WWDC wind-down

As everyone is getting on their respective planes and flying back to their respective homelands, it's time to look back on what happened and what the conference means.

The event itself was great fun, as ever. Meeting loads of new people (a big thank-you to the #paddyinvasion for my dishonourary membership) as well as plenty of old friends is always enjoyable - especially when everyone's so excited about what they're working on, what they've discovered and what they're up to the next day. It's an infectious enthusiasm.

Interestingly the sessions and labs content has more of a dual impact. On the one hand it's great to see how new things work, how I could use them, and to realise that I get what they do. The best feeling is taking some new information and being able to make use of it or see how it can be used. That's another reason why talking to everyone else is great - they all have their own perspectives on what they've seen and we can share those views, learning things from each other that we didn't get from the sessions. If you were wondering what the animated discussions and gesticulations were in the 4th Street Starbucks at 7am every morning, now you know.

On the other hand, it makes me realise that OS X is such a huge platform that there are parts I understand very well, and parts that I don't really know at all. My own code spreads a wide path over a timeline between January 1, 1970 and September 2009 (not a typo). For instance, it wasn't until about 2003 that I knew enough NetInfo to be able to write a program to use it (you may wonder why I didn't just use DirectoryServices - well even in 2003 the program was for NeXTSTEP 3 which didn't supply that API). I still have a level of knowledge of Mach APIs far below "grok", and have never known even the smallest thing about HIToolbox.

There are various options for dealing with that. The most time-intensive is to take time to study - I've got a huge collection of papers on the Mach design and implementation, and occasionally find time to pop one off the stack. The least is to ignore the problem - as I have done with HIToolbox, because it offers nothing I can't do with Cocoa. In-between are other strategies such as vicariously channeling the knowledge of Amit Singh or Mark Dalrymple and Aaron Hillegass. I expect that fully understanding Mac OS X is beyond the mental scope of any individual - but it's certainly fun to try :-).

Rootier than root

There's a common misconception, the book I'm reading now suffers from it, that single-user mode on a unix such as mac os x gives you root access. Actually, it grants you higher access than root. For example, set the immutable flag on a file (schg I think, but my iPhone doesn't have man). Root can't remove the flag, but the single user can.

What's new in 2009

Of course, it's a bit early for a retrospective of 2008, besides which I've already written 73 entries this year, my most prolific year to date on iamleeg. And that doesn't count numerous tweets, stack overflow contributions and of course the occasional piece of source code here or there for some security company. As the noise of fireworks and exploding media players sounds across the world, it's time to pre-emptively ditch 2008 and see what we can expect from 2009. Specifically, what you can expect from me.

It looks to me like the most popular pieces on this blog are the opinions and how-tos regarding Cocoa development, particularly my thoughts on properties and Cocoa memory management round-up. Don't worry, there's definitely more of this coming. As well as preparing for this Mac Developer Network conference talk I've been discussing recently, I've got another exciting - and unfortunately secret - project on the go now which should see plenty of collateral blog posting in the first half of the next year, all about Cocoa development. There'll also be a bit more of an iPhone mix-in; obviously for much of last year the SDK either didn't exist or was under non-disclosure, but now I've got more reasons to be using Cocoa Touch it will also be mentioned on here. I shall also be delving a bit deeper into Darwin and xnu than I have in previous times.

One example of Cocoa-related information is meetup announcements; I'm still involved in the local CocoaHeads chapter and I'll endeavour to post an advance warning for each meeting here. I know many of my readers are in the States but a few of you are local so please do come along! In fact, if you're not local (or "bissen't from rond theez partz", as we say here) then consider going to your nearest CocoaHeads or starting a new one. It's a great way to find out who's working on Mac or iPhone development in your area, share tips and stories and build up that professional contacts network.

Previously I've been concerned that readers here at iamleeg don't seem interesting in commenting on my posts, but these days I'm no longer worried. I can tell how many people are reading, and of those how many are regulars, and I have to say that the blog is doing pretty damn well. Of course, if you do feel inclined to join in the discussion (particularly if I've got something wrong, or missed an important point from a post) then you should feel perfectly at liberty to leave a comment.

Finally, have a happy new year!

Whither the codesign interface?

One of the higher-signal-level Apple mailing lists with a manageable amount of traffic is apple-cdsa, the place for discussing the world's most popular Common Data Security Architecture deployment. There's currently an interesting thread about code signatures, which asks the important question: how do I make use of code signatures?

Well, actually, it's not so much about how I can use code signatures, but how the subset of Clapham Omnibus riders who own Macs (a very small subset, as the combination of overheating batteries in the old G4 PowerBooks and combustible bendy-busses means they don't stay around very long) can use code signatures. Unfortunately, the answer seems to currently be "not by much", with little impression of that changing. The code signing process and capability is actually pretty damned cool, and a nice security feature which I'll be talking about at MacDev 09. It's used to good effect in the iPhone, where it and FairPlay DRM are part of that platform's locked-down execution environment.

The only problem is, there's not much a user can do with it. It's pretty hard to find out who signed a particular app, in fact the only thing you can easily do is discover that the same entity signed two versions of the same app. And that's by lack of interface, not by any form of dialogue or confirmation. That means that when faced with the "Foobar app has changed. Are you sure you still want to allow it to [whatever]" prompt, many users will be unaware of the implications of the question. Those who are and (sensibly) want to find out why the change has occurred will quickly become frustrated. Therefore everyone's going to click "allow", which rather reduces the utility of the feature :-(.

Is that a problem yet? Well, I believe it is, even though there are few components yet using the code signature information in the operating system. And it's precisely that allow-happy training which I think is the issue. By the time the user interfaces and access control capabilities of the OS have developed to the point where code signing is a more useful feature (and believe me, I think it's quite a useful one right now), users will be in the habit of clicking 'allow'. You are coming to a sad realisation; allow or deny?

Local KDC on Leopard

via Nigel Kersten, a great description of the operation of Leopard's built-in local KDC. I think the most exciting thing about the local KDC is the Bonjour support; could we see simple cross-system trust in the near future? Could there be someone in the world who can actually make Kerberos simple?

Broke track mounting

For some reason, CDs occasionally don't automount for me on my iMac. Luckily that's easy to work around:
kalevala:~ leeg$ diskutil list
[...]
/dev/disk3
#: TYPE NAME SIZE IDENTIFIER
0: CD_partition_scheme Audio CD *620.3 Mi disk3
[...]
kalevala:~ leeg$ diskutil mountDisk disk3
Volume(s) mounted successfully
Job is, as they say, a good 'un.

Mach-OFS: aforementioned polish and functionality

It's getting there, now has the ability to display load commands (though it only reports useful information for LC_SEGMENT and LC_SEGMENT_64 commands):

Again the screenshot depicts the OmniDazzle binary for no reason other than it's a nontrivial file. The directions in which to take the filesystem are now numerous: I can add info about the remaining load commands (v. useful), the raw data for each segment (somewhat useful), and the sections in each segment (v. useful). Whether the filesystem will eventually get to the level of symbol resolution, I'm not sure :-).

Mach-O FS (no really, MacFUSE does rule)

It needs some polishing and more functionality before I'd call it useful, then I have to find out whether I'm allowed to do anything with the source code ;-). But this is at least quite a cool hack; exploring a Mach-O file (thin or fat - in this case, I used the OmniDazzle executable which is a fat file) as if it's a file system. FUSE of course makes it easy, so thanks to Amit Singh for the port!

MacFUSE rules

One reason that microkernels win over everything else (piss off, Linus) is that stability is better, because less stuff is running in the dangerous and all-powerful kernel environment. MacFUSE, like FUSE implementations on other UNIX-like operating systems, takes the microkernel approach to filesystems, hooking requests for information out of the kernel and passing them to user-space processes to handle. Here's the worst that can happen when screwing up a FUSE filesystem:

Now that might sound not only like a recipe for lower-quality code, but also like I'm extolling the capability to create lower-quality code. Well no it isn't, and yes I am. The advantage is that now the develop-debug-fix cycle for filesystems is just as short as it is for other userland applications (and HURD translators and the like). This provides a lower barrier to entry (meaning that it's more likely that interesting and innovative filesystems can be created), but also a faster turnaround on bugfixes (no panic, restart, try to salvage panic log... no two-machine debugging with kdb...) so ultimately higher-quality filesystems.

Old news

So the Inquirer thinks they've got a hot potato on their hands, with this "security flaw" in OS X. I've been using this approach for years (like, since NeXTSTEP): boot into single-user and launch NetInfo manually, then passwd root. Or in newer Mac OS X, nicl means you don't have to launch NetInfo.

Of course, if you give physical access to the computer without a Firmware password, then the 'attacker' may as well just boot from external media and do whatever they want from there. But the solution, as well as setting the Firmware password, is to edit the /etc/ttys file, change the line:

console "/System/Library/CoreServices/loginwindow.app/Contents/MacOS/loginwindow" vt100 on secure onoption="/usr/libexec/getty std.9600"

to:

console "/System/Library/CoreServices/loginwindow.app/Contents/MacOS/loginwindow" vt100 on onoption="/usr/libexec/getty std.9600"

Now the root password is required in single-user mode (as the console is no longer considered a secure terminal).

Random collection of amazing stuff

The most cool thing that I noticed today ever is that Google Maps now allows you to add custom waypoints by dragging-and-dropping the route line onto a given road. This is great! I'm going to a charity biker raffle thing in Pensford next weekend, and Google's usual recommendation is that I stay on the M4 to Bristol, and drive through Bristol towards Shepton Mallet. This is, frankly, ludicrous. It's much more sensible to go through Bath and attack the A37 from the South, and now I can let Google know that.

Trusted JDS is über-cool. Not so much the actual functionality, which is somewhere between being pointy-haired enterprisey nonsense and NSA-derived "we require this feature, we can't tell you why, or what it is, or how it should work, but implement it because I'm authorised to shoot you and move in with your wife" fun. But implementing Mandatory Access Control in a GUI, and having it work, and make sense, is one hell of an achievement. Seven hells, in the case of Trusted Openlook, of which none are achievement. My favourite part of TJDS is that the access controls are checked by pasteboard actions, so trying to paste Top Secret text into an Unrestricted e-mail becomes a no-no.

There does exist Mac MAC (sorry, I've also written "do DO" this week) support, in the form of SEDarwin, but (as with SELinux) most of the time spent in designing policies for SEDarwin actually boils down to opening up enough permissions to stop from breaking stuff - and that stuff mainly breaks because the applications (or even the libraries on which those applications are based) don't expect to not be allowed to, for instance, talk to the pasteboard server. In fact, I'm going to save this post as a draft, then kill pbs and see what happens.

Hmmm... that isn't what I expected. I can actually still copy and paste text (probably uses a different pasteboard mechanism). pbs is a zombie, though. Killed an app by trying to copy an image out of it, too, and both of these symptoms would seem to fit with my assumption above; Cocoa just doesn't know what to do if the pasteboard server isn't available. If you started restricting access to it (and probably the DO name servers and distributed notification centres too) then you'd be in a right mess.

A bit of backup script

Good news - there's a handy tool in OS X called wait4path which can help when writing timed scripts to backup to removable media.

Bad news - it [at least in Tiger....] works slightly esoterically - if a path is already present, it will still wait for another mount kevent before exiting. It should therefore be used in a script like this:

#!/bin/sh

if [ ! -d /Volumes/Backups ]; then
echo "waiting for backup volume..."
/bin/wait4path /Volumes/Backups
fi

# do some backups

Note, however, that if you do this in a crontab job it could potentially wait for a very long time, so you should wrap all that with a /var/run style semaphore.