The problem

Upstream swank-clojure supports only ELPA and those who want to configure swank manually are on their own. I, for one, do not like those “conventions over configurations” stuff, atleast for my programming environment. Heck, that’s why I use emacs in the first place.

In the past, I was using an older version of swank-clojure which fully supported manual configuration. I then switched to a newer version of swank-clojure but with a totally different emacs lisp code to make configuration for easy.

Also, I use the clojure and clojure-contrib from the debian packages, which installs into /usr/share/java. The same directory has all the other java jars as well. But upstream swank-clojure supported somewhat different directory structure. If invoked on a project, it picks up dependencies from a subdirectory called “lib” (configurable). lein puts all the dependencies (including clojure*.jar) into that directory. It built up a new classpath from these (assuming that the project is self-contained, including dependencies).

This unfortunately does not work in my case. I almost never use a build tool while fooling around doing random with slime. For example, to use Slime with my SICP exercises where I do not have a build mechanism put in, it becomes difficult. Running tests for a namespace from slime is another reason why having the project namespace in the classpath is a very useful thing. For many simple projects that I do, all I need is clojure and clojure-contrib which is already in my classpath. But this made things difficult as adding the project being worked on into the classpath was messy. Becausee of the above setup, swank-clojure-project was unsable for me..

I also don’t like multiple copies of jars lying around. It is not the question of wasting disk space, which is very cheap anyway. But it is just a matter of taste and cleanliness. Multiple jars on each project directory is a convenience, but it is plain ugly.

Today I decided to clean up the whole mess. The end result is a much cleaner swank/slime.

Solution

Turned out the solution was very simple. When swank is started on a project, instead of changing swank-clojure-classpath, I build up another list using a variable called swank-clojure-extra-classpaths. This is then appended to the swank-clojure-classpath before starting up the JVM.

I also disabled the code which checks the presence of clojure*.jar files in some predefined location when swank starts up. This can always be called manually, if one wants to.

Code

For those wanting to follow similar approach, here is my github fork of swank-clojure with instructions on setting it up. The emacs repl (via ielm) was invaluable for debugging.

I will walk through creating a trivial clojure project using maven. I assume that maven is already installed on your machine.

First, run the following command:

$ mvn archetype:create -DgroupId=in.clj.hello -DartifactId=hello

This gives the following output on my machine:

[INFO] Scanning for projects...
[INFO] Searching repository for plugin with prefix: 'archetype'.
[INFO] ------------------------------------------------------------------------
[INFO] Building Maven Default Project
[INFO] task-segment: [archetype:create] (aggregator-style)
[INFO] ------------------------------------------------------------------------
[INFO] Setting property: classpath.resource.loader.class => 'org.codehaus.plexus.velocity.ContextClassLoaderResourceLoader'.
[INFO] Setting property: velocimacro.messages.on => 'false'.
[INFO] Setting property: resource.loader => 'classpath'.
[INFO] Setting property: resource.manager.logwhenfound => 'false'.
[INFO] [archetype:create {execution: default-cli}]
[WARNING] This goal is deprecated. Please use mvn archetype:generate instead
[INFO] Defaulting package to group ID: in.clj.hello
[INFO] ----------------------------------------------------------------------------
[INFO] Using following parameters for creating OldArchetype: maven-archetype-quickstart:RELEASE
[INFO] ----------------------------------------------------------------------------
[INFO] Parameter: groupId, Value: in.clj.hello
[INFO] Parameter: packageName, Value: in.clj.hello
[INFO] Parameter: package, Value: in.clj.hello
[INFO] Parameter: artifactId, Value: hello
[INFO] Parameter: basedir, Value: /tmp/hello-mvn
[INFO] Parameter: version, Value: 1.0-SNAPSHOT
[INFO] ********************* End of debug info from resources from generated POM ***********************
[INFO] OldArchetype created in dir: /tmp/hello-mvn/hello
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESSFUL
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1 second
[INFO] Finished at: Wed Apr 21 22:13:22 IST 2010
[INFO] Final Memory: 15M/150M
[INFO] ------------------------------------------------------------------------


A directory called hello is created. Let us look at the parameters to archetype:create. groupId is a unique way to identify the module. Think of it as the namespace. artifactId is the name of the project. groupId:artifactId:version uniquely identify an ‘artifact’. By default, maven will create a version number 1.0-SNAPSHOT for the project, if it is omitted from commandline.

The directory tree at this point looks like this:

$ tree
.
|-- pom.xml
`-- src
    |-- main
    |   `-- java
    |       `-- in
    |           `-- clj
    |               `-- hello
    |                   `-- App.java
    `-- test
        `-- java
            `-- in
                `-- clj
                    `-- hello
                        `-- AppTest.java

Let us get rid of the java files first.

$ rm -rf src/main/java
$ rm -rf src/test

Create the following directory structure:

$ mkdir -p src/main/clojure/in/clj

Note that we have created the structure according to our groupId specification.
Create a file called hello.clj at the leaf of this directory and put the following contents inside it:

(ns in.clj.hello)

(defn greet [ & args]
  (apply str args))

Now, open the pom.xml file. Some changes needs to be done here so that maven knows that we are building a clojure project and the dependencies. The following things are done:

1. The dependency on junit is removed.
2. A dependency on clojure is added.
3. A new plugin clojure-maven-plugin is added and configured.
4. Repositories from where, dependencies are fetched, are added.

At the end of this process, the pom.xml looks like this:

<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd">
  <modelVersion>4.0.0</modelVersion>
  <groupId>in.clj.hello</groupId>
  <artifactId>hello</artifactId>
  <packaging>jar</packaging>
  <version>1.0-SNAPSHOT</version>
  <name>hello</name>
  <url>http://maven.apache.org</url>
  <description>project to demonstrate maven</description>

  <build>
    <plugins>
      <plugin>
        <groupId>com.theoryinpractise</groupId>
        <artifactId>clojure-maven-plugin</artifactId>
        <version>1.3.2</version>
        <executions>
          <execution>
            <id>compile-clojure</id>
            <phase>compile</phase>
            <goals>
              <goal>compile</goal>
            </goals>
          </execution>
        </executions>
        <configuration>
          <sourceDirectories>
            <sourceDirectory>src/main/clojure</sourceDirectory>
          </sourceDirectories>
        </configuration>
      </plugin>
    </plugins>
  </build>

  <dependencies>
    <dependency>
      <groupId>org.clojure</groupId>
      <artifactId>clojure</artifactId>
      <version>1.1.0</version>
    </dependency>
  </dependencies>

  <repositories>
    <repository>
      <id>clojure-snapshots</id>
      <url>http://build.clojure.org/snapshots</url>
      <releases>
        <enabled>false</enabled>
      </releases>
      <snapshots>
        <enabled>true</enabled>
      </snapshots>
    </repository>
    <repository>
      <id>clojure-releases</id>
      <url>http://build.clojure.org/releases</url>
      <releases>
        <enabled>true</enabled>
      </releases>
      <snapshots>
        <enabled>false</enabled>
      </snapshots>
    </repository>
  </repositories>

</project>

Now do:
$ mvn package
This produces the following output on my machine:

[INFO] Scanning for projects...
[INFO] ------------------------------------------------------------------------
[INFO] Building hello
[INFO]    task-segment: [package]
[INFO] ------------------------------------------------------------------------
[INFO] [resources:resources {execution: default-resources}]
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory /tmp/hello-mvn/hello/src/main/resources
[INFO] [compiler:compile {execution: default-compile}]
[INFO] No sources to compile
[INFO] [clojure:compile {execution: compile-clojure}]
Compiling in.clj.hello to /tmp/hello-mvn/hello/target/classes
[INFO] [resources:testResources {execution: default-testResources}]
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory /tmp/hello-mvn/hello/src/test/resources
[INFO] [compiler:testCompile {execution: default-testCompile}]
[INFO] No sources to compile
[INFO] [surefire:test {execution: default-test}]
[INFO] No tests to run.
[INFO] [jar:jar {execution: default-jar}]
[INFO] Building jar: /tmp/hello-mvn/hello/target/hello-1.0-SNAPSHOT.jar
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESSFUL
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 4 seconds
[INFO] Finished at: Wed Apr 21 22:57:11 IST 2010
[INFO] Final Memory: 23M/121M
[INFO] ------------------------------------------------------------------------

A jar file is created under the target directory.
Now, we add the jar into our classpath and invoke the clojure repl:

$ java -cp /usr/share/java/clojure.jar:target/hello-1.0-SNAPSHOT.jar clojure.main
Clojure 1.1.0
user=> (use 'in.clj.hello)
nil
user=> (greet "Hello World")
"Hello World"
user=>

You can also invoke the repl from maven:

$ mvn clojure:repl

That wasn't really too hard. With emerging tools around maven like the Polyglot Maven which supports Clojure and a S-expression syntax for writing the POM, we don't even have to do as much as we did now.
I have put this example in this repository for anyone to play with.

This book is not strictly in the category of “self help” books meant to give a temporal high. It is a set of advice (54 of them) for any programmer to improve himself. Each of the chapter is only 2 or 3 pages long, so it is quite easy to read and digest. I have put some of my notes on the book in this wiki page. The author had earlier lived in India, apparently managing an outsourced team here and makes some good and correct observations on the general attitudes of Indian software programmers.

The book is worth a read. If you have no time to read the entire book and just need a summary, read my notes.

Suppose you have a map of people with their name and ages:

(def from-xml [{:name "Sean" :age 27} {:name "Ross" :age 27} {:name "Brian" :age 22}])

Now, you want to have the name and age of people in this list of those with the age 27. Clojure’s comp function comes to the rescue. comp takes some functions and applies them to the data from right to left. i.e., It takes the right most function, applies it to the input parameters, the result of which goes as input to the next right most function and so on.
Let us take one single map and pass it to the comp:

user> ((comp #{33} :age) {:name "Ram" :age 33})
33

What this does is, it takes the input map, applies the :age function on the map (Clojure’s hashmap keys are functions themselves). The result of this is used to see if it falls in the given set (which just has 33).  Now, sets are functions of their members. Here is an example:

user> (#{33 34} 34)
34
user> (#{33 34} 33)
33
user> (#{33 34} 32)
nil
user> (#{33} 33)
33


We use this composition as a predicate to the filter function which results in the elegant solution that Sean posted.

user> (filter (comp #{27} :age) from-xml)
({:name "Sean", :age 27} {:name "Ross", :age 27})

Note that we get back maps in the same format, which is really nice in certain situations. Sean’s Episode 16 also has quite a lot of other goodness as well.

One noteworthy suggestion in the book for every reader is to keep a personal wiki, which I had experimented with, sometime last year, with great success. I especially liked the meditation technique described in the book and can say with some confidence that it works very well. The chapter on “focus” is also very well researched. The book says that each context switch costs about 20 minutes. Interestingly, the book does not talk about the concept of flow.

I greatly enjoyed reading “Pragmatic Programmer: From Journeyman to Master” earlier (despite the extremely bad packing from Flipkart, which made the paper curly, sort of permanantly).

$ find ./ -name *.pdf -print | xargs -I xxx cp xxx /tmp

The key is the -I argument to xargs which replaces every occurance of the pattern in the command with the input from stdin.

Also don’t the pipes look like calling a function composition like in f(g(x) as mentioned by this post?

• read a section of the book.
• solve the exercises in the section.
• see the videos.
• come with doubts/insights etc for the next meetup.

We agreed to meetup once in 2 weeks.

So, here I am, (re)starting the SICP from today. Hopefully this time I will complete the book. I am very tempted to do the exercises in clojure, so that I won’t copy the code as it is from the text. I am not sure whether future chapters can become difficult for me  because of this, but I think I will give it a try with clojure. Also I believe I will need only a subset of clojure so that use only the fundamentals and no advanced facilities.

Here are some of the available resources that I plan to use:

• The book itself, which is available online and the exercises.
• The OCW lecture notes.
• The Berkeley CS61A videos (see the panel on the right for the following lectures, 44 of them).
• The Problem sets (or projects).
• The video lectures given by the authors at HP (based on first edition).
• Clojure/Emacs/Slime/Paredit.

My solutions will be available in this github repository.

Of course, I/we are not the first to do this. Here are some of the other people who have attempted this and have documented it (much better than I could ever do) on the web.

• Eli Bendersky’s SICP discussions.
• Drew Hess’s SICP wiki.
• Ken Dyck’s SICP solutions.
• Bill the lizard’s solutions.
• Weiqun Zhang’s solutions.
• … and many more.

For those who are using Scheme, this page will be of great help.

Linux kernel provides a wealth of information (through ACPI) about the Battery like the make, type, serial number etc. Look under
/sys/class/power_supply/BAT0/. This is what I have:

~$ cat /sys/class/power_supply/BAT0/status
Unknown
~$ cat /sys/class/power_supply/BAT0/technology
Li-ion
~$ cat /sys/class/power_supply/BAT0/type
Battery
~$ cat /sys/class/power_supply/BAT0/serial_number
32511
~$ cat /sys/class/power_supply/BAT0/manufacturer
SANYO
~$ cat /sys/class/power_supply/BAT0/model_name
42T4511
~$ cat /sys/class/power_supply/BAT0/energy_full_design
84240000
~$ cat /sys/class/power_supply/BAT0/energy_full
64530000
~$ cat /sys/class/power_supply/BAT0/energy_now
63160000

It looks like, my battery (this is a work laptop, Lenovo Thinkpad T61 running Debian x86-64 unstable) is functioning at a lesser capacity than it is designed. It will be nice to corelate these readings with the actual performance of the battery. I tend to run this machine mostly with AC, but I will try profiling it for battery performance and see how much time these numbers translate to.

$ git branch -m <old-branch-name> <new-branch-name>
$ git branch (should show the new name)

Now, do a git log and find out which is the commit before you started making the change and do:

$ git checkout -b <old-branch-name> <commit-sha1>

That’s it and you have the whole thing as you wanted it to be.

Thanks to my friend Anand for the tip on the -m option to git branch. Most git gems seem to be deep embedded in those switches of the sub commands of git.