Aaron Zang wrote:
> Alan Coopersmith wrote:
>> Aaron Zang wrote:
>>> Alan Coopersmith wrote:
>>>> Gary Winiger wrote:
>>>>> The vtdaemon service proposed a "rootunlock" property.  When the
>>>>> value of
>>>>> "rootunlock" was "true", vtdaemon allowed unlocking text virtual
>>>>> consoles
>>>>> using the root user's password instead of the locking user's password.
>>>>>
>>>>> The project team now considers the "rootunlock" property as 
>>>>> unnecessary
>>>>> because:
>>>>>
>>>>>     1) Neither xlock nor xscreensaver have such an unlocking feature.
>>>> I don't understand this claim - xlock, xscreensaver & CDE lockscreen
>>>> all have
>>>> a feature to allow unlocking the screen with the root password instead
>>>> of the
>>>> locking's users password, mostly because our customers demanded it be
>>>> there.
>>>> (Whether it's on or off by default differs by OS rev & program, but 
>>>> they
>>>>  all have it - see "-allowroot" in xlock(1), "Dtsession*keys" in
>>>> dtsession(1),
>>>>  and the allowRoot xscreensaver resource defined in LSARC 2006/446 
>>>> which
>>>>  we apparently forgot to document in the man page.)
>>>>
>>> Yes, these locks all have code supporting root unlock, and there are 
>>> macros
>>> functioning as switches to turn the feature on and off during 
>>> compilation.
>>> Current Solaris releases always turn off the switches (without compiling
>>> this
>>> feature in). That what we really mean.
>>
>> That is still wrong - they are all still built with the feature present.
>> In current Nevada builds the feature is not enabled by default, but
>> can be turned on at runtime, without a recompile, by just setting
>> the flag/properties mentioned.
>>
> 
> OK, I see. I tried xlock -allowroot on my system, it did work.
> And I found a "AllowRoot" entry in ~/.xscreensaver with "False" as default.
> 
> It seems that we should refine our claim like this:
> "Neither xlock nor xscreensaver support such an unlocking feature by 
> default"
> Is it correct?
> 
> Anyway, I still believe that the trend is to not using rootunlock. I still
> remember that around Nevada build 30, xscreensaver supported rootunlock by
> default, and now it is not the default behavior.

IMO it would make sense to retain the capability, but turn it off by default.

Michael
-- 
Michael Schuster        http://blogs.sun.com/recursion
Recursion, n.: see 'Recursion'

Reply via email to