Paraesthesia

Windows 7 Supported Audio/Video Formats

Travis Illig — Wed, 04 Nov 2009 19:59:51 GMT

In researching potential video formats for switching my DVD library from VIDEO_TS to something else (due to file access time issues) I found I had real trouble locating the list of audio and video formats that Windows 7 supports out of the box. The provided "supported formats" page really isn't technical enough.

While I haven't determined which video format to move my library to, I did get a response to my question about this on the forums. Here's the info from there, for your convenience:

Format Name	File Extensions	Container	Video Decoders	Audio Decoders
MPEG-4	.mp4 (A, V, A+V) .m4a (A) .mov	ISO MPEG-4, AVI	H.264, MPEG-4 Advanced Simple Profile (ASP) and Simple Profile (SP)	AAC, MP3
3GPP/3GPP2	.3gp, .3g2 (A, V, A+V)	3GP	H.264, MPEG-4 Simple Profile (SP)	AAC
AAC	.aac (A)	ADTS		AAC
ASP in AVI (Compatible with DivX 4-6 video, Xvid and 3ivx)	.avi (V, A+V)	AVI	MPEG-4 Advanced Simple Profile (ASP)	MP3, MS ADPCM
AVCHD	.m2t, m2ts, .mts (A, V, A+V)	MPEG-2 Transport Stream (TS)	H.264	Dolby Digital, LPCM
HDV	.m2t, m2ts, .mts (A, V, A+V)	MPEG-2 Transport Stream (TS)	MPEG-2	MPEG-1 L2 (Layer 2)

Or, split/arranged a different way:

Category	Supported Types
Video Decoders	H.264/AVC Baseline, Main and High Profiles (new in Windows 7) MPEG-4 Advanced Simple Profile (new in Windows 7) besides Simple Profile Subset of H.263 that overlaps with MPEG-4 SP Motion JPEG (new in Windows 7) WMV variants including VC1 and WMV Screen MPEG-1 and -2 DV
Audio Decoders	AAC, HE-AAC, HE-AAC v2 (new in Windows 7) Dolby Digital Plus (new in Windows 7) besides Dolby Digital/AC-3 WMA family of music decoders and speech decoders MP3 MPEG-1 and -2
Image Decoders	HD Photo (JPEG-XR) JPEG PNG BMP TIFF GIF ICO (Icon Format)
Containers	MP4 and its close cousins MOV, 3GP, M4A (new in Windows 7) Media Center's WTV (new in Win7 and Media Center TV Pack) and DVR-MS MPEG-2 Transport Streams and its cousin AVCHD MPEG-1 and -2 Program Streams AVI, WAV ASF DVD ADTS

Debugging Visual Studio Add-Ins and XmlSerialization Problems

Travis Illig — Wed, 04 Nov 2009 00:53:28 GMT

The quick version:

If you're writing a Visual Studio add-in, a DXCore add-in, or anything else where the debug environment is Visual Studio itself, be careful not to load the add-in project in the Visual Studio instance you're debugging. You'll get some weird errors.

This is a total edge case but it had me baffled for a while.

I'm working on a DXCore plugin that talks to a web service. It takes some selected code and sends it to a web service to do something with. Since you need code to work with, the easiest way to debug it was to:

Hit F5 to start another Visual Studio environment with the add-in installed.
Go to File -> Recent Projects and grab the first one in the list - the add-in project.
Test out the plugin.

At one point I made a call to the web service using a complex type (not a string or integer, but a data transfer object) like this:

ComplexObject response = serviceProxy.DoAnOperation(param1, param2);

...where param1 is an input object of type ComplexObject. Doing that I got a weird XmlSerializationException:

[A]CR_PluginName.WebServiceNamespace.ComplexObject cannot be cast to [B]CR_PluginName.WebServiceNamespace.ComplexObject. Type A originates from 'CR_PluginName, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null' in the context 'LoadFrom' at location 'C:\Documents and Settings\tillig\My Documents\DevExpress\IDE Tools\Community\PlugIns\CR_PluginName.dll'. Type B originates from 'CR_PluginName, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null' in the context 'LoadNeither' at location 'C:\Documents and Settings\tillig\Local Settings\Application Data\Microsoft\VisualStudio\9.0\ProjectAssemblies\wi3h64z601\CR_PluginName.dll'.

Huh?

The key was in the second location: "...\ProjectAssemblies\..." The object I was sending to the web service was defined in the plugin assembly... but it was trying to deserialize the response using the temporary assembly that the debug instance of Visual Studio had compiled in the background.

The solution ends up being "don't do that." Opening up a different solution, I no longer had issues debugging my web service calls and the plugin worked fine.

Hit F5 to start another Visual Studio environment with the add-in installed.
Go to File -> Open Project... and find some other project that is not the add-in project.
Test out the plugin.

How to Upgrade PerfectDisk for Windows Home Server

Travis Illig — Tue, 03 Nov 2009 18:18:09 GMT

UPDATE 11/4/2009: I got some feedback on this process from Raxco (PerfectDisk) Support. They say:

The update process is as follows:

While connected to your WHS and notified than an update to PD is available, choose Install

When notified that the Console is open and prompted to Retry, Ignore or Cancel, choose Ignore

When update has finished and prompted to reboot your WHS, reboot

Upon reboot of your WHS, PD should be updated.

Note that if you run PD and click on Product Resources, PD shows that it has been updated to new build. If you look in the add-ins, it still says that old PD build is installed. This is due to the way that the add-in manager is looking to see what is installed/available - it is not looking in the "normal" locations to determine what is actually installed - it is using a different set of registry keys. Also note that if you have HP WHS, after updating to a new build of PD, the Network Critical tab displaying PD update as available even though it is already installed. This is something specific to HP WHS and we are currently waiting to hear from HP on what exactly they are checking/doing.

I'm not so sure I like that there are "expected problems" remaining after this upgrade process. Doing it my way, you get the upgrade, the version reads correctly, and there's no remaining WHS "Network Critical" warning. For reference, the original entry I wrote and the process I followed for upgrade is here...

I've got PerfectDisk for Windows Home Server installed and I generally like it but the installation procedure is a little weird.

A fresh install goes like this:

Download the full installation package for PerfectDisk to your client computer.
Install PerfectDisk on your client computer.
As part of the installation, a folder will be created where the PerfectDisk add-in for Windows Home Server will be placed. The location is typically something like: C:\Program Files\Raxco\PerfectDisk10Install\PerfectDisk10_Home_Server\PerfectDisk_x86.msi
Drop the PerfectDisk_x86.msi file into the \\server\software\add-ins folder on your Windows Home Server.
Open the Windows Home Server console, go to the "Settings" tab, and under "Add-Ins" select to install the PerfectDisk add-in.

The last few steps - dropping the installer in the "add-ins" folder, opening the WHS console, and installing add-in - is pretty standard WHS stuff. The weird bit is that you can't just get the WHS add-in as a direct download - you have to install on a client computer.

Recently I found there was an upgrade to PerfectDisk and the upgrade process isn't documented. This is what the process is from what I can tell, after some panicking and manually uninstalling things.

PerfectDisk has a feature to "check for updates" from within the Windows Home Server console. It's OK to use the "check for updates" feature to see if there's an update, but do not install the update from there.

First, trying that will fail because the Windows Home Server console is open (Catch-22, eh?). Second, if you try to do the update from Windows Home Server via Terminal Services, it doesn't quite do what you expect. I think the engine gets updated, but the console add-in doesn't or something. Stuff gets messed up. Just don't do it.

Instead, it appears that this is the proper sequence of events for an upgrade:

Go to the PerfectDisk site and re-download the full [updated] installer for the client computer.
Install the new version on your client computer. This will update the installer files in the C:\Program Files\Raxco\PerfectDisk10Install\PerfectDisk10_Home_Server folder as well.
Log into the WHS console and uninstall the PerfectDisk add-in. You'll need to restart the WHS console and possibly reboot the WHS after this.
Copy the new add-in from the C:\Program Files\Raxco\PerfectDisk10Install\PerfectDisk10_Home_Server folder into the \\server\software\add-ins folder on your server. You may want to keep a backup copy of the old version of the add-in somewhere before you overwrite it with the new version. Don't put two copies of the add-in in the server's "add-ins" folder, though.
Log into the WHS console and install the new version of the PerfectDisk add-in. It will tell you to restart the WHS console. Do that.
On restarting the WHS console, go to the PerfectDisk add-in and re-enter your product key.

You'll notice that this is pretty close to what a fresh install looks like. You'd be correct. The in-place upgrade for PerfectDisk doesn't work for WHS.

This is all, unfortunately, documented nowhere on the PerfectDisk site. I have submitted a ticket to their support people about this. In the meantime, I figured out a bunch of things you shouldn't do to upgrade the add-in.

Things you should NOT do:

Do not try to install the PerfectDisk update on your WHS through the "Check for Updates" feature. This ends you up with a mismatched add-in version and engine version and somehow corrupts the way the add-in is registered with the WHS console.
- Don't do it from the WHS console.
- Don't do it from a Terminal Service session to the WHS.
Do not update your client computer's PerfectDisk install through the "Check for Updates" feature. That will update your client computer just fine, but it won't update the installers so you won't get the updated WHS add-in.
Do not uninstall PerfectDisk on your WHS through Add/Remove Programs. I didn't do this, but it will end up corrupting the way the add-in is registered with the WHS console. Use the WHS console to remove PerfectDisk.

243 Trick-or-Treaters

Travis Illig — Tue, 03 Nov 2009 16:19:36 GMT

We had five more trick-or-treaters this year than we had last year and the most popular time to visit was between 7:30p and 8:00p, which is a half hour after last year. From the graph, it appears that kids are coming out later and later - back in 2006 and 2007, 6:30p to 7:00p was the most popular time to visit.

Here's the graph:

And the cumulative data from this year and the other years we've tracked:

		Time Blocks
		6:00p - 6:30p	6:30p - 7:00p	7:00p - 7:30p	7:30p - 8:00p	8:00p - 8:30p
Years	2006	52	59	35	16	0
	2007	5	49	39	25	21
	2008	14	71	82	45	25
	2009	17	51	72	82	21

Halloween was on a Saturday this year so we prepared for this number of kids by getting two giant bags of candy at Costco. We did think there'd be more, so we have about half a bag of candy left, but at least we didn't run out like we did last year.

We put the Halloween projector out again this year and that was nice but I think I want to do more decorating for next year. I think Halloween's on Sunday next year, so I'm not sure if we'll have quite as big of a turnout, but it's fun to decorate and such.

Around the 6:30p time we had a couple of larger trick-or-treaters show up in something similar to Death Eater masks. They were also wearing some pretty heavy long black coats. They didn't say anything (which isn't out of the ordinary) and just held their bags out. Fine, some high school kids doing their last hurrah. When the next group of kids showed up, though, these larger ones were still there. OK, fine, we'll give you a second piece of candy and send you on your way. But they were there for a third time when kids showed up so I told them, no, they'd already been here twice and it was time for them to move on. As I shut the door, one of them stuck a foot in and got up in my face so I started getting ready to get nasty. "You'll be moving that foot now." Then the doorbell rang and they were there again. I'd had it... and then they said "Trick or Treat!" out loud.

It was my freaking parents.

Man, they got us so good. We had no idea. It was my mom who'd got up in my face because she thought we could see her eyes through the mask. You couldn't - it was total blackout from outside the mask. Wow, did they get us. Not sure we'll be able to get 'em back for that, but we owe them a big one now.

Manually Uninstalling a Windows Home Server Add-In

Travis Illig — Tue, 03 Nov 2009 01:01:03 GMT

I just had a bit of a scare with a misbehaving Windows Home Server add-in where the upgrade process went frightfully wrong. As such, I ended up with:

The .msi for the add-in in the \\server\Software\Add-Ins folder.
The list of add-ins saying the add-in was installed.
No add-in actually installed.

Thank goodness there's a great article over on HomeServerLand that tells you how to manually uninstall an add-in. I followed that process and the crisis was averted.

Charting Hanselminutes

Travis Illig — Fri, 30 Oct 2009 21:37:11 GMT

Before I even get into this, let me preface it by saying Scott's a friend of mine and he's a great guy. I told him I was posting this before I did it. It's just some interesting data that I got in an interesting way and thought folks would be interested. It's also intended to totally crush Scott's spirit. (I kid! I kid!)

So.

I was just writing about how I was seeing more and more video blogs and was thinking about the earlier days of Hanselminutes when it seemed like the show was shorter. I wanted to make sure it wasn't just my mind playing tricks on me so I did some data gathering. This is actually about the process I went through.

The idea: create a graph of the Hanselminutes podcast duration over time so a trendline can be established.

At first I thought it would be pretty straightfoward - I could grab the RSS feed and just parse out the duration info. Turns out they don't actually list how long each show runs, so I had to change my tack and analyze the MP3 durations directly.

Step 1: Getting the MP3s.

I'm not a Powershell guru but this sounded pretty Powershell-ish to me. The thing is, I already had some tools that would do some of the job for me, so I didn't write the whole thing in Powershell. It went like this:

Grab the RSS feed for the show by just right-click and save-as from the site.
Get the URLs for the MP3s. I used a command-line XPath query tool for that, looking at /rss/channel/item/enclosure/@url. That gave me a nice list of the URLs to the show.
Get the MP3s. This is where I did a little brute force Powershell scripting. I suppose I could have saved the list of URLs to a text file and then wrote a script that read in the lines from the text file, but I didn't. I did a regex search-and-replace to create a script that looks like this:

$client = new-object system.net.webclient;
$client.DownloadFile("http://perseus.franklins.net/hanselminutes_0185.mp3", "hanselminutes_0185.mp3");
$client.DownloadFile("http://perseus.franklins.net/hanselminutes_0184.mp3", "hanselminutes_0184.mp3");
$client.DownloadFile("http://perseus.franklins.net/hanselminutes_0183.mp3", "hanselminutes_0183.mp3");
$client.DownloadFile("http://perseus.franklins.net/hanselminutes_0182.mp3", "hanselminutes_0182.mp3");
...

Like I said, pretty brute force... but I'm not running this a bunch of times, I'm just doing it once.

Step 2: Getting the Duration from the MP3s.

This was harder than I thought. What you actually have to do for this is get the MP3 tag info and get the duration from that.

I used the open source TagLib# and wrote a tiny console app using SnippetCompiler that looked like this:

DirectoryInfo dir = new DirectoryInfo(@"C:\Documents and Settings\tillig\Desktop\Hanselminutes");
FileInfo[] files = dir.GetFiles("*.mp3");
foreach(FileInfo file in files)
{
  TagLib.File tag = TagLib.File.Create(file.FullName);
  Console.WriteLine("{0}\t{1}", file.Name, tag.Properties.Duration);
}

Again, could I have done that with Powershell? Sure, but I'm not too strong in Powershell and I haven't had a chance to get too far beyond pretty basic stuff. And, again, I'm running it once.

So that gets me a tab-delimited text file with the name of the MP3 file and the duration.

Step 3: The Graph.

This was a simple import into Excel and add a graph. I won't go through that.

The Result:

I was right - there is an upward trend in the Hanselminutes duration.

So... interesting.

UPDATE: If you want the data for your own enjoyment, here you go.

The Problem with Video Blogs

Travis Illig — Fri, 30 Oct 2009 18:06:16 GMT

A few blogs I read have started experimenting with video blogging and it's made me realize that I'm not a video blog... watcher(?).

If content is written, I can read it at my leisure. I can search through it, I can get it on my Blackberry during a boring meeting, I can do pretty much whatever. I can read a few paragraphs, switch to something else, and come back. Worst case scenario, I forget where I was and I can very quickly skim the article again to jog my memory.

Audio content is OK, but not great. It demands a bit more attention.

I'm not one of those folks who can write a term paper and watch a movie and talk to someone all at the same time. The result of me trying to multitask my I/O like that is nothing gets my full attention. I won't know what's going on in the movie, my paper will end up taking a long time to write and won't make sense, and the conversation will dwindle.

I bring that up because with audio content, I can't listen while I'm working. I'll lose my train of thought. (I can listen to music, but generally stick to stuff that doesn't have words or stuff I've heard so many times I don't focus on it.) I can't listen while I'm home, either, because generally the audio blogs/podcasts I'm into aren't things my wife's into.

That leaves my 15-minute-each-way daily commute. Given that, it takes me two or three days to get through an hour-long show like This American Life. Five days per week means I get about one-and-a-half podcasts in. I have to really pick and choose. In many cases, I end up doing a lot of deleting without listening because I can never catch up. (I'm looking at you, 30-minutes-three-times-a-week-Planet-Money. And has Hanselminutes been getting longer or is it just me?)

Which brings me to video blogs/podcasts/whatever. This is the worst of all worlds.

I can't do anything with them while I'm working because it's not even just audio content, it's video, too.
I can't do anything with them on my commute because it's video. Plus, most times the video is on a site like YouTube where you can't even download it and listen to the audio.
I can't do anything with them at home because, frankly, if I'm going to sit down and watch something, there are plenty of more entertaining things I can watch to help me unwind than technical videos.

It's the same problem I have with the phone. Instantly single-threaded. I might be able to do something that doesn't require much brainpower at all, but basically, phone + me = useless.

COMPLUS_Version and the .NET Framework Runtime

Travis Illig — Thu, 29 Oct 2009 15:39:43 GMT

I just spent a couple of days debugging a weird problem. We have a fairly large product that has several Visual Studio solutions in it, all of which target .NET 3.5. No, that's not the problem. The problem was that we were able to build each solution separately in the correct dependency order just fine, but when the whole thing ran together in an automated fashion, the build would fail.

The failure message indicated that an extension method was not being recognized. Something like:

'Foo' does not contain a definition for 'Bar'

Again, it would build on its own, but not in the larger environment. What gives?

I figured the problem had to be the targeted .NET environment - that the project was targeting .NET 2.0 when run in the larger build but .NET 3.5 when run alone. And I was right, but not how I thought.

As it turns out, a custom build task run in an earlier build was setting an environment variable called COMPLUS_Version to v2.0.50727, which forced everything after that to run in .NET 2.0.

I had no idea such an environment variable existed. Doing a quick Google search on it, the only documentation on it has to do with build and test environments forcing things to run in different .NET versions, like if you're building something for .NET 1.1 and want to see how it runs in .NET 2.0. I searched MSDN and other sites, but I can't actually find any "official" documentation on this. It's just one of those things you figure out.

Valid settings for COMPLUS_Version seem to be the same as the names of the folders you see when you go to the %WINDIR%\Microsoft.NET\Framework directory, like:

v1.1.4322
v2.0.50727
v3.5

...and so on.

Setting the value will force future processes in that space to use the specified .NET runtime, like:

set COMPLUS_Version = v3.5

That would force everything to run in .NET 3.5.

And we tried that - doing a set to .NET 3.5 to force everything to that runtime, but we then ran into another issue: We were using the vsdbcmd.exe program to do some database work during a build (that's another story) and if you force it to run in .NET 3.5 you get the error:

To run this application, you must first install one of the following versions of the .NET Framework: v3.5 Contact your application publisher for instructions about obtaining the appropriate version of the .NET Framework

That made no sense to me since I obviously have .NET 3.5 installed.

The answer was to get rid of COMPLUS_Version entirely. After the custom build task ran, set the variable, and completed its work, we used the MSBuildCommunityTasks "script" task to unset the environment variable:

<PropertyGroup>
  <SetCode>
<![CDATA[
  public static void ScriptMain() {
    System.Environment.SetEnvironmentVariable("COMPLUS_Version", null);
  }
]]>
  </SetCode>
</PropertyGroup>
<Script Language="C#" Code="$(SetCode)" Imports="System" />

Doing that removes the variable from the process space and later executables can allow the CLR to choose which environment to target automatically.