Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 12 Sep 2004 22:05:13 -0700
From:      Brooks Davis <brooks@one-eyed-alien.net>
To:        Dan Lukes <dan@obluda.cz>
Cc:        freebsd-bugs@freebsd.org
Subject:   Re: bin/71630: [PATCH] cleanup of the usr.sbin/pppd code
Message-ID:  <20040913050513.GA22473@odin.ac.hmc.edu>
In-Reply-To: <200409130110.i8D1AN1w061311@freefall.freebsd.org>
References:  <200409130110.i8D1AN1w061311@freefall.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help

--0F1p//8PRICkK4MW
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Sep 13, 2004 at 01:10:23AM +0000, Dan Lukes wrote:
> The following reply was made to PR bin/71630; it has been noted by GNATS.
>=20
> From: Dan Lukes <dan@obluda.cz>
> To: Giorgos Keramidas <keramida@freebsd.org>
> Cc: bug-followup@freebsd.org
> Subject: Re: bin/71630: [PATCH] cleanup of the usr.sbin/pppd code
> Date: Mon, 13 Sep 2004 03:07:18 +0200 (CEST)
>=20
>  On Sun, 12 Sep 2004, Giorgos Keramidas wrote:
> =20
>  > On 2004-09-12 04:38, Dan Lukes <dan@obluda.cz> wrote:
>  >> -     FILE *iffile;
>  >> +     FILE *iffile =3D iffile; /* to avoid "may be used unitialized" =
warning */
>  >
>  > Nope.  NULL is the proper initialization value for (FILE *) objects.
>  > This way any bugs that might exist further down will be exposed by a
>  > dereference of a NULL pointer which everyone will notice.  Hiding the
> =20
>   	True initialisation is not necesarry with current code, so
>  it's waste of resources now.
> =20
>   	I like defensive style of programming, but it has negative impact on
>  performance.
> =20
>   	Somebody should evaluate if we can live with unnecesarry
>  initialisation here. If yes, it's better patch to iffile=3DNULL, if not,=
 my
>  patch is better.

I find it extreamly hard to believe you could demonstrate the
performance difference in a microbenchmark, much less a real
application.  Frankly, unless you can prove a real initalization causes
a real performance problem, v=3Dv assignments should not be used because
they are bogus by inspection.

-- Brooks

--=20
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

--0F1p//8PRICkK4MW
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFBRSqIXY6L6fI4GtQRAonkAKCSIar0VX7wrD8bOaKVQ5fLr8wVBwCg54pY
LQMCX5bNudzFqVSLVFwRRDU=
=ejYQ
-----END PGP SIGNATURE-----

--0F1p//8PRICkK4MW--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040913050513.GA22473>