Continuous Integration Enters the Cloud

I came across this blog post in Google Reader and thought I’d share it.  The idea of being able to outsource the care and feeding of a continuous integration system to someone else is a very interesting one.  Having implemented and maintained such systems (which I’ve blogged about  in the past), I know it can be a lot of work (though using a product like TeamCity lightens the load considerably compared with CruiseControl.NET).  Stelligent isn’t the first company to come up the idea of CI in the cloud, but they may be the first using all free/open source tools to implement it.

I’ve read Paul Duvall’s book on continuous integration and highly recommend it to anyone who works with CI systems on a regular basis.  If anyone can make a service like this successful, Mr. Duvall can.

Set-ExecutionPolicy RemoteSigned

When you first get started with PowerShell, don’t forget to run ‘Set-ExecutionPolicy RemoteSigned’ from the PowerShell prompt. If you try to run a script without doing that first, expect to see a message like the following:

File <path to file> cannot be loaded because execution of scripts is disabled on this system.  Please see "get-help about_signing" for more details.
The default execution policy for PowerShell is "Restricted" (commands only, not scripts).  The other execution policy options (in decreasing order of strictness) are:
  • AllSigned
  • RemoteSigned
  • Unrestricted
When I first tripped over this, the resource that helped most was a TechNet article.  Later, I found a blog post that was more specific about the execution policies.

Can't launch Cassini outside Visual Studio? This may help ...

I’d been trying to launch the Cassini web server from a PowerShell script for quite awhile, but kept getting an error when I tried to display the configured website in a browser.  When I opened up a debugger, it revealed a FileNotFoundException with the following details:

"Could not load file or assembly 'WebDev.WebHost, Version=8.0.0.0, Culture=neutral, PublicKeyToken=...' or one of its dependencies..."
Since the WebDev.WebHost.dll was present in the correct .NET Framework directory, the FileNotFoundException was especially puzzling.  Fortunately, one of my colleagues figured out what the issue was.  WebDev.WebHost.dll wasn't in the GAC.  Once the file was added to the GAC, I was able to launch Cassini and display the website with no problems.

Can Google Find You?

Recruiters use Google.  Whether you’re actively seeking a new job or not, it’s important to use this fact to your advantage.  My friend Sandro gave me this advice years ago, when he told me to put my resume online and make it “googleable”.  For me, the result was contacts from recruiters and companies I might never have heard of otherwise.  In addition to putting your resume online, I would recommend blogging about your job–within reason.  Definitely do not write about company secrets or co-workers.  Putting such things in your blog doesn’t help you.  Instead, write about what you do, problems you’ve solved, even your process of problem-solving.  At the very least, when you encounter similar challenges in the future, you’ll have a reference for how you solved them in the past.  Your blog post about how you fixed a particular issue might be helpful to someone else as well.

There are many options available for putting a resume and/or blog online.  Sandro hosts his, mine, and a few others on a server at his house.  But for those of you who don’t have a buddy to host theirs, here are a couple of readily-accessible free options:

There's a ton of advice out there on what makes a great resume, so I won't try to repeat it all here.  You can simply put a version of your 1 or 2-page Microsoft Word resume on the web, or you can put your entire career up there.  Having your own blog or website means you aren't subject to any restrictions on length that a site like Monster or CareerBuilder might impose.  Consider linking your resume to the websites of previous employers, technologies you've worked with, schools you've attended, and work you've done that showcases your skills (especially if it's web-related).  I don't know if that makes it easier for Google to find you, but it does give recruiters easy access to details about you they might have to dig for otherwise.  Doing what you can to make this process easier for them certainly can't hurt.

Transforming Healthcare through Information Technology

Back on November 20, I attended a seminar at the Reagan Building on how healthcare in the U.S. could be improved through information technology.  As an alumnus of the business school, and someone who’d worked in healthcare IT before, I wanted to learn about a part of the healthcare debate that I hadn’t seen much coverage lately.  Dr. Ritu Agarwal gave the talk and answered questions during and after her presentation.

The main problem with healthcare in the U.S. could probably be summed up this way:

Despite spending more on healthcare than any other country in the world, our clinical outcomes are no better than in countries that spend far less.
Even more disturbing, of the 30 countries in the OECD, the U.S. has the highest infant mortality rate.

In the past 10 years, premiums for employer-based health insurance have risen 120%.  Over the same period, inflation grew 44%, while salaries grew only 29%.  So healthcare costs are increasing far faster than inflation (and our ability to pay for it with our salaries).

As far as healthcare IT goes, Dr. Agarwal gave the following reasons for the slow pace of adoption by healthcare providers:

  • inertia
  • it's a public good--patients get the benefits--not the healthcare providers
  • lack of common standards
Adding to the inertia point is the fact that healthcare in the U.S. has many stakeholders--patients, medical professionals, hospitals, pharmaceutical companies, insurance companies, and more.

Dr. Agarwal pointed to a number of countries with successful implementations of healthcare IT.  They included Canada, Australia, and the United Kingdom.  Australia in particular was singled out as being 5-10 years ahead of the U.S.

One thing I didn’t expect was that the Veterans Administration and the Department of Defense would be held up as native models of successful healthcare IT implementations.  One key factor noted by one of the other seminar participants was that the VA and DOD systems were closed.  Providers, specialists, hospitals, etc were all part of the government.  This enables them to enforce standards, in patient records and other areas.  Another point I considered later (which didn’t come up in the Q & A) was that the government model is non-profit as well.

Dr. Agarwal’s proposed solution to improving the current state of IT use in healthcare (as I recall it) was an regional exchange model.  Healthcare providers in a particular region of the U.S. would choose a standard for electronic health records (EHR) and other protocols.  Connections between these regional exchanges would ultimately form a national health information exchange.  Building on existing protocols and technologies (instead of attempting to build a national exchange from scratch) would be the most practical choice.

For more information, check out the slides from the presentation.

Unit testing strong-named assemblies in .NET

It’s been a couple of years since I first learned about the InternalsVisibleTo attribute.  It took until this afternoon to discover a problem with it.  This issue only occurs when you attempt to unit test internal classes of signed assemblies with an unsigned test assembly.  If you attempt to compile a Visual Studio solution in this case, the compiler will return the following complaint (among others):

Strong-name signed assemblies must specify a public key intheir InternalsVisibleTo declarations.
Thankfully, this blog post gives a great walk-through of how to get this working.  The instructions in brief:
  1. Sign your test assembly.
  2. Extract the public key.
  3. Update your InternalsVisibleTo argument to include the public key.

A visit to Iowa City

Last weekend, I visited my cousin Kevin at the University of Iowa to sit on his Ph. D defense.  For the past five years, he’s been working in pharmaceutical chemistry figuring out how to create vaccines that can be delivered directly to human genes.  I’m no chemist, so the bulk of his talk was way over my head, but it was very cool to see his command of the material and how well he presented.  When he came back from the private portion of his defense, we knew he’d succeeded.

After a celebratory lunch, Kevin took his brother Richard, sister Michelle, and me to a firing range to shoot.  By firing range, I don’t mean some shiny building with paper targets on motorized tracks.  We drove about an hour from Iowa City to a fenced-in area outdoors with some metal stands and a big pit.  You bring your own guns, ammo, and targets.  When other people are around, you have to signal them that you’re going to put targets out so they stop shooting.  We turned our fire on some empty steel solvent containers with four different weapons: a Ruger pistol (.22 LR ammunition), a Ruger rifle, a Springfield 1911 (.45 ammunition), and an M1 Garand (7.62mm rounds).  After spending a couple hours shooting, I will never look at Hollywood shoot-em-ups the same way again.  Movies seem totally fake compared with the noise and recoil of large-caliber weapons.  We had fun, and we turned out to be half-decent shots (for rookies).

StackOverflow Dev Days DC

In this case, DC = Falls Church, VA.  I went to the State Theatre to attend this conference.  Considering the cost ($99/person), the conference turned out to be a great value.  I wrote up a conference report to share with my bosses and co-workers and it’s included below.  It has footnotes because I typed it up in Word 2007 and pasted it in here with very little editing (because after all this writing, I’m feeling a bit lazy).

Summary The main reasons the creators of stackoverflow.com came up with this conference include the following:

  1. Bring together developers that are part of the Stack Overflow community[1]
  2. Teach developers something outside their immediate field[2]
  3. Accomplish 1 & 2 at low cost[3]
A fourth reason I would add is to pitch FogBugz 7.  It’s the primary product offering of Fogcreek Software, so it would have been odd for a conference it supports to not do at least a little advertising.  Spolsky also attempted to divide the venue by area for networking around certain topics, but I’m not sure how successful that was.

The conference succeeded in its main objectives.  At $99 per person, this conference was a bargain.  Given the diversity of topics and caliber of speakers, the price could have been higher and the venue would still have sold out.  Of the seven official topics presented (there was an eighth on Agile after the conference ended), only the ASP.NET MVC talk used technology that I had hands-on production experience with.  I was disappointed not to see a presentation on Android, but that was the only thing obviously missing from the day.

Keynote: Joel Spolsky If I were to boil down Joel Spolsky’s keynote to a single phrase, it would be this:

“Death to the dialog box!”

Spolsky’s talk argued persuasively that software often forces users to make decisions about things they don’t care about, or don’t know enough about to answer correctly.  Among his examples were modal dialog boxes for products like QuickBooks and Internet Explorer, and the Windows Vista Start button.  He talked about the other extreme (overly simple applications) as well, using the “build less” design philosophy of the 37signals team as an example.[4] Equating those kinds of applications with Notepad was a reach (and clearly played for laughs), but described the limitations of the alternative to complex applications pretty well.  The examples did a good job of setting up the choice between simplicity and power.

He cited an experiment in the selection of jam from The Paradox of Choice: Why More Is Less[5] to show the potential drawbacks of too many choices.  When the results of this experiment showed that a display with fewer choices resulted in an order of magnitude more sales of jam, it put a monetary value on the design decision to limit choices.

Predictably, his examples of the right kind of design were products from Apple.  It takes a lot more effort to put a Nokia E71 in vibrate mode than it does an iPhone.  Spolsky pointed to the iPod’s lack of buttons for Stop and Power as examples of addition by subtraction.  The best example he chose was actually Amazon’s 1-Click shipping.  In addition to offering the most reasoned defense I’ve heard yet of Amazon winning that patent, he explained how it works for multiple purchases.

A few other takeaways from the Spolsky’s keynote that I’ve tried to capture as close to verbatim as possible are:

  • The computer shouldn’t set the agenda.
  • Bad features interrupt users.
  • Don’t give users choices they don’t care about.
iPhone Development: Dan Pilone This talk successfully combined technical depth on iPhone development with information about business models for actually selling an app once it’s complete.  Pilone discussed which design patterns to use (MVC, DataSource, Delegate) as well as what paid applications are selling for in the App Store (the highest-grossing ones sell for between $4.99 and $9.99).

One of the most useful parts of the talk was about the approval process.  He gave his own experience of getting applications through the submission process, including one that was rejected and the reasons why.  According to him, 2 weeks is average time it takes Apple to accept or reject an application.  It’s even possible for upgrades of a previously-accepted app to be rejected.

Pilone did a good job of making it clear that quality is what sells applications.  He used the Convert[6] application (from taptaptap) as an example.  It’s just one of over 80 unit converter applications in the App Store, but it’s beating the competition handily.  OmniFocus was his second example. Revenue Models

  • Ad-supported (very difficult to make money with these)
  • Paid
  • In-app upgrades[7]
Dan Pilone is the co-author of Head First iPhone Development[8], which will be available for sale on October 29.

His recommendation for selling apps in the App Store is to release a paid version first, then an ad-supported version.  This advice seemed counterintuitive to me, but I suspect he suggested it because there’s no data on the in-app upgrades revenue model.  I see in-app upgrades as Apple’s most explicit support for the “freemium”[9] business model yet.

ASP.NET MVC: Scott Allen This talk was basically a demo of a preview version of ASP.NET MVC 2.  Allen wrote code for his demonstration on-the-fly (with the sort of mistakes that can happen using this approach), so the example was pretty basic.  The takeaways I thought were useful for getting started with the technology were:

  • External projects that add features to ASP.NET MVC
    • MVCContrib
    • MVC T4
    • You can combine standard WebForms and MVC in the same solution—particularly useful if you’re trying to migrate an application from ASP.NET to ASP.NET MVC.  Allen mentioned the blogging platform Subtext[10] as an example of one application attempting this kind of migration.
This talk left a lot to be desired.  StackOverflow is the most robust example of what can be built with ASP.NET MVC.  A peek inside even a small part of actual StackOverflow source using ASP.NET MVC would have made a far more compelling presentation.

FogBugz and Kiln Even though this was strictly a product pitch, I’ve included it in the report because of how they implement a couple of ideas: plug-in architecture and tagging.

Plug-in architectures as an idea aren’t new—what was different was the motivation Joel Spolsky described for it.  One of his intentions was to make it easier for people to extend the functionality of FogBugz in ways he didn’t want.  He isn’t a fan of custom fields, so instead of building them into the core product, they’re implemented as a plug-in.  He demonstrated Balsamiq integration (via plug-in) as well, so the architecture does enable extension in ways he likes as well.

Tagging isn’t a new idea either—what I found very interesting is how they apply them in FogBugz.  Spolsky pitched them as a substitute for custom workflow.  His idea (as I understood it) was that bugs could be tagged with any items or statuses outside the standard workflow.  There wasn’t much more detail than this, but I think the idea definitely is worth exploring further.

Python: Bruce Eckel His talk was supposed to be about Python, but Bruce Eckel covered a lot more than that.  The most important takeaways of his talk were these:

  1. In language design, maintaining backward compatibility can cripple a language.
  2. The best programming languages change for the better by not being afraid of breaking backward compatibility.
  3. “Threads are unsustainable.”
Eckel’s talk gave the audience a history of programming languages, as well as a hierarchy of “language goodness”.  For the main languages created after C, the primary motivation for creating them was to fix the problems of its predecessor.  So C++ was an attempt to fix the problems of C, while Java was an attempt to fix C++.  His assertion about the primary motivation behind Ruby was this (I’m paraphrasing):
Ruby intends to combine the best of Smalltalk and the best of Perl.
He made his point about the problems of backward compatibility by comparing an attempt to add closures to Java to language changes made by Ruby and Python.  An article titled “Seeking the Joy in Java” goes into greater detail on the Java side of things.[11] In the case of Java, the desire to maintain backward compatibility often prevents changes to a language which could fix things that are poorly implemented.  The authors of Python and Ruby aren’t afraid to break backward compatibility to make improvements, which makes them better languages than Java (in Eckel’s view).

Here’s Eckel’s hierarchy of programming languages:

Python (his favorite)
Ruby, Scala, Groovy (languages he also likes)
Java
C++
C
Eckel also mentioned Grails as a framework he likes.

Another one of his pronouncements that stood out was a hope that Java would become the COBOL of the 21st century.

Eckel’s argument regarding the difficulty of writing good multithreaded code is one I’ve heard before.  He pointed to Python as a language with libraries for handling both the single processor task-switching and multi-processor parallel execution models of concurrency.

Google App Engine: Jonathan Blocksom Jonathan Blocksom gave a great overview of Google App Engine.  He’s a member of Google’s Public Sector Project Team (not the App Engine team), and I think that helped him present the information from more of an audience perspective.  He did a nice job of describing the architecture and the advantages of using it.  Some of the applications running on Google App Engine include:

Blocksom also discussed some of the limitations of App Engine:
  • 30-second time limit on asynchronous tasks
  • No full text search
jQuery: Richard D. Worth This may not have been the best talk for those already familiar with jQuery, but for me (someone unfamiliar with jQuery), it was close to perfect.  The presenter did an excellent job of showing its advantages over regular ECMAScript.  He used a clever trick to minimize the amount of typing during his demos by using slides with only the changed lines highlighted.  The “find things then do stuff” model he explained made it very easy to grasp what he was doing as he increased the complexity of his examples.[12]

Wrap-up

After the conference ended, a “metaStackOverflow” question was added to collect reviews of the conference from its attendees.[13] The top answer (as of October 28, 2009) also includes links to slides for three of the talks, which I’ve included here:
[1] [blog.stackoverflow.com/2009/05/s...](http://blog.stackoverflow.com/2009/05/stack-overflow-developer-days-conference/)

[2] Ibid

[3] Ibid

[4] gettingreal.37signals.com/toc.php

[5] www.amazon.com/gp/produc…

[6] taptaptap.com

[7] This revenue model is brand-new—Apple only began to support this within the past week or so.

[8] www.amazon.com/Head-Firs…

[9] en.wikipedia.org/wiki/Free…

[10] www.subtextproject.com

[11] www.artima.com/weblogs/v…

[12] Worth used http://jsbin.com/ for some of the more complex demos.  It’s a very good tool I hadn’t seen before.

[13] meta.stackoverflow.com/questions…

A .NET Client for REST Interface to Virtuoso

For my current project, I’ve been doing a lot of work related to the Semantic Web.  This has meant figuring out how to write SPARQL queries in order to retrieve data we can use for testing our application.  After figuring out how to do this manually (we used this SPARQL endpoint provided by OpenLink Software), it was time to automate the process.  The Virtuoso product has a REST service interface, but the only example I found here for interacting with it used curl.  Fortunately, some googling revealed a really nice resource in the Yahoo Developer Network with some great examples.

I’ve put together a small solution in Visual Studio 2008 with a console application (VirtuosoPost) which executes queries against http://dbpedia.org/fct/ and returns the query results as XML.  It’s definitely not fancy, but it works.  There’s plenty of room for improvement, and I’ll make updates available here.  The solution does include all the source, so any of you out there who are interested in taking the code in an entirely different direction are welcome to do so.

Adventures in SPARQL

If this blog post seems different than usual, it’s because I’m actually using it to get tech support via Twitter for an issue I’m having.  One of my tasks for my current project has me generating data for use in a test database. DBPedia is the data source, and I’ve been running SPARQL queries to retrieve RDF/XML-formatted data against their Virtuoso endpoint.  For some reason though, the resulting data doesn’t validate.

For example, the following query:

PREFIX owl: <[www.w3.org/2002/07/o...](http://www.w3.org/2002/07/owl#>) PREFIX xsd: <[www.w3.org/2001/XMLS...](http://www.w3.org/2001/XMLSchema#>) PREFIX rdfs: <[www.w3.org/2000/01/r...](http://www.w3.org/2000/01/rdf-schema#>) PREFIX rdf: <[www.w3.org/1999/02/2...](http://www.w3.org/1999/02/22-rdf-syntax-ns#>) PREFIX foaf: <[xmlns.com/foaf/0.1/...](http://xmlns.com/foaf/0.1/>) PREFIX dc: <[purl.org/dc/elemen...](http://purl.org/dc/elements/1.1/>) PREFIX : <[dbpedia.org/resource/...](http://dbpedia.org/resource/>) PREFIX dbpedia2: <[dbpedia.org/property/...](http://dbpedia.org/property/>) PREFIX dbpedia: <[dbpedia.org/>](http://dbpedia.org/>) PREFIX skos: <[www.w3.org/2004/02/s...](http://www.w3.org/2004/02/skos/core#>) SELECT ?property ?hasValue ?isValueOf WHERE { { <[dbpedia.org/resource/...](http://dbpedia.org/resource/Bank>) ?property ?hasValue FILTER (LANG(?hasValue) = 'en') .} UNION { ?isValueOf ?property <[dbpedia.org/resource/...](http://dbpedia.org/resource/Bank>) } }
generates the RDF/XML output here.  If I try to parse the file with an RDF validator (like this one, for example), validation fails.  Removing the attributes from the output takes care of the validation issues, but what I'm not sure of is why the node ids are added in the first place.

Adding File Headers Made Easy

One of the things on my plate at work is a macro for adding a file header and footer to all the source files in a Visual Studio solution. The macro I put together from my own implementation, various web sources, and a colleague at work accomplished the goal at one time–but inconsistently. So I’d been exploring other avenues for getting this done when Scott Garland told me about the File Header Text feature of ReSharper. You simply put in the text you want to appear at the top of your source file, add a new Code Cleanup profile, check the Use File Header Text option, then run the new profile on your solution.

The result: if the filename ends in “.cs”, ReSharper will add the value in File Header Text as a comment to the top of the file. It’s even clever enough not to add duplicate text if a file already contains it in its header. So if you need to add copyright notices or any other text to the top of your source code files, and you use ReSharper, you’ve already got what you need.

Snow Leopard: Days 1-2

Thanks to a pre-order from Amazon on August 3, a copy of Snow Leopard arrived on my doorstep August 28. The install was uneventful–typical of Mac OS X installs. I put in the DVD, clicked through a few dialog boxes, went to run a couple of errands. When I got back, I logged in as usual.

So far, I haven’t noticed many differences between Leopard and Snow Leopard.  The few of note:

  • Hard drive space.  Before installing Snow Leopard, I had around 14GB of free space.  After installing Snow Leopard (and the latest version of XCode) I have 27GB of free space.  It's quite a bit more freed space than the 7GB Apple advertised
  • Printing.  I have a HP LaserJet 1022.  I had to re-install it after upgrading to Snow Leopard and use an Apple driver.  It still works just fine.
  • Battery Status.  Apple added information on battery health.  Since my MacBook Pro is closing in on 3 years old, the "Service Battery" message is most likely correct.  Apple Support already has a thread about it.  Another thing I've noticed which may also be new to Snow Leopard is that I'm getting battery life percentages for my Bluetooth keyboard and mouse as well.
  • Character/Keyboard Viewer.  A new widget in the upper-right of the screen.  I haven't found any particular use for it yet.
  • Mail.  When I first started it, the app prompted me for some sort of upgrade.  Once it was done, the notes from my iPhone showed up under a Reminders item.
  • Quicken.  I'm still using Quicken 2007 for Mac, so I saw a little prompt about Rosetta when I first launched it.  What I really need to do is get out of Quicken 2007 into something else, but that's a subject for another post.
I can't say I've noticed any speed differences one way or the other so far--but it's only been a couple of days.

Random SQL Tricks (Part 2)

In my previous random SQL tricks post, I discussed how to generate random alphanumeric strings of any length.  A slight variation on that idea that also proved useful in generating test data is the following stored procedure (which generates a varchar consisting entirely of numbers):

CREATE PROCEDURE [dbo].[SpGenerateRandomNumberString] @randomString varchar(15) OUTPUT AS BEGIN SET NOCOUNT ON DECLARE @counter tinyint DECLARE @nextChar char(1) SET @counter = 1 SET @randomString = ''

WHILE @counter <= 15 BEGIN SELECT @nextChar = CHAR(48 + CONVERT(INT, (57-48+1)*RAND()))

SELECT @randomString = @randomString + @nextChar SET @counter = @counter + 1 END END GO

The range in the select for @nextChar maps to ASCII values for the digits 0-9.  Unlike the stored procedure from my first post, there’s no if statement to determine whether or not the random value retrieved is allowed because the ASCII range for digits is contiguous.  The needs of my application restricted the length of this numeric string to 15 characters.  For more general use, the first refactoring would probably add string length as a second parameter, so the numeric string could be a variable length.

A long overdue upgrade

I’m finally running the latest version of WordPress (I’ve been way behind on upgrading).  I’d be curious to hear from those of you who visit regularly what you think of the new look.  I’d considered a couple others:

I ultimately chose this one for the random post feature that appears in the upper-right of the page.

Build Server Debugging

Early in June, I posted about inheriting a continuous integration setup from a former colleague.  Since then, I’ve replaced CruiseControl.NET and MSTest with TeamCity and NUnit 2.5.1, added FxCop, NCover, and documentation generation (with Doxygen).  This system had been running pretty smoothly, with the exception of an occasional build failure due to SQL execution error.  Initially, I thought the problem was due to the build restoring a database for use by some of our integration tests.  But when replacing the restore command with a script for database creation didn’t fix the problem, I had to look deeper.

A look at the error logs for SQL Server Express 2005 revealed a number of messages that looked like:

SQL Server has encountered <x> occurrence(s) of cachestore flush ...
Most of what I found in my initial searches indicated that these could be ignored.  But a bit more googling brought me to this thread of an MSDN SQL Server database forum.  The answer by Tom Huleatt that recommended turning off the Auto-Close property seemed the most appropriate.  After checking in database script changes that included the following:
ALTER DATABASE <database name> SET AUTO_CLOSE OFF

GO

none of the builds have failed due to SQL execution errors.  We’ll see if these results continue.

Random SQL Tricks (Part 1)

One of my most recent tasks at work has been generating test data for integration tests of a new application.  We don’t have the version of Visual Studio which does it for you, and rather than write an app that did it, I spent the past week hunting for examples that just used Transact-SQL.  The initial post that I found the most useful is this one, in which the author provides five different ways of generating random numbers.  I use his third method quite often, as you’ll see in this post (and any others I write on this topic).

One of our needs for random test data was alphanumeric strings of varying lengths.  Because the content of the text mattered less than the need for text, it didn’t have to resemble actual names (or anything recognizable).  The first example I found of a T-SQL stored procedure for generating a random string was in this blog post by XSQL Software.  The script does generate random strings, but they include non-alphanumeric characters.  To get the sort of random strings I wanted, I took the random number generation method from the first post and the stored procedure mentioned earlier and adapted them to this:

CREATE PROCEDURE [dbo].[SpGenerateRandomString] @sLength tinyint = 10, @randomString varchar(50) OUTPUT AS BEGIN SET NOCOUNT ON DECLARE @counter tinyint DECLARE @nextChar char(1) SET @counter = 1 SET @randomString = ''

WHILE @counter <= @sLength BEGIN SELECT @nextChar = CHAR(48 + CONVERT(INT, (122-48+1)*RAND()))

IF ASCII(@nextChar) not in (58,59,60,61,62,63,64,91,92,93,94,95,96) BEGIN SELECT @randomString = @randomString + @nextChar SET @counter = @counter + 1 END END END

The range in the select for @nextChar is the set of ASCII table values that map to digits, upper-case letters, and lower-case letters (among other things).  The “if” branch values in the set are those ASCII table values that map to punctuation, brackets, and other non-alphanumeric characters.  Only alphanumeric characters are added to @randomString as a result.  Having a stored procedure like this one available makes it much easier to generate test data, especially since it can be called from other stored procedures.

Introducing Doxygen

Last Wednesday evening, I gave a presentation on Doxygen at RockNUG.  I didn’t actually bother with slides in order to give as much time as possible to actually demonstrating how the tool worked, so this blog post will fill in some of those gaps.

Doxygen is just one of a number of tools that generate documentation from the comments in source code.  In addition to C# “triple-slash” comments, Doxygen can generate documentation from Java, Python, and PHP source.  In addition to HTML, Doxygen can provide documentation in XML, LaTeX, and RTF formats.

Getting up to speed quickly is pretty easy with Doxywizard.  It will walk you through configuring all the necessary values in wizard or expert mode.  When you save the configuration file it generates, the purpose and effect of each setting is thoroughly documented.  One thing I will note that may not be readily apparent is that you can run Doxygen against multiple directories with source code to get a single set of documentation.  It just requires that the value of your input (INPUT) property contain all of those directories (instead of a single one).

Converting MSTest Assemblies to NUnit

If you wanted to convert existing test assemblies for a Visual Studio solution from using MSTest to NUnit, how would you do it?  This post will provide one answer to that question.

I started by changing the type of the test assembly.  To do this, I opened the .proj file with a text editor, then used this link to find the MSTest GUID in the file and remove it (the guid will be inside a ProjectTypeGuids XML tag).  This should ensure that Visual Studio and/or any third-party test runners can identify it correctly.  Once I saved that change, the remaining steps were:

  • replace references to Microsoft.VisualStudio.QualityTools.UnitTestFramework with nunit.framework
  • change unit test attributes from MSTest to NUnit (you may find a side-by-side comparison helpful)
  • delete any code specific to MSTest (this includes TestContext, DeploymentItem, AssemblyInitialize, AssemblyCleanup, etc)
After the above steps, NUnit ran the tests without any further modifications.  All of my calls to Assert worked the same way in NUnit that they did in MSTest.

MSBuild Transforms, Batching, Well-Known Metadata and MSTest

Thanks to a comment from Daniel Richardson on my previous MSTest post (and a lot more research, testing, & debugging), I’ve found a more flexible way of calling MSTest from MSBuild.  The main drawback of the solution I blogged about earlier was that new test assemblies added to the solution would not be run in MSBuild unless the Exec call to MSTest.exe was updated to include them.  But thanks to a combination of MSBuild transforms and batching, this is no longer necessary.

First, I needed to create a list of test assemblies.   The solution is structured in a way that makes this relatively simple.  All of our test assemblies live in a “Tests” folder, so there’s a root to start from.  The assemblies all have the suffix “.Test.dll” too.  The following CreateItem task does the rest:

<CreateItem Include="$(TestDir)**bin$(Configuration)*.Test.dll" AdditionalMetadata=“TestContainerPrefix=/testcontainer:"> <Output TaskParameter=“Include” ItemName=“TestAssemblies” /> </CreateItem>

The task above creates a TestAssemblies element, which contains a semicolon-delimited list of paths to every test assembly for the application.  Since the MSTest command line needs a space between each test assembly passed to it, the TestAssemblies element can’t be used as-is.  Each assembly also requires a “/testcontainer:” prefix.  Both of these issues are addressed by the combined use of transforms, batching, and well-known metadata as shown below:

<Exec Command=""$(VS90COMNTOOLS)..IDEmstest.exe” @(TestAssemblies->'%(TestContainerPrefix)%(FullPath)',' ‘) /runconfig:localtestrun.testrunconfig" />

Note the use of %(TestContainerPrefix) above.  I defined that metadata element in the CreateItem task.  Because it’s part of each item in TestAssemblies, I can refer to it in the transform.  The %(FullPath) is well-known item metadata.  For each assembly in TestAssemblies, it returns the full path to the file.  As for the semi-colon delimiter that appears by default, the last parameter of the transform (the single-quoted space) replaces it.

The end result is a MSTest call that works no matter how many test assemblies are added, with no further editing of the build script.

Here’s a list of the links that I looked at that helped me find this solution: