I've had a couple of folks ask me about the form validation I mentioned in my last post, so here's the code.
It's fairly basic IMO. This is public domain code; use as you will and at your own risk. You'll be able to make a form that looks like this:
I've spent some of my personal time over this past week looking into Node.js. I'll be up-front in stating that my efforts have not been particularly broad, but I have gone reasonably deep into the parts that I looked at.
I've been building evented systems for more than a decade, mostly in the "C" language, so Node is particularly interesting to me; it blends an evented I/O model with the Javascript language in such a way that it doesn't feel like a continual struggle.
To get a feel for Node, I decided to look at what it would take to get some kind of mtrack implementation running on Node; don't get too excited, I didn't finish anything worthy of public scrutiny.
For the purposes of the experiment I selected Riak for the data storage solution; it has a compelling mixture of document storage, full-text searching, secondary indices and map-reduce facilities.
After reading this scary blog entry about domain hijacking I've been a bit concerned about brute forcing of credentials and have been turning on the two-factor authentication facilities that folks like Google provide for my gmail and personal domains.
I've just found out about Duo Security, a service that allows you to add two-factor authentication to your SSH server, Juniper VPN and even Wordpress blogs. Their service is free for up to 10 users and they start charging when you pass that threshold.
Read on to find out how to set it up.
I'm pleased to announce that my proposal for the OmniTI Surge 2011 Conference has been accepted!
TL;DR: read the session description at the Surge site, otherwise, read on for some background and additional commentary.
Brendan Gregg raised this question on Twitter: Cruel interview question: write a program on Solaris that sleeps for < 1 ms.
The first choice that springs to mind is nanosleep(3RT), but you'll be frustrated because your requested sleep interval is implemented in terms of the standard clock frequency, making you sleep for longer.
So how do you sleep (not busy wait!) for time periods shorter than the regular clock?
Adam Leventhal mentioned on twitter that he'd found my umem slab allocator port on macports. I commented that he might miss some of the umem debugging features that are available on its native Solaris platform in mdb, and we dreamed briefly of a portable version of mdb.
This weekend I took a peek at the mdb sources (via the Illumos Gate) to see how hard this might be. As it happens, this isn't the first time I've looked at porting mdb, so I was bit better prepared to go snooping around in there.