GNU bug report logs - #20029
'yes' surprisingly slow

Previous Next

Full log

Message #14 received at 20029 <at> debbugs.gnu.org (full text, mbox):

From: Eric Blake <eblake <at> redhat.com>
To: Pádraig Brady <P <at> draigBrady.com>,
 Ole Tange <tange <at> gnu.org>, 20029 <at> debbugs.gnu.org
Subject: Re: bug#20029: 'yes' surprisingly slow
Date: Mon, 09 Mar 2015 14:02:16 -0600

[Message part 1 (text/plain, inline)]

On 03/09/2015 01:47 PM, Pádraig Brady wrote:

> 
> Note this change also ensures that yes will only write complete lines
> for lines softer than BUFSIZ.

s/softer/smaller/

> 
> * src/yes.c (main): Build up a BUFSIZ buffer of lines,
> and output that, rather than having stdio process each item.
> * tests/misc/yes.sh: Add a new test for various buffer sizes.
> * tests/local.mk: Reference the new test.
> Fixes http://bugs.gnu.org/20029
> ---
>  src/yes.c         | 43 +++++++++++++++++++++++++++++++++++++++++--
>  tests/local.mk    |  1 +
>  tests/misc/yes.sh | 28 ++++++++++++++++++++++++++++
>  3 files changed, 70 insertions(+), 2 deletions(-)
>  create mode 100755 tests/misc/yes.sh
> 
> diff --git a/src/yes.c b/src/yes.c
> index b35b13f..91dea11 100644
> --- a/src/yes.c
> +++ b/src/yes.c
> @@ -58,6 +58,10 @@ Repeatedly output a line with all specified STRING(s), or 'y'.\n\
>  int
>  main (int argc, char **argv)
>  {
> +  char buf[BUFSIZ];

Do you really want this stack-allocated?  BUFSIZ can be larger than a
page, which can then interfere with stack overflow detection.

> +. "${srcdir=.}/tests/init.sh"; path_prepend_ ./src
> +print_ver_ yes
> +
> +for size in 1 4095 4096 8191 8192 16383 16384; do

Should you also test 4097 8193 and 16385 (that is, a likely
1-more-than-BUFSIZ in the mix)?

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[signature.asc (application/pgp-signature, attachment)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.