UDID apocalypse… now?

36 views
Skip to first unread message

Chris Adamson

unread,
Mar 21, 2012, 9:25:35 PM3/21/12
to a2-cocoaheads
Multiple reports from Twitter claim that Apple is rejecting apps and updates that use the UDID (via -[UIDevice uniqueIdentifier]) and even pulling existing apps that use it.

Can't verify, just passing along. I'm blocked on this problem myself, and fully expect tomorrow to be Screaming Panic Nightmare Day.

--Chris

Jeff Kelley

unread,
Mar 21, 2012, 9:32:57 PM3/21/12
to a2-coc...@googlegroups.com
I wouldn't be surprised if libraries such as Flurry, Crashlytics, or Google Analytics used the UDID. Even if they've updated their SDKs to not fetch that information, your apps may have old versions. Best to check before having an unhappy weekend.

Jeff Kelley

Chris Adamson

unread,
Mar 21, 2012, 9:35:13 PM3/21/12
to a2-coc...@googlegroups.com
Flurry sent out an e-mail to all users last Friday saying to update the library and re-submit by Friday March 23. I even featured it in the CFUUID section of one of my CocoaConf talks on Saturday.

Wonder if Flurry got a nudge from Apple?

--Chris

--
You received this message because you are subscribed to the Google Groups "CocoaHeads Ann Arbor" group.
To post to this group, send email to a2-coc...@googlegroups.com.
To unsubscribe from this group, send email to a2-cocoahead...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/a2-cocoaheads?hl=en.

Paul Bruneau

unread,
Mar 21, 2012, 9:43:43 PM3/21/12
to a2-coc...@googlegroups.com
Google analytics totally uses it. I haven't seen them come up with an alternative a la Flurry.

If this is really happening it's too late to save this weekend, but I'm highly skeptical of this. It's very not like Apple to go from "deprecated" to "pull your app" in one OS version. 

But not to say it's impossible

Sent while drivinAAAAAAAGGGGGHHHH!!!!!
--

Chris Adamson

unread,
Mar 21, 2012, 9:45:32 PM3/21/12
to a2-coc...@googlegroups.com
I figured we had until iOS 6, but Apple is nothing if not capricious.

--Chris

Keith Bourne

unread,
Mar 21, 2012, 9:55:54 PM3/21/12
to a2-coc...@googlegroups.com
Wouldn't there be a mass exodus of apps if they instituted this that quickly?  I'd think that would be a problem, even for Apple?

Keith Bourne
Tel: 734.926.8208

Find me @:
Web: http://keithbourne.com
Twitter: @keithbourne



On Wed, Mar 21, 2012 at 9:45 PM, Chris Adamson <inval...@gmail.com> wrote:
nothing if not capricious.

Eric Shapiro

unread,
Mar 21, 2012, 10:41:35 PM3/21/12
to a2-coc...@googlegroups.com
On Mar 21, 2012, at 9:25 PM, Chris Adamson <inval...@gmail.com> wrote:

> Multiple reports from Twitter claim that Apple is rejecting apps and updates that use the UDID

The work-around being a hash of the MAC address + bundle id?

Can we always find the MAC, even if that network interface is offline? And can we differentiate between multiple MACs?

Seems like a PITA. I wonder if they're going to pull Adhoc Helper too.

Eric

------------------
Eric Shapiro
er...@relium.com

Chris Adamson

unread,
Mar 21, 2012, 10:56:50 PM3/21/12
to a2-coc...@googlegroups.com
The workaround being CFUUID. Way simpler than running your own algorithm or unique-number-server.

Original versions of the UUID used MAC + date, but this was rejected because using the MAC address could identify the source machine, which is potentially bad. Most implementations of UUID today, including Apple's CFUUID, use 128-bit random numbers. Yes, the random number is good enough: with that many bits, everyone on earth could have 600 million UUIDs and there'd only be a 50% chance of a duplication.

CFUUID docs here:
https://developer.apple.com/library/mac/#documentation/CoreFoundation/Reference/CFUUIDRef/Reference/reference.html

My CFUUID demo app from CocoaConf here:
http://dl.dropbox.com/u/12216224/conferences/cocoaconfmar12/UUIDStarfield.zip

--Chris

Paul Bruneau

unread,
Mar 21, 2012, 11:07:14 PM3/21/12
to a2-coc...@googlegroups.com
I wouldn't expect an exodus, if by exodus you mean a departure by choice of a lot of devs. But a lot of apps would be temporarily affected I think

Eric Shapiro

unread,
Mar 21, 2012, 11:07:27 PM3/21/12
to a2-coc...@googlegroups.com
On Mar 21, 2012, at 10:56 PM, Chris Adamson wrote:

> Original versions of the UUID used MAC + date, but this was rejected because using the MAC address could identify the source machine, which is potentially bad. Most implementations of UUID today, including Apple's CFUUID, use 128-bit random numbers. Yes, the random number is good enough: with that many bits, everyone on earth could have 600 million UUIDs and there'd only be a 50% chance of a duplication.

Yeah, I'm sure we'd never get a duplicate. Hah!

That's when jetliners start falling out of the sky.

Luckily, I'm not writing any jetliner control software. :-)

-Eric

-----------------------------------------------
Eric Shapiro sha...@relium.com
-----------------------------------------------

Voratima Orawannukul

unread,
Mar 22, 2012, 12:38:09 AM3/22/12
to a2-coc...@googlegroups.com, a2-coc...@googlegroups.com
Besides Google Analytics, what are the alternatives out there? WebTrends? Any suggestions?

Sent from my iPhone

Eric Shapiro

unread,
Mar 22, 2012, 12:44:49 AM3/22/12
to a2-coc...@googlegroups.com
On Mar 22, 2012, at 12:38 AM, Voratima Orawannukul wrote:

> Besides Google Analytics, what are the alternatives out there? WebTrends? Any suggestions?

We use Flurry.

Not sure I'd recommend them. They don't track the OS version, which is annoying. Nor do you get up-to-the-minute results, which is annoying when you launch a new app.

We also use Comscore, but I don't think I have access to those reports so can't comment on how good they are.

We've had issues with both SDKs, although I think everything has been fixed now.

I was tempted to roll my own, but the advertisers wanted standard Flurry and Comscore accounts.

Reply all
Reply to author
Forward
0 new messages