DanielLeens commented on PR #10988:
URL: https://github.com/apache/seatunnel/pull/10988#issuecomment-5229130726

   Thanks @SEZ9 for the independent full pass on the current head. A few notes 
tying this back to where things stand:
   
   +1 on Issue 1 (docs gap) and Issue 2 (regex FROM-clause fallback fragility 
for joins/subqueries, plus the MySQL-style backtick in the character class) — 
these line up with what I already flagged as my own non-blocking Issue 2 and 
Issue 4 in my 2026-08-06 review, so good to have independent confirmation on 
both.
   
   Issue 3 (bare `Integer.parseInt` on the dimension string) and Issue 5 
(fully-qualified `java.nio.ByteBuffer` instead of an import) are new catches 
from your pass that I didn't call out — agreed on both, worth folding into the 
same fix commit. Issue 4 (PR description/checklist placeholders) is a process 
nit, agreed as well.
   
   One thing I want to make sure doesn't get lost in the combined list: none of 
Issue 1-5 here cover the actual merge blocker from my 2026-08-06 
`CHANGES_REQUESTED` review — `PostgresTypeConverter.convert()` at 
`PostgresTypeConverter.java:249-251` still sets `builder.sourceType(PG_VECTOR)` 
as the bare literal `"vector"` instead of baking the dimension in (the pattern 
every other parameterized type in this file already follows, e.g. 
`varchar(255)`). Since `PostgresCreateTableSqlBuilder.buildColumnSql()` prefers 
`column.getSourceType()` verbatim whenever source and sink share the same 
PostgreSQL catalog identifier, the Postgres→Postgres auto-create-sink path — 
the exact scenario `JdbcPostgresPgVectorIT#testPgVectorAutoCreateSink` is built 
to verify — will still emit `vector` instead of `vector(N)`. There's been no 
new commit on this PR since I raised that, so it remains open and is still the 
primary blocker for me.
   
   So, from my side: please fold Issues 3 and 5 above into the same pass as my 
Issue 1 fix (dimension into `sourceType`), plus docs and the regex-fallback 
hardening we're both flagging, and I'm happy to do another full re-review once 
that lands.


-- 
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