NetBSD Problem Report #27140

Received: (qmail 13872 invoked by uid 605); 4 Oct 2004 14:04:47 -0000
Message-Id: <20041004140411.05C82294A2@dmath5.geometrie.tuwien.ac.at>
Date: Mon,  4 Oct 2004 16:04:11 +0200 (CEST)
From: wiz@NetBSD.org
Sender: gnats-bugs-owner@NetBSD.org
Reply-To: wiz@NetBSD.org
To: gnats-bugs@gnats.NetBSD.org
Subject: "sleep time" doesn't error out
X-Send-Pr-Version: 3.95

>Number:         27140
>Category:       bin
>Synopsis:       "sleep time" doesn't error out
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    bin-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Oct 04 14:05:00 +0000 2004
>Closed-Date:    Mon Oct 17 10:12:20 +0000 2005
>Last-Modified:  Fri Mar 30 09:30:07 +0000 2012
>Originator:     Thomas Klausner
>Release:        NetBSD 2.0_BETA
>Organization:

>Environment:


Architecture: i386
Machine: i386
>Description:
"sleep time" produces no error, even though sleep(1) gets a non-numeric
argument.
>How-To-Repeat:
# sleep time
#
>Fix:
Fix the argument handling in sleep(1); not provided.
>Release-Note:
>Audit-Trail:

From: SODA Noriyuki <soda@sra.co.jp>
To: wiz@NetBSD.org
Cc: gnats-bugs@gnats.NetBSD.org
Subject: Re: bin/27140: "sleep time" doesn't error out
Date: Mon, 4 Oct 2004 23:44:59 +0900 (JST)

 >>>>> On Mon,  4 Oct 2004 16:04:11 +0200 (CEST), wiz@NetBSD.org said:

 > "sleep time" produces no error, even though sleep(1) gets a non-numeric

 Isn't this trivial to fix?
 e.g.

 Index: sleep.c
 ===================================================================
 RCS file: /cvsroot/src/bin/sleep/sleep.c,v
 retrieving revision 1.19
 diff -u -r1.19 sleep.c
 --- sleep.c	7 Aug 2003 09:05:41 -0000	1.19
 +++ sleep.c	4 Oct 2004 14:38:09 -0000
 @@ -60,10 +60,10 @@
  int
  main(int argc, char *argv[])
  {
 -	char *arg, *temp;
 -	double fval, ival, val;
 +	char *ep;
 +	double ival, val;
  	struct timespec ntime;
 -	int ch, fracflag;
 +	int ch;

  	setprogname(argv[0]);
  	(void)setlocale(LC_ALL, "");
 @@ -83,37 +83,30 @@
  		usage();

  	/*
 -	 * Okay, why not just use atof for everything? Why bother
 +	 * Okay, why not just use strtod() for everything? Why bother
  	 * checking if there is a fraction in use? Because the old
  	 * sleep handled the full range of integers, that's why, and a
  	 * double can't handle a large long. This is fairly useless
  	 * given how large a number a double can hold on most
  	 * machines, but now we won't ever have trouble. If you want
  	 * 1000000000.9 seconds of sleep, well, that's your
 -	 * problem. Why use an isdigit() check instead of checking for
 -	 * a period? Because doing it this way means locales will be
 -	 * handled transparently by the atof code.
 +	 * problem.
  	 */
 -	fracflag = 0;
 -	arg = *argv;
 -	for (temp = arg; *temp != '\0'; temp++)
 -		if (!isdigit((unsigned char)*temp))
 -			fracflag++;
 -
 -	if (fracflag) {
 -		val = atof(arg);
 -		if (val <= 0)
 +	ntime.tv_sec = strtol(argv[0], &ep, 10);
 +	if (argv[0][0] != '\0' && ep[0] == '\0') {
 +		if (ntime.tv_sec < 0)
 +			exit(0);
 +		ntime.tv_nsec = 0;
 +	} else {
 +		val = strtod(argv[0], &ep);
 +		if (argv[0][0] == '\0' || ep[0] != '\0' ||
 +		    isinf(val) || isnan(val))
 +			    errx(2, "invalid time interval %s", argv[0]);
 +		if (val < 0)
  			exit(0);
  		ival = floor(val);
 -		fval = (1000000000 * (val-ival));
  		ntime.tv_sec = ival;
 -		ntime.tv_nsec = fval;
 -	}
 -	else{
 -		ntime.tv_sec = atol(arg);
 -		if (ntime.tv_sec <= 0)
 -			exit(0);
 -		ntime.tv_nsec = 0;
 +		ntime.tv_nsec = (1000000000 * (val - ival));
  	}

  	if (nanosleep(&ntime, NULL) == -1)

 P.S.
 Maybe it's better to convert isinf() case to LONG_MAX, but it seems
 currently our nanosleep() doesn't support LONG_MAX as tv_sec.
 --
 soda

From: "Dheeraj Reddy" <dheeraj@ece.gatech.edu>
To: gnats-bugs@netbsd.org
Cc:  
Subject: Re: bin/27140
Date: Mon, 4 Oct 2004 18:48:22 -0400

 Or maybe just the following patch does it ?

 Index: sleep.c
 ===================================================================
 RCS file: /cvsroot/src/bin/sleep/sleep.c,v
 retrieving revision 1.19
 diff -r1.19 sleep.c
 105c105,106
 <               if (val <= 0)
 ---
 >               if (val <= 0) {
 >                       usage();
 106a108
 >               }

 -- 
 It is better to be silent and be thought a fool than to speak and remove
 all doubt
From: Elad Efrat <elad@netbsd.org>
To: gnats-bugs@netbsd.org
Cc: 
Subject: PR/27140 CVS commit: src/bin/sleep
Date: Mon, 17 Oct 2005 10:11:46 +0000 (UTC)

 Module Name:	src
 Committed By:	elad
 Date:		Mon Oct 17 10:11:46 UTC 2005

 Modified Files:
 	src/bin/sleep: sleep.c

 Log Message:
 PR/27140: "sleep time" doesn't error out.
 Apply fix from Dheeraj Reddy.


 To generate a diff of this commit:
 cvs rdiff -r1.19 -r1.20 src/bin/sleep/sleep.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

State-Changed-From-To: open->closed
State-Changed-By: elad@netbsd.org
State-Changed-When: Mon, 17 Oct 2005 10:12:20 +0000
State-Changed-Why:
Fixed, thanks.


From: "Jukka Ruohonen" <jruoho@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/27140 CVS commit: src
Date: Fri, 30 Mar 2012 09:27:11 +0000

 Module Name:	src
 Committed By:	jruoho
 Date:		Fri Mar 30 09:27:10 UTC 2012

 Modified Files:
 	src/distrib/sets/lists/tests: mi
 	src/etc/mtree: NetBSD.dist.tests
 Added Files:
 	src/tests/bin/sleep: Makefile t_sleep.sh

 Log Message:
 Add regression tests for PR bin/3914 and PR bin/27140.


 To generate a diff of this commit:
 cvs rdiff -u -r1.461 -r1.462 src/distrib/sets/lists/tests/mi
 cvs rdiff -u -r1.70 -r1.71 src/etc/mtree/NetBSD.dist.tests
 cvs rdiff -u -r0 -r1.1 src/tests/bin/sleep/Makefile \
     src/tests/bin/sleep/t_sleep.sh

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

>Unformatted:

Home	PR Database Search