| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Sami Imseih <samimseih(at)gmail(dot)com> |
| Cc: | David Steele <david(at)pgbackrest(dot)org>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>, Anthonin Bonnefoy <anthonin(dot)bonnefoy(at)datadoghq(dot)com>, jian he <jian(dot)universality(at)gmail(dot)com> |
| Subject: | Re: Regression in statement locations |
| Date: | 2025-05-19 23:38:47 |
| Message-ID: | aCvBB88wk4TWdKfO@paquier.xyz |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, May 19, 2025 at 05:10:14PM -0500, Sami Imseih wrote:
> I am still not sure why this is the case, but wanted to share this
> for now.
Hmm. Something seems to not be compiling well for the final query of
a stmtmulti in gram.y with updateRawStmtEnd(), as we rely on the
position of the semicolon to decide what the stmt_len should be with
multiple querues. We don't set the stmt_len without the semicolon,
causing pgss to reuse the full query length when storing the entry of
the last query.
With the semicolon in place, stmt_len gets set for the last query of
the string. Still digging more..
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Matthew Sterrett | 2025-05-19 23:54:09 | Re: Proposal for enabling auto-vectorization for checksum calculations |
| Previous Message | Jim Jones | 2025-05-19 23:11:07 | Re: [PoC] XMLCast (SQL/XML X025) |