treehugger is a library to write Scala source code programmatically. It’s also an implementation of Scala AST based on Reflection API, now available from github eed3si9n/treehugger.
Edit: I’ve expanded this into a complete guide using awesome n8han/pamflet.
Scala the language is one of the most elegant, expressive, consistent, and pragmatic languages. From pattern matching to the uniform access principle, it got so many things right. And Scala the ecosystem and Scala the community only makes it better.
In Scala 2.9.1, locally declared implicits are preferred over imported ones. The problem is that the spec does not cover such behavior. My original hypothesis was that either I did not understand the spec correctly, or the spec was wrong. Based on the assumptions, I set out to explore the implicits resolution precedence last week. Like MythBusters say, the best kind of result is when you get something totally unexpected. It turns out that both of the hypotheses were wrong.
Northeast Scala Symposium 2012 is coming up in a few months, but I want to revisit a talk from this year’s nescala to wrap up 2011. One after the other, nescala had amazingly high quality of talks. You can check them all out here. With Daniel’s Functional Data Structure and Jonas’s Akka each having an hour-long key notes, the symposium left an impression on me that actors and FP are two major forces within Scala community. (Paul declaring that sending messages to actors is not referentially transparent was a hint too, I guess) There were also earlier signs of how the year turned out, like Mark’s sbt 0.9 presentation and Nermin’s Scala performance consideration. One talk that stood out in terms of immediate impact to change my code was Josh’s talk: Implicits without the import tax: How to make clean APIs with implicits.
version 2.0
When the original version was published on 06/19/2011, the motive for writing this guide was to aid the effort of moving people over to sbt 0.10 from 0.7, inspired by Mark’s sbt 0.10 demos that I was able to see live (first at northeast scala, and second at scala days 2011). At the time, the plugins were considered to be a major roadblock to the migration, since build users can’t move to 0.10 without the plugins. So my strategy was to port the plugins myself if they weren’t there, ask questions on the mailing list when I get stuck, and write up the results. I’ve gotten many positive feedbacks, and it’s helped people get on to 0.10. However, as it turns out, my understanding of sbt 0.10 wasn’t always complete, and downright wrong and misleading at times. I take responsibility of my writing. Instead of leaving old contents in, I’ve decided to push it into github, make a new version and move on. The most up-to-date knowledge of writing plugins is compiled in Plugins Best Practices written mostly by Brian and Josh, and a tiny section by me.
NOTE: Official docs: http://www.scala-sbt.org/1.x/docs/Testing-sbt-plugins.html
Let’s talk about testing. Once you write a plugin, it turns into a long-term thing. To keep adding new features (or to keep fixing bugs), writing tests makes sense. But how does one go about testing a plugin to a build tool? We fly, of course.
sbt comes with scripted test framework, which let’s you script a build scenario. It was written to test sbt itself on complex scenarios such as change detection and partial compilation:
I wish there was a common super trait for various future implementations in the standard library, so I can express the concept without tying the code to a specific platform stack. I am not sure if there are others who feel the same, but I think it would be useful for library authors. That’s my motivation of writing sff4s.
You’ve probably come across the notion before but let’s go over it quickly. A future value (also known as promise) represents an incomplete calculation.
On 7/13/2011 I hosted a ny-scala meetup on sbt 0.10 migration with doug tangren and rose toomey. Here is the slides.
I’ve tried other IDEs for Scala, but always ended up going back to TextMate. As I am getting on the IntelliJ IDEA wagon, I’ve made a Twilight theme to make the ride smoother.
Ultimately, the users of scalaxb are interested the real problems that the entity objects express, not how they persist into XML. That’s why I knew I eventually had to vacate the singleton/companion object of the case class to implement the data binding. Until recently it has been generating the data binding implementation as follows:
object Address extends rt.ElemNameParser[Address] {
val targetNamespace = "http://www.example.com/IPO"
def parser(node: scala.xml.Node): Parser[Address] =
...
def toXML(__obj: Address, __namespace: String, __elementLabel: String, __scope: scala.xml.NamespaceBinding): scala.xml.NodeSeq =
...
}
Now scalaxb has hijacked the prime real estate for the purpose of XML data binding, which has little to with Address.
See Create a global .gitignore.
$ git config --global core.excludesfile ~/.gitignore_global
See Checking out pull requests locally.
[remote "origin"]
url = git@github.com:sbt/sbt.git
fetch = +refs/heads/*:refs/remotes/origin/*
fetch = +refs/pull/*/head:refs/remotes/origin/pr/*
For some reason, I can’t keep OSGi in my head. Everything I read about it slips away in a few weeks, and I have re-read the guides and tutorials.
Here’s a memo of setting up OSGi bundle written in Scala using NetBeans, following Neil Barlett’s OSGi in Practice, except the book uses Eclipse.
Skip this section if you use non-Mac. Install MacPorts. Run the following from Terminal:
One of my favorites from msdn is Application Architecture for .NET: Designing Applications and Services by Eduardo Jezierski. The version 2.0 is Microsoft Application Architecture Guide by J.D. Meier et al, but it’s much more beefier than the original. Currently 1.0 is put under Retired node in msdn.
Once you use any technology to a significant extent, you understand its strength and shortcomings. I’ve probably written tens of thousands of lines of code in Delphi or C++, Java and C# too to some extent. I’ve depended on those languages, but gripe about them too. The new language obviously has the advantage of just being new. Since I haven’t written anything significant in Scala besides Tetris, I haven’t hit the shortcomings yet.
I haven’t read the original book, so this is purely based on the film work by Spike Jonze.