From: | <korryd(at)enterprisedb(dot)com> |
---|---|
To: | "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | "Mark Kirkwood" <markir(at)paradise(dot)net(dot)nz>, "Bruce Momjian" <bruce(at)momjian(dot)us>, <pgsql-patches(at)postgresql(dot)org> |
Subject: | Re: [pgsql-patches] Patch to avoid gprofprofilingoverwrites |
Date: | 2007-02-01 14:19:53 |
Message-ID: | 1170339593.6941.122.camel@sakai.localdomain |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
> > Right - but LINUX_PROFILE was added to correct Linux specific oddities
> > with the time counter accumulation, whereas your patch is not Linux
> > specific at all. So I think a more representative symbol is required.
>
> Yeah, that was my problem with the patch too. The issue that it's
> addressing isn't Linux-specific in the least.
>
> Is there a way to detect via #if that profiling is enabled? I wouldn't
> expect a really portable answer, but maybe there's something that works
> for gcc?
You're right - I hadn't really looked beyond the problem that I was
trying to solve.
The technique should work for gprof on other platforms, but, as you
point out, LINUX_PROFILE is unique to Linux (no, I hadn't noticed that,
but it seems pretty obvious in retrospect).
I like Neil's idea of a using --enable-profiling configure flag. Every
time I need to profile, I have to go search the archives for 'gprof' -
it would be nice to see --enable-profiling when I configure --help.
-- Korry
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2007-02-01 14:33:15 | Re: [pgsql-patches] Patch to avoid gprof profilingoverwrites |
Previous Message | Peter Eisentraut | 2007-02-01 08:12:54 | Re: [pgsql-patches] Patch to avoid gprof profilingoverwrites |