Nuo Yan : 5. Programming

Class instance variable for ruby modules

Nuo Yan — Fri, 21 Aug 2009 05:54:00 GMT

Sorry if the title is a bit misleading, but this post talks about the scenario such that I have several classes, all extending the same module, and I need to avoid using shared or class variables (@@) in the module because the classes extending the module may change the shared or class variable at the same time and cause problem.

So basically we have:

class Test

extend TestModule

end

class TestTwo

extend TestModule

end

module TestModule

@@val

def somemethod

#do something with @@val

end

And the goal is the replace the @@val class variable with a class instance variable so that if the Test.somemethod and TestTwo.somemethod happens at the same time, there won't be any problem.

The easiest way (I can think of) of doing this is to set the variable within the class, instead of setting it in the module. For example:

class Test

@val = "abc"

class << self; attr_accessor :val; end

extend TestModule

end

But the problem of doing this is that everybody else using TestModule will need to update their code after this update, which is unacceptable.

Is there a way to move the variable definition and accessor to the module and has exactly the same effect as defining them in the class? Yes. define the self.extended(base) method.

module TestModule

def self.extended(base)

class << base; attr_accessor :val; end

base.val = "abc"

end

def somemethod

#do something with self.val

end

class Test

extend TestModule

end

Here base refers to the class extending this module. self. extended is called when the module is initialized, and the accessors for the base class is then created (class << base; attr_accessor :val; end), base.val = abc sets the default value.

By doing it this way, the module works exactly the same as before, and multiple classes can extend it and call the same method without affecting each other.

Hope this helps.

Don't forget the port number when replicating CouchDB

Nuo Yan — Fri, 17 Jul 2009 02:52:00 GMT

In the afternoon I needed to replicate a database from my co-worker's CoudhDB using CouchDBX. In the replicator, I entered my co-worker's database's path like the following:

http://10.0.0.3/TestDatabase

And I selected to replicate to one of my existing databases. Everything seemed to be right (I could ping and telnet my co-worker's IP and CouchDB port 5984 without problem), but CouchDBX returned error 500, could not connect to remote database.

After a series of diagnostics, we finally realized, we forgot to put the port number in the path of remote database. When we changed the database path to http://10.0.0.3:5984/TestDatabase, everything just worked very well.

I'm sharing this since I bet many people may encounter the same problem. Hope this is helpful.

Creating a Merb Rack middleware for running in Merb applications

Nuo Yan — Tue, 14 Jul 2009 03:54:00 GMT

There are a lot of resources on the Internet showing how to write and run a standalone rack application or middleware, but few showing how to write one to run in other Merb applications. This is actually very easy, so I'm going to share how I did it. Hopefully to be helpful to people in such need.

The goal is to have a rack middleware in a merb application, which can be installed as gem, and then can be ued in other merb applications.

First, create the merb application:

$ merb-gen app  application_name

where application_name is your proposed application name.

Second, create your code folder and ruby file. I created a "lib" folder to be the container of my middlewares.

Third, open the ruby file and write your middleware. (In this example, let's name it samplemiddleware.rb within the "lib/test" folder)

The class could inherit Merb::Rack::Middleware if you want. Below is an example.

module Test
class SampleMiddleware < Merb::Rack::Middleware

    def initialize(app)
      @app = app
    end

    def deferred?(env)
      @app.deferred?(env) if @app.respond_to?(:deferred?)
    end

    def call(env)
      puts "do something here"
      @app.call(env)
    end

end
end

Fourth, write your rake file, so the rack middleware merb application can be installed as a gem.

Fifth, install the gem.

So far we finished the middleware, now we go to the other merb application which we want the middleware to run within. Assuming the merb application has a rack configuration file (rack.rb) under the "config" folder. Open this file.

We need to require our middleware, in this case:

require 'test/samplemiddleware.rb''

And then use the middleware in the same file:

use Test::SampleMiddleware

Finally remember to have a line to run the Merb::Rack application (should already have it):

run Merb::Rack::Application.new

Now when the merb application runs, the rack middleware you wrote and installed will run as well. It's handy to use the middleware to process requests in your Merb applications before Merb handles the requests.

Modify bind address for couchdb 0.9.0

Nuo Yan — Thu, 28 May 2009 23:21:00 GMT

By default couchdb only binds to 127.0.0.1 so it's not accepting requests from other IP addresses. The bind address setting can be modified in the configuration file. For 0.9.0, by default, the default configuration file is located at /usr/local/etc/couchdb/default.ini, and the local configuration file is located at /usr/local/etc/couchdb/local.ini. Settings in the local configuration file will override those in the default one.

Bind address is in the [httpd] section

[httpd]

port = 5984

bind_address = 127.0.0.1

Note that in the local configuration file, this may be commented out with a leading ";", if you want to set it in the local configuration file, you need to uncomment the bind_address line and modify 127.0.0.1 to some other IP address.

If you modify the bind address to 0.0.0.0, it will bind all interfaces.

Making an executable standalone Merb Slice gem

Nuo Yan — Thu, 21 May 2009 07:31:00 GMT

I have a standalone Merb Slice which can be started by running the "slice" command in its top directory. Today I needed to make a gem file for this Slice so that people can install the gem and start it directly by running the Slice's name, for example, "foo".

So I created the Rakefile, in which I specified the executable file in the following way:

--- Code: Rakefile ---

spec = Gem::Specification.new do |s|
s.rubyforge_project = 'merb'
s.name = GEM_NAME
s.version = GEM_VERSION
s.platform = Gem::Platform::RUBY
s.has_rdoc = false
s.summary = SUMMARY
s.description = s.summary
s.author = AUTHOR
s.email = EMAIL
s.homepage = HOMEPAGE
s.executables = "foo"
s.require_path = 'lib'
s.autorequire = GEM_NAME
s.files = %w(Rakefile TODO) + Dir.glob("{config,lib,spec,app,data,public}/**/*")
end

Then I created a bin folder within the Slice's top directory, and created a file named "foo" within the bin folder I put Merb.start in the file "foo", and I tried to make a gem and installed it. But the Slice did not start correctly this way.

Then I spent a while trying different possibilities, and with somebody's help, I finally got the right way of doing it.

1. look up the location of merb-slices. This can be done by executing gem list - d | grep merb-slices.

2. locate a file named "slice" under the bin folder of merb-slices.

3. copy the contents of this file to the file you have in your Slice's bin folder, and amend from that (you may want to remove or change the Merb::Config.use block since your own Slice may have that set somewhere else, for example, in init.rb).

Depending on how you amend the slice binary file, you may encouter a situation that your executable is only able to run within the source folder. For example, if you run "foo" in your Slice's source folder, it becomes equivalent as running "slice", but if you run it anywhere else, it fails with the error "No slice found". If this happens, what you need to do is change the following things in your "foo" binary file:

1. Change __DIR__ (originally assigned to current directory):

Dir.chdir File.join(File.dirname(__FILE__),"..")
__DIR__ = Dir.getwd

2. Update slice_name:

slice_name = "foo"

3. Add a line before Merb.start to specify the right init file to load:

ARGV.push *[ "-I", File.join(File.dirname(__FILE__), "..", "config", "init.rb") ]

Note: This assumes init.rb is under config folder of your Slice project.

Testing exceptions with RSpec

Nuo Yan — Mon, 20 Apr 2009 22:54:00 GMT

I'm new to ruby and rspec. I wrote a ruby function and wanted to check whether it throws an exception or not in a certain situation with rspec, but I realized should_raise function no longer worked in the rspec version I was using (1.2.2). I tried to find information on the Internet and I found that I should use raise_error.

However, when I wrote something like:

it "description" do
ObjectName.new("parameterName").should raise_error
end

It did not work. Finally I learned the correct way - using lambda:

it "description" do
lambda{ObjectName.new("parameterName")}.should raise_error
end

Wanted to post this here and hope it can be helpful for people in the similar need.

Using Eclipse to add/update files to CVS

Nuo Yan — Wed, 11 Feb 2009 01:31:00 GMT

The command-line based CVS tool is hard to use for first time users; many GUI-based CVS tools are not really well designed to be easy to use either. I’m not against the point that developers and testers should learn how to use a technical tool in no time even though the tool is hard to use, but developers and testers have their schedules and need to be efficient. Having an efficient way to add and update files to CVS is very necessary.

Eclipse is the tool I use to deal with CVS file operations. It works really well not only with source code, but any other file type.

For example, if you want to upload a Microsoft Word document (docx) to a remote CVS repository, simply follow the following steps:

1. In Eclipse, import the remote CVS repository by using the Import wizard:

a. Click File->Import

b. Choose “Projects from CVS” and click Next

c. Depends on real situation, either select “Create a new repository location” or “Use an existing repository location”. In this example, we use an existing repository location because we want to add the Microsoft Word document to an existing CVS repository. Click Next.

d. Select “Use an existing module”, and navigate the tree list to find the repository you want to add the file to. Click Next.

Note: If you need to upload different files to different subfolders (repositories) within a folder (repository), it’s often useful to select the parent folder in this step so you don’t need to run this wizard every time for each of the subfolders.

e. Follow the remaining steps in the wizard to finish the import process. Selecting “Check out as a project in the workspace” is the easiest way to import the repository to a new project. If you want to import it into your existing project, select “Check out into existing project.”

2. Import the file(s) you want to add/update to the CVS repository to the appropriate folders:

a. In the Project Explorer, expand the folder you just checked out from CVS, right click the folder you want to add the new file to, and click “Import” from the context menu.

b. Select “File System” and click Next.

c. Click “Browse…” to browse the folder containing the file you want to add or update to CVS.

d. Select the files, and click “Finish”.

3. Now the file(s) you want to add or update to CVS are in the appropriate folders. Right click each file, and click Team-> Commit to commit the file to the corresponding remote CVS repository.

Hope this helps.

Book Recommendation

Nuo Yan — Wed, 30 May 2007 06:04:00 GMT

Writing Solid Code: Microsoft's Techniques for Developing Bug-Free C Programs (Microsoft Programming Series)

Yes it's an old book which was published in 1993, but it's really awesome! I first got this book in last Saturday night, then I spent several hours finished reading it. I just couldn't stop until I finished!

I recommend every C programmer to read it, if you haven't done so yet.

Nuo

Calculating Required Storage In a Direct-Mapped Cache

Nuo Yan — Fri, 04 May 2007 04:58:00 GMT

Suppose a 1-level direct-mapped cache need to have 64 KB's data, each block is 16-byte, and the memory address is 32-bit.

To calculate the total bits needed for the cache, we could go through the following steps:

We need to have 64 KB's data (2^16 bytes) in the cache, and each block is sized as 16 bytes (2^4), so we totally need 2^12 blocks for the data, let's call the "12" as "a." This means a bits are used for the index.
Each word has 4 bytes, then in this case each block has 2^2 words, let's call the exponent 2 as "b." This means b bits are used for word in each block.
The size of the tag is determined by the equation "length of address - (a+b+2)," the 2 bits here are the byte offset of the address. So, in this case, it's "32 - (12+2+2) = 16."
Each block has 16 bytes, each byte has 8 bits, so the amount of data for each block is 16 * 8 + tag + valid bit = 128 + 16 + 1 = 145 bits.
Times the amount of data for each block (145) to the number of blocks (2^12). Then we get 593920 Bits. Convert to KB, it is 72.5 KB.

So 72.5 KB's of data is required to make a 64KB 1-level direct-mapped cache .

In general, the idea is very simple: total bits required = number of blocks * bits required for each block. Mistakes can usually be made during the process of calculating the size of the tag. Sometimes it's also easy to forget to add the valid bit. But if doing it carefully, we could realize that the way of calculating it is quite simple.

Notes on Compiler Constructions

Nuo Yan — Tue, 30 Jan 2007 01:13:00 GMT

Introduction:

This document briefly introduces the concepts and construction of programming language compilers. The document uses a simple programming language we named “D” for the purposes of this project, then goes through the process of constructing a full compiler for the “D” programming language. “D” only represents the name of the sample programming language used in this project, and has no relationship with any popular existing programming language. Code samples will be included in this document; however, since this document will be published on the Internet, to prevent future computer science students with similar assignments from plagiarizing any part of this paper, the full source code for the compiler project will not be included in this document. All source codes for the compiler in this document are written in C, and compiled using gcc in Unix/Linux. The source code has also been tested for compilation under Windows NT (including 2000, XP, and Vista) using Visual C++ (including 6.0, 2003, and 2005). The X86 assembly language code presented in this documents uses the Microsoft/ Intel standard, which is different from the GNU/AT&T standard in some respects. This document is neither a course requirement, nor an assignment. All contents represent the author’s personal interests and the author provides no warranty.

To view the full document, either:

Download the attachment of this post, which is in *.zip format.
Click on the following link: http://students.washington.edu/~nuoyan/compiler.pdf.

Nuo Yan : 5. Programming

Class instance variable for ruby modules

Don't forget the port number when replicating CouchDB

Creating a Merb Rack middleware for running in Merb applications

Modify bind address for couchdb 0.9.0

Making an executable standalone Merb Slice gem

Testing exceptions with RSpec

Using Eclipse to add/update files to CVS

Book Recommendation

Calculating Required Storage In a Direct-Mapped Cache

Notes on Compiler Constructions

Table of Contents:

Introduction: