Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 27 Mar 2004 04:28:09 -0600
From:      Jon Noack <noackjr@alumni.rice.edu>
To:        Don Bowman <don@sandvine.com>
Cc:        Kris Kennaway <kris@obsecurity.org>
Subject:   Re: LOR status page?
Message-ID:  <40655739.8010507@alumni.rice.edu>
In-Reply-To: <FE045D4D9F7AED4CBFF1B3B813C85337045D8701@mail.sandvine.com>
References:  <FE045D4D9F7AED4CBFF1B3B813C85337045D8701@mail.sandvine.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 3/24/2004 7:15 PM, Don Bowman wrote:
> From: Garance A Drosihn [mailto:drosih@rpi.edu]
>> At 3:46 PM -0800 3/24/04, Kris Kennaway wrote:
>>> On Wed, Mar 24, 2004 at 02:03:28PM -0600, Jon Noack wrote:
>>>> Would it be helpful to put up a web page with all known lock 
>>>> order reversal false positives (or better yet all known lock 
>>>> order reversals with a status indication)?  This would allow 
>>>> people to check there before reporting, saving everyone time.
>>> 
>>> Clearly we need to do something to stop people reporting the same 
>>> non-bugs every day, the problem is that it needs to be somewhere 
>>> people are likely to check.  Maybe a pointer to your proposed 
>>> webpage in UPDATING will help.
>> 
>> Could we do something so we don't PRINT the false-positives?  If 
>> we're about to turn 5.x-current into 5.x-stable, then it is not 
>> good to tell users "Here are a bunch of error messages that you 
>> should just ignore".  At least in my experience, what happens is 
>> that users are much more likely to ignore *all* error messages.
>> 
>> I have no idea what would need to be done, of course.  I'm just 
>> uneasy at telling users to ignore scary-looking error messages.
>> 
>> I do agree that a web page saying exactly which ones to ignore 
>> would be better than expecting end-users to figure that out by 
>> scanning the mailing lists...
> 
> How about make the first line of the error message be: "This 
> technique sometimes produces false positives... See http://.../ for 
> more details"

I like this idea.  Anyone willing to commit such a change?  If so, I'll 
get going on the web page and let you know when it's ready.  Without 
such a blatant pointer I don't think the page would be useful enough to 
warrant creating it.

Jon



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40655739.8010507>