Horton

Re: [HACKERS] Cutting initdb's runtime (Perl question embedded)

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@lists.postgresql.org
Date: 2018-09-12T07:38:30Z
Lists: pgsql-hackers 
On Fri, Aug 10, 2018 at 02:31:25PM -0400, Tom Lane wrote:
> The cfbot points out that this has suffered bit-rot, so here's a rebased
> version --- no substantive changes.

+   /*
+    * Try to read the symlink.  If not there, not a symlink, etc etc, just
+    * quietly fail; the precise reason needn't concern us.
+    */
+   len = readlink(linkname, link_target, sizeof(link_target));

One thing that I can see changing with this patch is how timezone is set
in postgresql.conf.  For example, on HEAD I get 'Japan' while this patch
gives back 'Asia/Tokyo'.  Could it be an issue for countries with
multiple timezones?  I am not sure how Russian systems would react on
that for example.

The time cut for initdb is interesting for buildfarm members anyway, and
the patch looks in nice shape to me.
--
Michael

Commits

  1. Attempt to identify system timezone by reading /etc/localtime symlink.

  2. Make sure to run one initdb TAP test with no TZ set

  3. Use one transaction while reading postgres.bki, not one per line.

  4. Move bootstrap-time lookup of regproc OIDs into genbki.pl.