GNU bug report logs - #21880
Possible bug in sort --check --key

Previous Next

Full log

View this message in rfc822 format

From: help-debbugs <at> gnu.org (GNU bug Tracking System)
To: Alexander Kindyakov <akindyakov <at> gmail.com>
Subject: bug#21880: closed (Re: bug#21880: Possible bug in sort --check --key)
Date: Wed, 11 Nov 2015 20:02:03 +0000

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

Your bug report

#21880: Possible bug in sort --check --key

which was filed against the coreutils package, has been closed.

The explanation is attached below, along with your original report.
If you require more details, please reply to 21880 <at> debbugs.gnu.org.

-- 
21880: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=21880
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems

[Message part 2 (message/rfc822, inline)]

From: Eric Blake <eblake <at> redhat.com>
To: Alexander Kindyakov <akindyakov <at> gmail.com>, 21880-done <at> debbugs.gnu.org
Subject: Re: bug#21880: Possible bug in sort --check --key
Date: Wed, 11 Nov 2015 13:01:07 -0700

[Message part 3 (text/plain, inline)]

tag 21880 notabug
thanks

On 11/11/2015 11:50 AM, Alexander Kindyakov wrote:
> Hello!
> 
> In sort man page nothing told about working sort --check with '--key'
> option. IMHO this behaviour is strange:

Thanks for the report.

> 
> This case is succeeded:
> 
>  $ echo -e '1\t2\n1\t1' | LC_ALL=C sort --key=1b,1 --check --stable
> --field-separator=$'\t'

'echo -e' is not portable; better is using printf.

Let's see what sort was looking at here (it's a shame that --check and
--debug can't be used at the same time; but seeing _what_ gets sorted
without --check explains whether --check will succeed):

$ printf '1\t2\n1\t1\n' | LC_ALL=C sort --debug -k1b,1 -s -t$'\t'
sort: using simple byte comparison
1>2
_
1>1
_
$

So, in this version, you were checking JUST the first field, and nothing
else; among the portion of the line checked, everything was identical,
so --check has nothing to complain about.

> 
> But this (without option '--stable') is failed:
> 
> $ echo -e '1\t2\n1\t1' | LC_ALL=C sort --key=1b,1 --check
> --field-separator=$'\t'
> sort: -:2: disorder: 1  1

Again, look at the debug output:

$ printf '1\t2\n1\t1\n' | LC_ALL=C sort --debug -k1b,1 -t$'\t'
sort: using simple byte comparison
1>1
_
___
1>2
_
___
$

Notice that the removal of --stable behaves as if you had added another
--key=1 (that is, sort the ENTIRE line as a last-resort key); and now
that you have the entire line involved, --check can indeed see an
out-of-order difference in the input.

> 
> It seems to me both of situation is equal with relation of line key, why
> does the output is different?

The output differs because of your use of --stable.  Therefore, this is
not a bug in sort, and I'm closing this bug.  However, do feel free to
add further comments to this thread.

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

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

[Message part 5 (message/rfc822, inline)]

From: Alexander Kindyakov <akindyakov <at> gmail.com>
To: bug-coreutils <at> gnu.org
Subject: Possible bug in sort --check --key
Date: Wed, 11 Nov 2015 21:50:14 +0300

Hello!

In sort man page nothing told about working sort --check with '--key' 
option. IMHO this behaviour is strange:

This case is succeeded:

 $ echo -e '1\t2\n1\t1' | LC_ALL=C sort --key=1b,1 --check --stable 
--field-separator=$'\t'

But this (without option '--stable') is failed:

$ echo -e '1\t2\n1\t1' | LC_ALL=C sort --key=1b,1 --check 
--field-separator=$'\t'
sort: -:2: disorder: 1  1

It seems to me both of situation is equal with relation of line key, why 
does the output is different?

Sincerely yours, Alexander K

Previous Next

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