Re: [BUG] CRASH: ECPGprepared_statement() and ECPGdeallocate_all() when connection is NULL

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Shruthi Gowda <gowdashru(at)gmail(dot)com>
Cc: SATYANARAYANA NARLAPURAM <satyanarlapuram(at)gmail(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Nishant Sharma <nishant(dot)sharma(at)enterprisedb(dot)com>, Mahendra Singh Thalor <mahi6run(at)gmail(dot)com>, Fujii Masao <masao(dot)fujii(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL Development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: [BUG] CRASH: ECPGprepared_statement() and ECPGdeallocate_all() when connection is NULL
Date: 2026-05-25 23:34:43
Message-ID: ahTckwyxLba5ofwq@paquier.xyz
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, May 25, 2026 at 01:16:20PM +0530, Shruthi Gowda wrote:
> Thanks Satya for looking into this. However, we don't need to add the
> validation here because the caller ecpg_do_prologue*()* has already taken
> care of it. Adding it again would be redundant. What do you think?

Right, but I think that this points at a different issue: it is
wasteful to call a second time ecpg_get_connection() to get a
connection while ecpg_do_prologue() has already done a cache lookup
for the connection.

Wouldn't it be better to pass directly the connection obtained by the
first call of ecpg_get_connection() as a function argument of
ecpg_auto_prepare(), replacing the connection name? If we need the
extra init() at the end, please show if this can go wrong in the shape
of a regression test.
--
Michael

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Xuneng Zhou 2026-05-26 01:48:26 Re: Implement waiting for wal lsn replay: reloaded
Previous Message jian he 2026-05-25 23:28:06 Re: Rename Postgres 19 to Postgres 26 (year-based)?