeens commented on PR #12619:
URL: https://github.com/apache/seatunnel/pull/12619#issuecomment-6411593403

   Thanks @goutamadwant for the checklist, and @SEZ9 for the focused questions. 
I re-checked the unchanged head `441ee3e5f8331c0038f7d1fb2a024087f4ce20ac` 
(source reading only, nothing executed locally).
   
   F1, F2, F3/F5, F6, F7 and F8: the exact pointers are in my 2026-10-18 
comment above (the `JobLineageProjector` per-node and per-table-path 
retained-bytes checks, `acceptsAnyNonNullJobId`, 
`reportsDefaultOnlyPathAsUnavailable`, the `JobLineageException` Javadoc, the 
`JobLineageLimits.DEFAULT` Javadoc, and `requiresLimits`), so I will not repeat 
them. One detail that may help for F3/F5: in `JobLineageProjector` the loop 
over `vertex.getTablePaths()` skips `null` and `TablePath.DEFAULT` entries and 
sets `pathsOmitted`, and only non-transform vertices go through that loop, so 
an omitted-only path set cannot end up as `REPORTED`.
   
   On F4 (a), the TRANSFORM / `NOT_APPLICABLE` branch is exercised against real 
engine output: `projectsTransformFromPhysicalDagInfo` builds the logical plan 
from `lineage_transform.conf` via `TestUtils.createTestLogicalPlan`, passes it 
through `DAGUtils.getJobDAGInfo(...)`, and asserts one `SOURCE`, one 
`TRANSFORM` (name contains 
`Transform[0]-FilterRowKind->Transform[1]-FilterRowKind`, `datasetMetadata` is 
`NOT_APPLICABLE`, no table paths) and one `SINK`, the two edges, zero warnings, 
and that a second projection of a freshly built DAG is identical. 
`projectsSnapshotsGeneratedByTheEngine` does the same for 
`fake_to_console.conf` (4 nodes, sorted edges, no leaked address). Both go 
through `DAGUtils`, so they exercise the shape the engine actually produces.
   
   On F4 (b), I can only confirm the covered side: the tests above use the 
physical DAG info built by `DAGUtils`. Which logical-DAG scenario is deferred, 
and what the existing bug is, is something only @goutamadwant can spell out. A 
link to a tracking issue from the checklist would make the deferral traceable, 
as @SEZ9 asked.
   
   From my side there is still no source-level blocker on this head, and the 
required Build check is green. This is a targeted reply, not a new full review; 
the remaining gates are @SEZ9's verification and a committer's approval.
   


-- 
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