vbhanuchander-lang commented on PR #17628:
URL: https://github.com/apache/iceberg/pull/17628#issuecomment-5274568517
**Correction to my own description above, and full prior art.** I wrote that
"six weeks had
passed with no linked PR". That was wrong, and I should have caught it
before opening this — I
apologise. There are two earlier PRs against #17070:
- **#17071 by @eye-gu** (2026-07-03) — the same one-character fix *plus* a
test. Complete work,
opened the day after the issue. It received **no human review** and the
stale bot closed it on
2026-08-11, one day before I opened this.
- **#17130 by @Mahnoor-Zaffar** (2026-07-07) — still open, an auto-generated
`[Auto-fix]` PR. It
changes the production line only, with no test.
My Gate-A check missed both: I searched for `#17070` and #17130 references
the issue by full URL
rather than by number, while #17071 was already closed.
So the credit is @eye-gu's — the diagnosis and the fix are theirs, and I
have said so in the commit
message. I have also **taken their seven boundary pairs into
`testDoubleOrderingForBoundaryPairs`** with attribution, because they cover
things my sweeps did
not: `-Double.MIN_VALUE` vs `0.0`, `0.0` vs `Double.MIN_VALUE`, and
`±Double.MAX_VALUE`. With those
included, **3 of the 4 new tests fail without the production change** (18
tests, 3 failed) and all
18 pass with it.
**How I would like to see this resolved, in order of preference:**
1. Revive **#17071** and merge that — it is @eye-gu's work and it was first.
I will happily close
this PR. If it helps, my three sweep tests and the note on why the
existing random tests miss
the bug can be moved over there, or dropped entirely; the one-line
production fix is what
matters and it is identical in all three PRs.
2. Merge this one, which is the union of both test approaches, and close
#17130.
3. Merge #17130 for the fix alone — though I would push back gently on that,
since it lands the
change with no test, and the reason this bug survived for so long is
precisely that the existing
tests cannot reach it.
Whichever a committer prefers, the substantive point is that the same
one-character fix has now
been independently derived three times and has sat unreviewed for six weeks.
It would be good to
land one of them.
--
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]
