Re: Compiler warnings

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Joe Conway <mail(at)joeconway(dot)com>
Cc: Stephen Frost <sfrost(at)snowman(dot)net>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Compiler warnings
Date: 2017-01-02 18:18:29
Message-ID: CA+TgmoYeuEny2R0iUyy8zy9909VkcUQxrj6+nKzodJpUeUx-0A@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Dec 29, 2016 at 10:47 AM, Joe Conway <mail(at)joeconway(dot)com> wrote:
> Shouldn't this be back-patched? The plancache warning goes back through
> 9.2 (at least) and the lwlocks warning through 9.5 (or maybe it was 9.4).

Warnings are going to be different for each individual developer, but
I am cautiously in favor making more of an effort to fix back-branch
warnings provided that it doesn't generate too much code churn. For
example, if your toolchain generates only these two warnings on 9.2
then, sure, let's back-port these two fixes; making things
warning-clean is great. But if there are dozens or hundreds of
warnings currently, fixing only a handful of those warnings probably
isn't valuable, and fixing all of them is probably a little more risk
than we necessarily want to take. Someone could goof and make a bug.
On my MacBook Pro with my toolchain, we're warning-clean back to 9.3
or 9.4, and before that there are some problems -- most annoyingly the
fact that 73b416b2e41237b657d29d8f42a4bb34bf700928 couldn't be easily
backported to older branches. I don't think it would be crazy to try
to get all of the warnings I see fixed up and it would be convenient
for me, but I haven't been willing to do the work, either.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2017-01-02 18:25:35 Re: Cluster wide option to control symbol case folding
Previous Message Ryan Murphy 2017-01-02 18:14:34 Re: Clarifying "server starting" messaging in pg_ctl start without --wait