I recently ran into a problem on my fancy new machine while trying to commit a rather large number of files into an SVN repository. The error message stated that some of the files in the .svn control directory had become corrupted and unreadable.

After Googling around a bit I came across a post on the CollabNet issue log which identified this as an issue with the NTFS stack on Windows 7. This post included another link to a Microsoft technet discussion about the issue.

The long and the short of it is that this is an identified issue in the NTFS implementation in all editions of Windows 7 (both 32 and 64 bit versions). The indexing service is locking files which SVN is trying to move. This only appears to be a problem with large batch transactions. Smaller ones, for me at least, have been working just fine but YMMV.

In case you don’t want to read the whole discussion thread, here’s the response from the NTFS team developer who responded to the community reports:

This is a known regression in Windows 7 in the NTFS file system.  It occurs when doing a superceding rename over a file that has an atomic oplock on it (atomic oplocks are a new feature in Windows 7).  The indexer uses atomic oplocks which is why it helped when you disabled the indexer.  Explorer also uses atomic oplocks which is why you are still seeing the issue.  When this occurs STATUS_FILE_CORRUPT is incorrectly returned and the volume is marked “dirty” which is a signal to the system that chkdsk needs to be run.  No actual corruption has occured.

Neal Christiansen
NTFS Development Lead

The identified workaround for this issue is to stop the indexing service. If you don’t use search very often you can disable it. If you do, you can just stop the service and allow it to restart the next time you restart Windows.

The next trick, of course, is finding the indexing service.  In Windows 7 the service has been renamed “Windows Search”. It serves essentially the same functions as the old “Indexing Service”.

There have been some reports that this issue affects Windows Vista as well, but I don’t have a Vista machine to test with.

It’s not new information, but since I just figured this out recently, I thought I’d punch it up here.  When you add a new bookmark in FireFox there are a few fields that you get prompted to fill out:

These fields are only a portion of the information that Firefox can keep for a particular bookmark. Among the most useful of these is the keyword field.

For people like me who prefer to use the keyboard over any form of pointing device, this is fantastic.  I can open a new tab and type in the keyword of the bookmark I want, in this case ‘kdm’ and firefox will load the bookmark in question.

Even though most browsers have a built-in search bar, that’s another toolbar that eats up real-estate in your browser. If you’re screen-space-challenged or you have a disdain for toolbars.  Take your bookmarks to the next level with a parameter.

By adding in the string %s you can create search bookmarks for your favourite searching sites. As a helper, I’ve included a few search links below which already contain the %s. Just right-click the links below and use the ‘Copy Link Location’ or ‘Copy Shortcut’ options.

Google (ca, au, de, jp)

Bing

Yahoo

Wikipedia (de, fr, pl, it)

Stack Overflow

Memory Alpha

Over the past few days I’ve put some time into working on a library to create EPUB documents in .NET.  When I first did a search for this a few months ago I really didn’t find anything that suited my needs: a library that I could use to create EPUB documents on the fly, in code.

So I said to myself: “Self! You can write code, build the damn thing yourslef!”. So I did.

The initial version of the library has been published up on Google Code and is probably full of holes. If you’re interested, have a look and let me know what you think.  I’ll try to publish some more detailed specs for what the library does in the coming weeks.

For now, it can get got at: http://code.google.com/p/epubsharp/ — and yes, the documentation on that page is as sparse as it is here. 

The future of Short Order code is very much up in the air right now. With a baby on the way in the near future and what seems like no time to work on things for the podcast already There’s a good chance that there won’t be another episode of Short Order code for at least another few months.

I have been considering using the SOC blog as more of a blog and posting all of my programming related posts there and leaving this one as more of a general blog. I have some posts coming soon for programming related things and I’m trying to figure out what makes more sense: posting them here where they’ll get a few more eyeballs, or posting them on SOC where the site is all about programming — albeit a programming podcast that hasn’t seen an episode since late 2009.

In short, the future is muddy. Hopefully I’ll be able to shed some more light on things as the next month or two roll on.

I’ve got no doubt I’ll take some heat for going this route, but my solution to the build it or buy it dilemma was in fact to buy a new box. I decided to go for a machine from their business lineup, the Vostro 430 MT. This machine comes with a bunch of cores and an ample amount of RAM will allow this box to become the base for my home computing activities.

This machine’s primary duty will be to take over the day-to-day computing tasks from my Macbook freeing it up for mobile-only duties so that I’m not carrying around the primary copies of most of my data every time I leave the house.  In addition, I’ll be virtualizing the two desktop machines I have in the basement today, one of which has succumbed to old age in the past couple of months.  The VMs will provide some added flexibility as far as Operating Systems and some single-duty machines.

For now I’m evaluating a couple of virtualization options to use on the new box when it arrives, and I’m re-working the layout of the home network, splitting out VLANs for the servers to help isolate them a bit from the rest of the network. The main goal of some of this is to play around and learn the technologies, of course.

Some of the VM applications I’m considering include test machines for development, SSH tunnel/proxy for remote access, web server, Mozilla Weave server, secure machine for my office VPN and whatever else my dreams, or nightmares (or conversations with Knightwise) inspire.

The machine should arrive in the next week, unboxing and other additional posts will follow.

Google Docs today expanded the functionality of the service by offering 1GB of cloud storage for any type of file you want to upload. The only restrictions seem to be the 1GB cap on storage (you can buy more at $0.25 USD / GB) and the limit of 250MB per individual file.

Though it doesn’t quite match some of the other features we’ve become accustomed to seeing in cloud storage like an OS client with drag & drop functionality, it certainly provides a reliable place to store some of your larger files that won’t fit under the smaller 20MB cap that Google places on email. This is quite possibly the precursor to the much lauded and speculated about G-Drive cloud storage service.

via LifeHacker

A super fast way to create ISO images (or any kind of bit for bit copies) from the OS X (or linux, or unix) command line is to use the DD command. This isn’t without it’s risks since the DD command will duplicate byte for byte all data from one location to another, but once you’ve found the necessary information you can easily replicate data.

I needed to create a copy of my Windows XP installation CD to use for setting up new virtual machines. Since it’s quicker to install VMs from an ISO image rather than from the installation media I wanted an ISO of my CD. The built in OS X disk utility can make this work… in theory. My copy of disk utility (on Snow Leopard) didn’t want to recognize the CD. So I turned to my command line knowledge.

I should also note that before you can read the CD using DD, you will need to unmount it:

sudo umount /dev/disk1s0

You’ll be prompted for your administrator password, then you’re in business.

To create an image of a CD (or DVD or anything else…) use the following:

dd if=/dev/disk1s0 of=/Users/kdmurray/Desktop/WindowsXP.iso bs=64k

• if= input file, in this case our Windows XP CD
• of= output file, the actual ISO file you want created
• bs= block size, the size of chunks the tool will work with

I did a few speed tests at 16k, 32k, 64k, 128k, 256k and 1M and found that the 64k seemed to be the ideal block size on my machine. This will vary by CPU, and possibly by hard drive controller, so do a few contolled tests first. To do a test, start the copy process, then stop it after 20-30s (ctrl-c). Whichever one copied the most in 30s is the one you should use.

That’s all she wrote, folks. Happy command-lining.

The time has finally come for me to pick up a new system. This is usually a very happy moment in a geek’s life; the time when you get to spec out a new system with some of the latest tech. Reading product reviews, selecting components… it’s fun!

One of the true marks of geekdom is the ability to build your own machines from the ground up. Everything from ordering the parts, to gingerly placing in the CPU and glopping on just the right amount of thermal-conductive paste right through to powering up the system for the first time then installing and pimping out the OS of your choice. I’ve been there, its a ton of fun when you finally get everything working just right. It’s also a fair bit of work.

The more pragmatic geek can also buy a system, spec’d to order from companies like Dell or Lenovo. These systems still have pretty much everything you want, professionally assembled and sent to your door. These usually take a few weeks to ship so you need to be a bit more patient but there’s a lot less work involved.

So, what about price? Most of the time the price is pretty comparable for a complete system. I priced out a system with the following basic requirements from both Dell and a local computer retailer:

• Intel Core i7 860
• 8GB of Memory
• Windows 7 Pro

The two systems came out less than $80 apart with some minor differences in specs, hard drive size, optical drive, video card etc. For the most part it was the same system, and essentially the same price.

So what does it come down to? Is one option truly better than any other? The best advice I can give is that you need to get yourself a system you can trust. If that means you want professionals to put the system together for you and certify it, then that’s the route you need to take. If, however, you’re the Richard Stallman of computer builds and you feel that you need to free each of the components and build a computer that’s independent of “the man” and that’s where you derive the most value, then that should be your choice.

Image courtesy Frigante on Flickr.

For the longest time I simply lived with the fact that my system clock didn’t behave properly in my Windows partition when I had BootCamp setup on the macbook. It didn’t matter what I did, I couldn’t seem to find a way to keep the clock properly synchronized.  The problem stems from how Apple decided to implement the system clock which is different from the way that most PC manufacturers do.

Apple’s hardware always returns time in UTC, most PCs update the system clock to local time and won’t require a conversion of that time so of course the default Windows settings aren’t really set up for BootCamp (though you’d think the BootCamp installer could correct this…)

The solution, as it turns out, is a simple registry key change to tell Windows that the system’s clock is using UTC time.  One value in one key, that’s all it takes:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\TimeZoneInformation]
"RealTimeIsUniversal"=dword:00000001

BootcampUtcClockFix.reg

Another solution that I found over at superuser.com is to modify OS X to change the time at startup and at shutdown. To me, this seems like more of a hack than the “unsupported” registry key. Windows 7 has been working well under this configuration and has even successfully interpreted the daylight saving time change.

It’s been a long time since I had an active presence on GeoCities (about 13 years by my count) but its still a site that holds a place in my heart. I got  started with website design, as did thousands of other people, by having a site on GeoCities many years ago. And within a few months of joining what was in some ways one of the earliest social networks of the web, I’d entered the “GeoCities Community Leader” program to help other people learn the technology, and learn to do more with the tools the free service provided.

Today Yahoo is shutting down GeoCities officially. For good. It’s been 15 years since the service got its start in 1994, and 10 years since Yahoo bought the company for over $3 Billion in stock near the peak of the tech bubble in late 1999. Yahoo was fairly patient with the service leaving accounts active in one form or another until today.  It is in many ways the end of an era.

Some sites are paying tribute to the venerable property. I can’t tell if xkcd is paying tribute, or reminding those of us who are a bit sentimental just what the web looked like when GeoCities was the major force behind free web space on the internet.  Either way, have a look at the xkcd homepage today, or if you’ve missed the boat on that one, click the image in the right for a look at their 1990′s inspired design.

RIP GeoCities.