Rails - Feb 2006 - UUIDTool and weird timestamp split error

Hi,

I am using UUIDTools to generate my Guids. It works fine on my computer 
(Windows Xp), but doesn''t work on my host (TextDrive)

The error is:

Loading production environment.
>> UUID.timestamp_create.to_s
/sbin/ifconfig: Permission denied
NoMethodError: private method `split'' called for nil:NilClass
        from ./script/../config/../config/../lib/uuidtools.rb:236:in 
`timestamp_create''
        from ./script/../config/../config/../lib/uuidtools.rb:226:in 
`synchronize''
        from ./script/../config/../config/../lib/uuidtools.rb:226:in 
`timestamp_create''
        from (irb):8
>>

Any ideas what this could be? I am using ver 1.0 of UUIDtools.

Any help would be a life saver
Joerg

-- 
Posted via http://www.ruby-forum.com/.

Hi-

I''m completely new to rails - and ruby for that matter - but your
problem seems to be that UUIDTools try getting the MAC Address of
your network card by calling the program /sbin/ifconfig , which the user rails
runs under cannot execute.
Some soultions:
- Convince your provider to give oyur rails user execute permission on
/sbin/ifconfig
- install ifconfig for your rails user (you might also have to convince the UUID
code to use your local copy of ifconfig instad)
- convince the maintainer of the UUIDTools to implement another way to retrieve
the MAC address
- overwrite UUID.get_mac_address to return teh mac adress from a config file
instead of actually getting it from the network card

HTH,
-markus
> -----Original Message-----
> From: rails-bounces@lists.rubyonrails.org 
> [mailto:rails-bounces@lists.rubyonrails.org] On Behalf Of 
> Joerg Diekmann
> Sent: Wednesday, February 22, 2006 10:34 AM
> To: rails@lists.rubyonrails.org
> Subject: [Rails] UUIDTool and weird timestamp split error
> 
> Hi,
> 
> I am using UUIDTools to generate my Guids. It works fine on 
> my computer 
> (Windows Xp), but doesn''t work on my host (TextDrive)
> 
> The error is:
> 
> Loading production environment.
> >> UUID.timestamp_create.to_s
> /sbin/ifconfig: Permission denied
> NoMethodError: private method `split'' called for nil:NilClass
>         from ./script/../config/../config/../lib/uuidtools.rb:236:in 
> `timestamp_create''
>         from ./script/../config/../config/../lib/uuidtools.rb:226:in 
> `synchronize''
>         from ./script/../config/../config/../lib/uuidtools.rb:226:in 
> `timestamp_create''
>         from (irb):8
> >>
> 
> 
> Any ideas what this could be? I am using ver 1.0 of UUIDtools.
> 
> Any help would be a life saver
> Joerg
> 
> -- 
> Posted via http://www.ruby-forum.com/.
> _______________________________________________
> Rails mailing list
> Rails@lists.rubyonrails.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

Cool. Just found another way ... you can use UUID.random_create to get a 
Guid ... doesn''t use the MAC address.


Markus Strickler wrote:
> Hi-
> 
> I''m completely new to rails - and ruby for that matter - but your 
> problem seems to be that UUIDTools try getting the MAC Address of
> your network card by calling the program /sbin/ifconfig , which the user 
> rails runs under cannot execute.
> Some soultions:
> - Convince your provider to give oyur rails user execute permission on 
> /sbin/ifconfig
> - install ifconfig for your rails user (you might also have to convince 
> the UUID code to use your local copy of ifconfig instad)
> - convince the maintainer of the UUIDTools to implement another way to 
> retrieve the MAC address
> - overwrite UUID.get_mac_address to return teh mac adress from a config 
> file instead of actually getting it from the network card
> 
> HTH,
> -markus

-- 
Posted via http://www.ruby-forum.com/.

Joerg Diekmann wrote:
> Cool. Just found another way ... you can use UUID.random_create to get a 
> Guid ... doesn''t use the MAC address.
Doesn''t that defeat the purpose?  I thought the only way you could
guarantee
that an ID was truly unique was to include another unique ID, e.g. the MAC 
address, CPU ID, or some other hardware serial number.

A random ID, even combined with date/time, could still theoretically collide 
(e.g. if two machines happened to pick the same random value at the same time).

-- Steve

Steve Sloan wrote:
> A random ID, even combined with date/time, could still theoretically
> collide (e.g. if two machines happened to pick the same random value
> at the same time).
To have a ~50% chance of a collision, you must generate ~2^64 random
128-bit numbers. Considering the small subset of those 2^64 that will
actually be generated, the chance of duplicates happening randomly can
be ignored --- and its far less than the chance of either bugs in the
implementation or even duplicate MAC addresses accidentally being
assigned.