From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | pgsql-committers(at)lists(dot)postgresql(dot)org |
Subject: | pgsql: Reduce memory leakage in initdb. |
Date: | 2023-03-22 18:28:53 |
Message-ID: | E1pf3Ce-004rpK-KK@gemulon.postgresql.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-committers |
Reduce memory leakage in initdb.
While testing commit 3e51b278d, I noted that initdb leaks about a
megabyte worth of data due to the sloppy bookkeeping in its
string-manipulating code. That's not a huge amount on modern machines,
but it's still kind of annoying, and it's easy to fix by recognizing
that we might as well treat these arrays of strings as
modifiable-in-place. There's no caller that cares about preserving
the old state of the array after replace_token or replace_guc_value.
With this fix, valgrind sees only a few hundred bytes leaked during
an initdb run.
Discussion: https://postgr.es/m/2844176.1674681919@sss.pgh.pa.us
Branch
------
master
Details
-------
https://git.postgresql.org/pg/commitdiff/4fe2aa7656dce2bd31d4807a6843ff495b9deb80
Modified Files
--------------
src/bin/initdb/initdb.c | 87 +++++++++++++++++++------------------------------
1 file changed, 34 insertions(+), 53 deletions(-)
From | Date | Subject | |
---|---|---|---|
Next Message | Andrew Dunstan | 2023-03-22 19:33:15 | Re: pgsql: Teach verify_heapam() to validate update chains within a page. |
Previous Message | Tom Lane | 2023-03-22 17:49:13 | pgsql: Add "-c name=value" switch to initdb. |