Monday, October 29, 2012

Standardized Markdown?

Having used Markdown styles at both GitHub and StackOverflow, I really like this idea:
I propose that Stack Exchange, GitHub, Meteor, Reddit, and any other company with lots of traffic and a strategic investment in Markdown, all work together to come up with an official Markdown specification, and standard test suites to validate Markdown implementations. We've all been working at cross purposes for too long, accidentally fragmenting Markdown while popularizing it.

Wednesday, October 10, 2012

Java command line app with Spring and Gradle

Many Java developers have had to build a command line app for one reason or another over the years, and there are many(!) ways to do it.  There are a bevy of command line libraries, some people roll their own when they get to this point (which is one reason why there are so many command line libraries!), and others just pull in the arguments simply and directly, eschewing a library.

However, writing a tool, a developer should just focus on the meat of the problem rather than having to worry about the enabling functionality that allows it to work.  The Spring Framework very much supports this approach to development with its inversion of control (IoC) container and a myriad of other features.

To support that approach for command line applications, the spring-cmdline library provides a bridge between the Spring framework and the JOpt Simple command line library.

This tutorial will show how to leverage the spring-cmdline library to stand up a very simple command line application, using Gradle to accelerate getting a build up and running.

Wednesday, March 7, 2012

Groovy Mocks and Java Classes

Trying to convert a Java/EasyMock based JUnit test to Groovy in the last couple days led me to find what I think is a bug in Groovy's MockFor class.

The documentation for MockFor says to use the proxyDelegateInstance method when testing Java classes, but unfortunately the delegate proxy obtained using that method is type-incompatible with the Java class you're trying to mock if that class is concrete.  It works for Groovy classes, all abstract classes, and all interfaces, but it appears to fall down when given a plain, concrete Java class.

Monday, March 5, 2012

Groovy and String Comparisons

Though I've recently come to love working with Groovy, I'm deep enough where I'm starting to see some of its warts.

The one that hit me nearly right away was the nastiness about groovy.lang.GString not extending java.lang.String.  This is mainly the fault of the latter being final and not extendable, but it means that GString instances aren't automatically comparable with plain Java strings.
In Groovy, string literals can be quoted using both single and double-quotes.  GString appears when you use double-quotes around a string and expand variables into it.  See "Strings and GStrings" for more information.

Monday, November 21, 2011

As much as I like Windows 7 relative to previous versions of Windows, I'm again taking a stab at using Linux as a development desktop OS.  This is complicated by the fact that I still have to work in a shop where the supported IS boxes are all Windows 7.

I'm a bit disappointed that open source support for Microsoft Exchange seems to be waning, as Evolution doesn't support newer versions of Exchange.

I'm very happy with Linux Mint 11, as it seems to correctly do a number of things that vanilla Ubuntu 11 doesn't do out of the box.

Right now I'm working on getting my development environment functional again.  I've used Component Software Diff on Windows for years, since it's no-nonsense, visually appealing, and usually just works.  Under Linux, I'm trying out Meld and followed a tutorial to get this working with git, though firing up Python just to launch another command-line tool seems a little overkill when this would suffice:
#!/bin/bash
meld $2 $5

Wednesday, October 26, 2011

manifest classpath + taglib jars + Tomcat = FAIL


I found an interesting side-effect of using manifest classpaths in our application jar files as it applies to taglib jar dependencies and embedded Tomcat 6.

·         Manifest classpaths do not appear to be expanded in the URLClassLoaders that load the jars with the manifests, so only the jar file referenced directly on the classpath are included in its URLs. The taglib search mechanism in Tomcat 6 (haven't checked Tomcat 7 to see if there's a difference) only walks the classpath looking for URLClassLoaders, looking at the underlying JAR URLs. Thus, if the taglib jar is referenced only as a dependency in the manifest of the main jar file and is not specifically on the classpath, it will not be spotted and use of that taglib by the JSP engine will throw an exception.

This problem won’t actually show up in tests as IDE environments and builds generally include all the jar files in the classpath, so this one will bite you at runtime.

The workaround for this is to include taglib jar files specifically on the classpath, which is something that using the manifest classpath is supposed to help avoid.

Le sigh.

Friday, September 17, 2010

Difficult time voting in primaries this year

For the primaries this year, I went to our local precinct polling center early in the morning. I discovered that there were some barriers to voting. Specifically, the roads and sidewalks all around the polling place were under active construction. A front-end loader was driving around dumping large loads of dirt right next to the torn-up sidewalks to get inside. The main sidewalk right outside the side door where we were detoured to was fresh enough to look like you shouldn't walk on it.

(Posted late as I only recently managed to get this image off my phone.)
Posted by Picasa