Re: New compiler warnings in buildfarm - Mailing list pgsql-hackers

From Andres Freund
Subject Re: New compiler warnings in buildfarm
Date
Msg-id 20240731183256.5y6ed4riir2oe7na@awork3.anarazel.de
Whole thread Raw
In response to Re: New compiler warnings in buildfarm  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: New compiler warnings in buildfarm
List pgsql-hackers
Hi,

On 2024-07-31 10:11:07 -0400, Tom Lane wrote:
> Peter Eisentraut <peter@eisentraut.org> writes:
> > On 30.07.24 18:19, Tom Lane wrote:
> >> Sometime in the last month or so, flaviventris's bleeding-edge
> >> version of gcc has started whining[1] about truncation of a
> >> string literal's implicit trailing '\0' in contexts like this:
> >> ../pgsql/src/backend/commands/copyto.c:106:41: warning:
> >> initializer-string for array of 'char' is too long
> >> [-Wunterminated-string-initialization]
> >> 106 | static const char BinarySignature[11] = "PGCOPY\n\377\r\n\0";
> >> |                                         ^~~~~~~~~~~~~~~~~~~~
>
> > According to the gcc documentation, this warning is part of -Wextra.
> > And indeed flaviventris runs with -Wextra:
>
> > 'CFLAGS' => '-O1 -ggdb -g3 -fno-omit-frame-pointer -Wall -Wextra
> > -Wno-unused-parameter -Wno-sign-compare -Wno-missing-field-initializers
> > -O0',
>
> Ah --- and it was not doing that a month ago.

Hm? I've not touched flaviventris config since at least the 26th of March. And
a buildfarm run from before then also shows -Wextra:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=flaviventris&dt=2024-03-17%2011%3A17%3A01


> So maybe the compiler has had this warning for longer.

It's very new:

commit 44c9403ed1833ae71a59e84f9e37af3182be0df5
Author:     Alejandro Colomar <alx@kernel.org>
AuthorDate: 2024-06-29 15:10:43 +0200
Commit:     Martin Uecker <uecker@gcc.gnu.org>
CommitDate: 2024-07-14 11:41:00 +0200

    c, objc: Add -Wunterminated-string-initialization


It might be worth piping up in the gcc bugtracker and suggesting that the
warning isn't issued when there's an explicit \0?


> > So I think the appropriate fix here for now is to add
> > -Wno-unterminated-string-initialization to this buildfarm configuration.
>
> Agreed; our policy so far has been to treat -Wextra warnings with
> suspicion, and there is not anything really wrong with these bits
> of code.
>
> It looks like serinus needs this fix too.

Added to both.  I've forced runs for both animals, so the bf should show
results of that soon.

Greetings,

Andres Freund



pgsql-hackers by date:

Previous
From: Christophe Pettus
Date:
Subject: Re: CI, macports, darwin version problems
Next
From: Andres Freund
Date:
Subject: Re: New compiler warnings in buildfarm