Mailing-List: contact commits-help@subversion.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@subversion.apache.org
Received-SPF: pass (athena.apache.org: local policy)
Message-Id: <1321191811.10712.140660998285445@webmail.messagingengine.com>
From: "Daniel Shahaf" <d.s@daniel.shahaf.name>
To: dev@subversion.apache.org
Cc: commits@subversion.apache.org, philip@apache.org
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset="us-ascii"
References: <20111111180752.6FDE32388978@eris.apache.org>
Subject: Re: svn commit: r1201002 -
 /subversion/trunk/subversion/libsvn_diff/diff_file.c
In-Reply-To: <20111111180752.6FDE32388978@eris.apache.org>
Date: Sun, 13 Nov 2011 15:43:31 +0200


On Friday, November 11, 2011 6:07 PM, philip@apache.org wrote:
> Author: philip
> Date: Fri Nov 11 18:07:52 2011
> New Revision: 1201002
> 
> URL: http://svn.apache.org/viewvc?rev=1201002&view=rev
> Log:
> Fix an unitialised memory read identified by valgrind.  This occurs
> during merge_tests.py 35.
> 
> * subversion/libsvn_diff/diff_file.c
>   (find_identical_prefix): Don't look beyond endp.
> 
> Modified:
>     subversion/trunk/subversion/libsvn_diff/diff_file.c
> 
> Modified: subversion/trunk/subversion/libsvn_diff/diff_file.c
> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_diff/diff_file.c?rev=1201002&r1=1201001&r2=1201002&view=diff
> ==============================================================================
> --- subversion/trunk/subversion/libsvn_diff/diff_file.c (original)
> +++ subversion/trunk/subversion/libsvn_diff/diff_file.c Fri Nov 11 18:07:52 2011
> @@ -470,8 +470,9 @@ find_identical_prefix(svn_boolean_t *rea
>           too many for the \r. */
>        svn_boolean_t ended_at_nonmatching_newline = FALSE;
>        for (i = 0; i < file_len; i++)
> -        ended_at_nonmatching_newline = ended_at_nonmatching_newline
> -                                       || *file[i].curp == '\n';
> +        if (file[i].curp < file[i].endp)
> +          ended_at_nonmatching_newline = ended_at_nonmatching_newline
> +                                         || *file[i].curp == '\n';

Does the comment saying:

       * Determine how far we may advance with chunky ops without reaching
       * endp for any of the files.

needs to be updated? (s/reaching/exceeding/)

>        if (ended_at_nonmatching_newline)
>          {
>            lines--;
> 
> 
>