| From: | Noah Misch <noah(at)leadboat(dot)com> |
|---|---|
| To: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> |
| Cc: | Andres Freund <andres(at)2ndquadrant(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com>, Jim Nasby <jim(at)nasby(dot)net>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: dynamic shared memory |
| Date: | 2013-09-30 23:33:51 |
| Message-ID: | 20130930233351.GB125986@tornado.leadboat.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Tue, Sep 24, 2013 at 08:58:36AM +0530, Amit Kapila wrote:
> On Tue, Sep 24, 2013 at 12:32 AM, Noah Misch <noah(at)leadboat(dot)com> wrote:
> > I don't know whether writing it as binary will help or hurt that situation.
> > If nothing else, binary gives you one less variation to think about when
> > studying the code.
>
> In that case, shouldn't all other places be consistent. One reason I
> had in mind for
> using appropriate mode is that somebody reading code can tomorrow come
> up with a question or a
> patch to use correct mode, then we will again be in same situation.
There are cases that must use binary I/O (table data files), cases that
benefit notably from text I/O (log files, postgresql.conf), and cases where it
doesn't matter too much (dsm state file, postmaster.pid). I don't see a need
to make widespread changes to other call sites.
--
Noah Misch
EnterpriseDB http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Ants Aasma | 2013-10-01 00:51:50 | Re: [PERFORM] Cpu usage 100% on slave. s_lock problem. |
| Previous Message | Peter Geoghegan | 2013-09-30 22:45:47 | Re: INSERT...ON DUPLICATE KEY LOCK FOR UPDATE |