Liah Hansen

Checking A Column's Datatype in a Rails 3 Migration

2011-04-28T00:00:00-07:00

Checking A Column's Datatype in a Rails 3 Migration

Today we were temporarily perplexed because Facebook ids were being saved in MySQL as completely different numbers than we were expecting. The problem ended up being that the datatype was Int instead of BigInt.

We wanted to be able to determine the datatype of a column through ActiveRecord in our migration. This is how to get that info:

Widget.columns_hash["id"].sql_type

Rails Console - Show Error Messages for All Object Attributes

2011-04-15T00:00:00-07:00

Rails Console - Show Error Messages for All Object Attributes

To view all of the errors on each of the attributes of your ActiveRecord object (Rails 3), use Object.errors.full_messages:

 > u = User.new
 => #<User id: nil, email: "", encrypted_password: "", reset_password_token: nil, remember_created_at: nil, sign_in_count: 0, current_sign_in_at: nil, last_sign_in_at: nil, current_sign_in_ip: nil, last_sign_in_ip: nil, confirmation_token: nil, confirmed_at: nil, confirmation_sent_at: nil, created_at: nil, updated_at: nil> 
 > u.save
 => false 
 > u.errors.full_messages
 => ["Email can't be blank", "Password can't be blank"]

Testing Model Associations in Rspec (Rails 3)

2011-04-14T00:00:00-07:00

Testing Model Associations in Rspec (Rails 3)

I live with five really awesome housemates. Five warm bodies tend to make a mess by the end of the week so we have a chore chart on the fridge. I tend to be the one who badgers everyone else into doing their weekly chore - not my favorite thing to do...so, I figure it's time to put Rails to work sending those emails for me.

I named our fridge chore chart replacement Choretle. I'm excited to get it finished so that it can do all the dirty work of emailing reminders and doling out appropriate praise for a job well done.

I've got my data model pretty much sorted out at this point and I just started generating my models. I decided to do some googling about testing model associations since it has never been as streamlined as I'd like. I found a couple of ways to test associations, some of them better than others.

The way I've tested associations in the past has gone something like this:

describe Group do
  it "should respond to users" do
    g = Group.new
    g.should respond_to(:users)
  end
end

What I don't like about it is that it is a round-about way of testing for an association. It tests a side effect of the association, not that the association exists.

The next solution I found is built into ActiveRecord::Reflection::AssociationReflection, so no need to install extra gems:

>  g = Group.reflect_on_association(:users)
 => #<ActiveRecord::Reflection::ThroughReflection:0x00000100d65208 @macro=:has_many, @name=:users, @options={:through=>:affiliations, :extend=>[]}, @active_record=Group(id: integer, name: string, creator: integer, created_at: datetime, updated_at: datetime), @collection=true> 
> g.macro
 => :has_many

The macro method makes it fairly easy to write an association test:

describe Group do
  it "should have many users" do
    g = Group.reflect_on_association(:users)
    g.macro.should == :has_many
  end
end

It is still kinda ugly don't you think? Definitely not an elegant way to indicate the relationship. Luckily, I found a couple of recommendations for gems that make for more descriptive syntax: remarkable and shoulda. Both gems have similar syntax:

#both remarkable and shoulda use this syntax
    describe Group do
      it { should belong_to(:supergroup) }
      it { should have_many(:users) }
    end

Remarkable seems cool, but from what I could tell, it isn't well maintained. The last commit is from about a year ago on the carlosbrando fork.

I decided to use Shoulda for my tests. It's easy to get working - just add gem 'shoulda-matchers' to the Gemfile.

Here is my (short and symantically relevent) model association test:

  describe Group do
    it { should have_many(:users)}
  end

In my research, I encountered some debate about whether testing associations is worthwhile in the first place. Some people argue that it is testing Rails instead of your own code. I'd argue that testing associations isn't testing Rails. It is testing that your code is hooked up the way you want it. Testing associations also gets me into the TDD frame of mind as soon as I start a project. Plus, when Shoulda makes it so easy, I don't have any excuses not to test associations.

NERDTree for Macvim - Everything I've Learned So Far

2011-04-07T00:00:00-07:00

NERDTree for Macvim - Everything I've Learned So Far

To toggle the vertical NERDTree window on and off, use :NERDTreeToggle
To open a tab and move into it use j or k to navigate the file you want to open, then hit t from the NERDTree window
To open a tab but stay in the current tab, use j or k to navigate the file you want to open, then hit T from the NERDTree Window
To move a file in the NERDTree window, use m
To add a file in the NERDTree window, use a
To navigate to the next tab, use gt
To navigate to the previous tab, use gT
All window commands start with ctrl w
To move back to the NERDTree window from another split window in the current tab, use ctrl ww or ctrl wh
To move between windows in general, ctrl wh takes you left, ctrl wl takes you right, ctrl wj takes you down and ctrl wk takes you up.
To quit a window use ctrl wq

Scheme - Generalizing Functions

2011-04-03T00:00:00-07:00

Scheme - Generalizing Functions

These are notes I took while watching the third lecture of Brian Harvey's 61A CS Class

Topic: Generalizing Functions

The code below can also be found on the Berkeley website.

First we saw how to generalize our sumsquare function so that it can also work for sumcube:

(define (sumsquare a b)
  (if (> a b)
      0
      (+ (* a a) (sumsquare (+ a 1) b)) ))

(define (sumcube a b)
  (if (> a b)
      0
      (+ (* a a a) (sumcube (+ a 1) b)) ))

(define (sum fn a b)
  (if (> a b)
      0
      (+ (fn a) (sum fn (+ a 1) b))))

(define (square x)
  (* x x))

(define (cube x)
  (* x x x))

Next we generalized functions that perform specific tasks such as returning a list of the even numbers from a list of both even and odd numbers:

(define (evens nums)
  (cond ((empty? nums) '())
        ((= (remainder (first nums) 2) 0)
         (se (first nums) (evens (bf nums))) )
        (else (evens (bf nums))) ))

This function returns a list of words that have an 'e':

(define (ewords sent)
  (cond ((empty? sent) '())
        ((member? 'e (first sent))
         (se (first sent) (ewords (bf sent))) )
        (else (ewords (bf sent))) ))

This function returns the pronouns of a sentence:

(define (pronouns sent)
  (cond ((empty? sent) '())
        ((member? (first sent) '(I me you he she it him her we us they them))
         (se (first sent) (pronouns (bf sent))) )
        (else (pronouns (bf sent))) ))

This function generalizes the above functions so that we can use just one function to perform the same behavior as the above three functions:

(define (filter pred sent)
  (cond((empty? sent) '())
       ((pred (first sent))(se (first sent)(filter pred (bf sent))))
       (else (filter pred (bf sent)))))

The most important word in computer science is Abstraction. The second and third most important words are Domain and Range.

Domain: What kind of arguments does it take

Example: domain of = is numbers whereas domain of equal? is anything

Range: Set of possible values a function can return

Example: predicate is a function whose range is true or false

Higher Order Function: A function that has a function as an argument and/or a function as a return value

First Class Data Type:

can be the value of a variable
can be an argument to a procedure
can be a value returned by a procedure
can be an element of a data aggregate

Data aggregate is like an array or a list, etc.

The way we define functions...like this:

(define square (x) (* x x))

is actually shorthand for this:

(define square(lambda (x) (* x x)))

The lambda is added in by the Scheme interpreter

Unix - find /path_to_search -print | grep filename

2011-03-30T00:00:00-07:00

Unix - find /path_to_search -print | grep filename

I was using ls -R | grep filename trying to find a particular file deep within the font bowls of Tex. It found the file just fine, but then I needed to know where the file lives. Here is the magic code: $ find /path_to_search -print | grep filename

Scheme - Sum of Biggest Squares, Lambda Calculus (Noisebridge Notes)

2011-03-27T00:00:00-07:00

Scheme - Sum of Biggest Squares, Lambda Calculus (Noisebridge Notes)

I just got home from the Noisebridge Schemers meetup. We started off talking about problem 1.2 (1985 version) / 1.3 (1996 version). Here's the problem:

Define a procedure that takes three numbers as arguments and returns the sum of the squares of the two larger numbers

Here is our first implementation:

(define ssq2(a, b, c)
  (and (a<=b) (a<=c))(sum_of_squares b c)
  (and (b<=a) (b<=c))(sum_of_squares a c)
  (and (c<=a) (c<=b))(sum_of_squares a b)
)

We talked about how writing this function in scheme involves thinking in a different way than if we were to write it in Ruby. In Ruby we might sort the arguments using a sort method, then eliminate the smallest one. Ruby gives us sort and min/max methods for free. In Scheme it isn't quite so easy. In Scheme we think about the problem more in terms of a tree whereas in Ruby we manipulate our data in memory using methods. To illustrate this point, we considered the following code which is how we might think about the problem in a Rubyish way. It assumes that there is a sort method that takes a greater than or less than symbol to determine whether to sort ascending or descending.

(define(ssq2 a b c)
  (+ (* (car(sort(list a b c) >)
        (car(sort(list a b c) >)))
      (* car(cdr(sort(list a b c) >)))
         (car(cdr(sort(list a b c) >)))))

Here is my implementation in Ruby using a Scheme-like design.

def sum_square_two_biggest(a,b,c)
  if a<=b && a<=c
    sum_of_squares b c
  elsif b<=c && b<=a
    sum_of_squares c a
  else
    sum_of_squares a b
  end 
end

This is an asci art tree that demonstrates how a Schemer might think about the problem. The left branches are the result of the node being true.

                a<b
              /      \
            /           \
          |                |   
        b<c                  a<c     
      |     |              |       |
      |     a<c            |       c<b
      |     |    \         |     |      |      
   a<b<c   a<c<b   c<a<b  b<a<c   c<b<a   b<c<a

We talked about functional programing and how using set! breaks functional programming. Don't use set! if you want to use applicative evaluation as it forces evaluation immediately. On the other hand, using define expands into a lambda expression and does not break functional programming rules.

Lambda example:

((lambda(n) (+ n 1)) 41)

=> 42

Lambda calculus example:

(λx|x) a
=> a

(λx|x)(λy|yy)
=> (λy|yy)
(λy|yy)(λx|x)
=> (λx|x)(λx|x)
=> (λx|x)

Book recommendation for lambda calculus: algorithmics by david harel

Scheme - Functional Programming, Applicative Order, Normal Order

2011-03-20T00:00:00-07:00

Scheme - Functional Programming, Applicative Order, Normal Order

These are notes I took while watching the first lecture of Brian Harvey's 61A CS Class

Functional programming: Output will always be the same if the input is the same - this is helpful in ensuring consistency when we run the same program on different processors (parallel processing)

Applicative order evaluation: All the arguments to Scheme procedures are evaluated when the procedure is applied in other words, arguments are evaluated before procedures

Normal order evaluation: Evaluation of procedure arguments is delayed until the actual argument values are needed. Delaying evaluation of procedure arguments until the last possible moment (e.g., until they are required by a primitive operation) is called lazy evaluation.

You are only assured of getting the right answer with both applicative and normal order evaluation when using functional programming. The following example illustrates how if we don't follow the rules of functional programming (same output for same input), we may get different results for normal vs applicative evaluation.

def zero(x)
    x - x
end

Applicative: will always give zero because the argument is evaluated before the procedure is applied:

zero(10.rand)

Normal: will give varying answers because the entire 10.rand expression is substituted for each of the x’s in the procedure:

zero(10.rand)

In this case by using rand we are breaking the rules of functional programming. rand is not a function since a function by definition gives the same outputs for the same inputs. Since it isn’t a function and we aren’t doing functional programming, the order in which things happen matters.

Moral of the story: if you are doing functional programming, order doesn’t matter.

Efficiency can also be different for applicative vs. normal

Efficiency difference using the square function:

def square(x)
    x*2
end

applicative:

square(2+3)

2+3 is evaluated once

normal:

square(2+3)

2+3 is evaluated four times

There is a way to do things so that normal order comes out to be more efficient, but in this case applicative was more efficient. In any case, the point is that you need to think about the order in which operations are performed.

Bundle Config & Bundle Open

2011-02-25T00:00:00-08:00

Bundle Config & Bundle Open

I wanted to be able to open up gems easily from the command line. I tried

 
bundle open unicorn

and got the following error:

To open a bundled gem, set $EDITOR or $BUNDLER_EDITOR

To get rid of this error I added the following to .profile:

 
export BUNDLER_EDITOR=mate

This causes TextMate to open the gem of choice when I call:

 
bundle open unicorn

Escaping and Unescaping in HAML

2011-02-04T00:00:00-08:00

Escaping and Unescaping in HAML

To turn escaped HTML into HAML, use != instead of just = To escape unescaped HTML, use &= instead of just =.

> Haml::Engine.new('!= "<p> Hello World</p>"').render<br>
=> "<p> Hello World</p>\n" <br>
> Haml::Engine.new('%p Hello World').render<br>
=> "<p>Hello World</p>\n" <br>
> Haml::Engine.new('= "<p> Hello World</p>"').render<br>
=> "<p> Hello World</p>\n" <br>

Resources:

Unescaping HTML
Escaping HTML

redirect_to vs. render - Rails

2011-02-02T00:00:00-08:00

redirect_to vs. render - Rails

redirect_to:

redirect_to causes the browser to receive a 302.
The browser makes a brand new request and you will need to supply the browser with the information it needs to make the request such as the object attributes.
Any existing variables will be lost.
redirect_to is the right choice when you are posting a completely filled out form with valid data.

render:

render causes the browser to return a 200.
The browser does not make a new request and instead just renders the page you specify without sending any information to it.
The code in the rendered action will not be executed, it simply returns the view
render is the right choice when a form was incorrectly filled out and you want to display errors on the page you were just on as well as saving state of all the data that had already been entered into the form.

A link that helped me:
http://www.helloworlder.com/?p=6

Remove Files Older Than a Particular Date From a Directory in Unix

2010-12-15T00:00:00-08:00

Remove Files Older Than a Particular Date From a Directory in Unix

I have a folder filled with hundreds of images. Most of the images were first added to the folder on May 19th. The folder I'm moving them to already has all the images that were added on May 19th, so to save time, I don't want to move the May 19th images. I made a back up of my images folder that and am using the below commands within it. I only want to keep images added after May 19th in the folder so that I can zip them up and move them to another server. I'm using a dummy folder I made locally to test this out. Here are the unix commands I will use:

[Liah-Hansens-MacBook-Pro:~/test]$ ls -al
total 0
drwxr-xr-x   6 liahhansen  staff   204 Dec 15 15:01 ./
drwxr-xr-x+ 75 liahhansen  staff  2550 Dec 14 19:18 ../
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 goodbye
drwxr-xr-x   2 liahhansen  staff    68 Dec 14 19:30 happy/
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 hello
-rw-r--r--   1 liahhansen  staff     0 May 24  2010 timefilealittleolder
[Liah-Hansens-MacBook-Pro:~/test]$ touch -am 05192010 timefile

[Liah-Hansens-MacBook-Pro:~/xargs_test]$ ls -al
total 0
drwxr-xr-x   7 liahhansen  staff   238 Dec 15 15:11 ./
drwxr-xr-x+ 75 liahhansen  staff  2550 Dec 14 19:18 ../
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 goodbye
drwxr-xr-x   2 liahhansen  staff    68 Dec 14 19:30 happy/
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 hello
-rw-r--r--   1 liahhansen  staff     0 May 19  2010 timefile
-rw-r--r--   1 liahhansen  staff     0 May 24  2010 timefilealittleolder

The important part:

[Liah-Hansens-MacBook-Pro:~/test]$ find . -mtime +208 -type f | xargs rm

Now timefile is gone. Yay!

[Liah-Hansens-MacBook-Pro:~/xargs_test]$ ls -al
total 0
drwxr-xr-x   6 liahhansen  staff   204 Dec 15 15:11 ./
drwxr-xr-x+ 75 liahhansen  staff  2550 Dec 14 19:18 ../
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 goodbye
drwxr-xr-x   2 liahhansen  staff    68 Dec 14 19:30 happy/
-rw-r--r--   1 liahhansen  staff     0 Dec 15 13:58 hello
-rw-r--r--   1 liahhansen  staff     0 May 24  2010 timefilealittleolder

Yay!

When I performed these commands on the real files, I ran into a problem where many of the file names had whitespace in them. This caused rm to fail and I still had many files left in the folder from May 19th. Here is the command that got rid of them once and for all:

 
find . -mtime +208 -type f | xargs -I{} rm '{}'

Checking just to make sure:

 
ls -al |grep "May 19" | wc -l
0

Golden!

First Program in C

2010-10-04T00:00:00-07:00

First Program in C

Now that I've got Ruby and Rails fairly well learned and practiced, I decided it was time to start working towards learning some lower level languages. Here is my first baby step towards learning C:

touch foo.c

mate foo.c`

In foo.c I wrote (with the help of my roommate Sean):

#include <stdio.h>
int main() {
    int i;
    printf("Hello World!!!!!!\n");
    for(i=0; i<5; i++) {
        printf("%d\n", i);
    }
    return 0;
}

Next, I compiled my program:

gcc foo.c

This created a file called a.out

Next I ran a.out

./a.out

And it out put:

Hello World!!!!!!
0
1
2
3
4

Sean also taught me how to define the name of the out file:

gcc -o foo foo.c

This let me do the following to execute my program:

./foo

Open Government at GoGaRoCo

2010-09-22T00:00:00-07:00

Open Government at GoGaRoCo

Eric Mill gave an inspiring talk at Golden Gate Ruby Conference. Below are my notes about it.

Innovative websites written in Rails and based off of open government data:

http://datasf.org/showcase/
http://www.nyc.gov/html/datamine/html/home/home.shtml
http://data.dc.gov/
http://www.appsfordemocracy.org/
http://www.data.gov/
http://flyontime.us/
http://usaspending.gov/
http://sunlightfoundation.com/clearspending/
http://www.thomas.gov/
http://www.govtrack.us/ -- scrapes thomas
http://www.opencongress.org/ -- built off of govtrack.us, rails app
- inspired congress to publish votes in xml
- senate also puts out info in xml
- no where near all 50 states give out info
sunlight is trying to make state apis the norm
- mongo db is necessary for it - like some states have only one house, not two

Federal Register - one of the more important publications of the government

rulemaking
public notices
chances to comment
published by the office of the public register

http://www.govpulse.us/ was one result of a contest - open source rails app on github, EC2

it shows what is going on in your area
government contacted them and contracted the develpers
ditched some red tape
pushedthem to take risks
buy-in from the top of the agency
made something happen in 3 months...much faster than most gov projects
they cut their hours by 40% at regular jobs
result of their work: http://www.federalregister.gov/

Gov is seeing devs as their customers

America

first country to have a freedom of information act
all govs info is in the public domain
we have the freedom to set amazing examples - what we do matters - talk to other people

2010 census - machine readable, but have to get through 300 pages of pdf instructions - people have written ruby libraries around the data.

Interesting Questions Students Asked Me

2010-09-22T00:00:00-07:00

Interesting Questions Students Asked Me

How Ruby language is structured...Kernal is mixed into Object. What impact does that have (sleep, method_missing, puts)? What about Enumerable and the classes it is mixed into...they all have '#each' but how does it work that the methods defined in Enumerable can work for all the different classes?
Threads in Ruby
Scaling in Ruby
DSLs and metaprogramming
Connection to the database...how is this done internally...does it have to do with IO
IO
Procs & Lambdas...what are some real world applications of them.
what happens when two file objects are created with the same file and try to update the file at the same time
VMs - is Ruby 1.8.6 the 'main' Ruby...are JRuby, Groovy and other specialized VMs using 'main' Ruby and altering it...or are they rewriting Ruby in a different language...what does virtual machine even really mean in this context...
regular expressions
duck typing vs dynamically typed

Voicenotes - My First Twilio App

2010-07-25T00:00:00-07:00

Voicenotes - My First Twilio App

Sometimes I wonder about my sanity when I voluntarily spend a sunny Saturday in a basement writing code. Yesterday was one of those days at the Social App Workshop held at Heroku's new headquarters. The morning was an action packed string of talks from engineers at Twitter, Twilio, Facebook, Apigee and Heroku. I was really impressed with Jeff's talk about Twilio and thought I'd try making an app with it's api.

The results of my efforts are here: http://voicenotes.heroku.com/voice_to_text

My goal was to write an app that you could call, record a message, and the message would be transcribed and put on a webpage. The transcribe feature is pretty inaccurate. I left a voicenote of 'Remember to take out the trash' and it transcribed it as 'Remember to take out the plan'. It got 'two laundry' when I said 'do laundry'. But still, Twilio is a very cool idea and it got my creative juices flowing.

Twilio & Rails - 502 Bad Gateway Error Solution

2010-07-25T00:00:00-07:00

Twilio & Rails - 502 Bad Gateway Error Solution

I ran into an issue while writing a Twilio app in Rails where Twilio kept getting a 502 bad gateway error. The url I was using was definitely working, so that wasn't the issue. I did a little google sleuthing and the answer is putting the following code into the controller:

skip_before_filter :verify_authenticity_token

Push to Heroku from a Branch of Your Local Git Repository

2010-02-25T00:00:00-08:00

Push to Heroku from a Branch of Your Local Git Repository

We spent some time searching for a bit of git hocus pocus to solve this one...here's the spell:

git push heroku <branch_name>:master

Our situation:

A production server
A staging server
A sandbox server

We have a branch in our git repo for the sandbox server & needed to be able to push straight from that, not from master. The little colon did the trick!

A couple resources:

Git Workshop Retrospective

2010-02-25T00:00:00-08:00

Git Workshop Retrospective

I was elated with how well the workshop was received. All 22 open spots were taken on meetup.com and we had 100% of those people show up! There were still around 12 people on the waiting list. Five women participated (including me), making the gender balance 23% female...not too shabby! Everyone was snug in our small space at Blazing Cloud, but it worked.

I think that the hands on approach was very effective. First I had everyone pair off and work together to research one of the following:

when to use cherry-pick
when to use checkout (branch & files)
when to use merge
when to use reset
when to use revert
when to use rebase
when to use branch vs stash
how to ignore files
find as many ways to add & commit as you can (all the options & flags, what they do)
how to delete files properly

Then each pair reported back with their findings. The discussion was great, made even better when Brian Colfer, my co-organizer & git expert came by and explained complex git concepts articulately.

We switched gears and each pair completed an exercise I prepared. All the materials for the exercise are online at gitworkshop.com. The scenario included a boss who barked out indecisive orders that gave us opportunities to git reset, git revert, git stash, git cherry-pick, git merge and git branch. I learned a ton creating the tutorial (spent at least 12 hours doing it...plenty of time to get a good grasp of git).

My impression was that the participants learned a lot and had a good time!

Git Workshop

2010-02-20T00:00:00-08:00

Git Workshop

I've been using git since I started learning Ruby in July 2009. I've gone through many of the more advanced topics - creating branches, checking them out & merging them is clear to me...that is until it isn't clear. Sometimes git does things I'm not expecting. So, I figured I'd host a little hands-on workshop to try to work out git intricacies. There is nothing like teaching a topic to learn it inside & out.

More info about the workshop...

Understanding How Ruby Blocks Work

2010-01-28T00:00:00-08:00

Understanding How Ruby Blocks Work

I've been preparing the curriculum for a class about Ruby blocks. The word "block" just means a block of code that gets passed to a method right after other parameters. When a block is sent to a method, the code within it is then available to the method. The method can call the block of code any time it wants to using the "yield" keyword. To sum it up, the two important pieces of blocks, are:

the code block

the yield keyword

Look below...first I write my method, dos_veces. When I call dos_veces, I add some code to the end of it between curly braces {...}. That code gets executed by the method whenever I typed "yield" within my method definition.

I've found a few extremely helpful websites including: Eli Bendersky's blog post about blocks procs and methods, Robert Sosinski's blog post about blocks, procs and lamdas, Ruby Learning: Blocks Tutorial, and Jamis Buck's blocks rock post.

Double Colon

2009-11-07T00:00:00-08:00

Double Colon

I'm reading Refactoring: Ruby Edition and I came across the double colon (::) operator. I pretty much understand what it means, but I thought I'd do a little more research to get a better handle on its intricacies. I came across the code below at http://en.wikibooks.org/wiki/Ruby_by_examples#::

CONST = ' out there'

class Inside_one
    CONST = proc {' in there'}
    def where_is_my_CONST
       ::CONST + ' inside one'
    end
 end
 
 class Inside_two
    CONST = ' inside two'
    def where_is_my_CONST
       CONST
    end
 end

I thought I'd try it out in irb. Here is the output:

>> puts Inside_one.new.where_is_my_CONST
 out there inside one

>> puts Inside_two.new.where_is_my_CONST
 inside two

>> puts Object::CONST + Inside_two::CONST
 out there inside two

>> puts Inside_two::CONST + CONST
 inside two out there

>> puts Inside_one::CONST
#<Proc:0x00000001012ac4a0@(irb):3>

>> puts Inside_one::CONST.call + Inside_two::CONST
 in there inside two

Search and Product Listing in Zen Cart

2009-11-04T00:00:00-08:00

Search and Product Listing in Zen Cart

I've been working out some bugs in the search functionality that arose from some customization of the product listing. The files I've been editing are scattered across many different folders and I thought it would be helpful to list all the files and what is in them.

The heart of the product listing for both searches and category displays is held in a variable called $listing_sql. There are two files where $listing_sql is built:

includes/index_filters/default_filter.php
includes/modules/pages/advanced_search_result/header_php.php

These files contain conditional statements that determine what SQL query should be executed.

Both files send $listing_sql to:

includes/modules/product_listing.php

This file processes the SQL output into strings of html.

Symbol to Proc

2009-11-03T00:00:00-08:00

Symbol to Proc

I was tri-programing with S & B today, writing Selenium tests. S typed:

People.all.map(&:name)

into irb to demonstrate a database call. I've seen this syntax before, but never understood what it meant. He said he didn't know, but that it worked. The output is an array of the names of all the People:

["Sandra", "Bob", "Fred", "Rosey"]

B explained that the ampersand followed by a symbol is a concise syntax for mapping the symbol object to a block.

Here is a bit of playing with irb that illustrates how Symbol#to_proc works:

>> People.all
=> [#<People id: 2, name: "Sandra">, #<People id: 3, name: "Bob">, #<People id: 4,  name: "Fred">, #<People id: 5,  name: "Rosey">]

>> People.all.map(&:name)
=> ["Sandra", "Bob", "Fred", "Rosey"]

>> People.all.collect { | t | t.name }
=> ["Sandra", "Bob", "Fred", "Rosey"]

>> :name.to_proc
=> #<Proc:0x00002b720b0f45f8@/usr/local/lib/ruby/gems/1.8/gems/activesupport-2.3.4/lib/active_support/core_ext/symbol.rb:11>

>> mything = proc { | t | t.name }
=> #<Proc:0x00002b720f91e950@/home/heroku_rack/lib/console.rb:133>

>> People.all.map(&mything)
=> ["Sandra", "Bob", "Fred", "Rosey"]

>> People.create!(:name => "Ben")
=> #<People id: 9, created_at: "2009-11-02 13:25:55", updated_at:
"2009-11-02 13:25:55", name: "Ben", creator_id: nil>

>> People.all.map(&mything)
=> ["Sandra", "Bob", "Fred", "Rosey", "Ben"]

In a nutshell, the following two lines of code have the same result:

People.all.map(&:name)
People.all.map( | t | t.name)

Let the Blogging Begin!

2009-09-06T00:00:00-07:00

Let the Blogging Begin!

UPDATE: As of March 2011 my blog has migrated from Radiant to Jekyll.

I've been attending some amazing Ruby & RoR workshops and classes for the last month. The whole time I've been wanting to blog about all of my breakthroughs with programming. One of my RoR extracurriculars has been an informal meeting of 12-15 newbies every Sunday from 1-5 in a cottage belonging to our RoR teacher, Alex Chaffee. We've built a humble - very, very humble blog app. I was hoping to get it ready for prime time sooner and use it for my personal blog. After three weeks we finally implemented authentication, but it still lacks in many areas, so I thought I'd use Radiant in the meantime. It took me all of 10 minutes to install Radiant! It is simple - much more barebones than Wordpress or Drupal, but it gets the job done while giving me a chance to watch RoR in action. Here are all 12 lines of commands it takes to install:

radiant -d sqlite3 liah-hansen
cd liah-hansen/
rake db:bootstrap
git init
heroku create
heroku addons:add custom_domains
heroku domains:add liahhansen.com
heroku domains:add www.liahhansen.com
git add .
git commit -m "initial commit"
git push heroku master
heroku db:push

Ta Da!

Plus I changed this line in environment.rb:

config.middleware.use ::Radiant::Cache

to this:

config.middleware.use ::Radiant::Cache,
:entitystore => "radiant:tmp/cache/entity",
:metastore => "radiant:tmp/cache/meta"

and I added a .gems file to the root directory:

rspec --version 1.2.8
radiant --version 0.8.1