BioNuc Technical

MySQL GROUP_CONCAT function

2011-07-07T05:16:00.000-07:00

I knew about this MySQL function called "GROUP_CONCAT" lately and I think it is really a very useful function to use

To make simple for you to understand, I will use a small example to illustrate on

Consider we have a table called listings which has many to many relationship with another table called geographies knowing that the join table name is listings_geographies

Now look at this query

select listings.id, listings.display_name, geographies.name as geography_name from listings LEFT OUTER JOIN listings_geographies ON listings.id = listings_geographies.listing_id LEFT OUTER JOIN geographies ON listings_geographies.geography_id = geographies.id WHERE status = 3 GROUP BY listings.id

In my example I have only two listings: the first has 3 geographies and the second has 2 geographies. By running this command, you will get 2 records but the field called "geography_name" will have one of the three values and not all the values

What if you want to grab all the three values and add them in 1 field so that you get 2 records but with "geography_name" field containing all geographies names values. If you want that, then you can rely on this nice function "GROUP_CONCAT"

The above function can be re-written to be this one

select listings.id, listings.display_name, GROUP_CONCAT(geographies.name SEPARATOR ' ') as geography_name from listings LEFT OUTER JOIN listings_geographies ON listings.id = listings_geographies.listing_id LEFT OUTER JOIN geographies ON listings_geographies.geography_id = geographies.id WHERE status = 3 GROUP BY listings.id

That's it
I needed this function as I was using Sphinx and I wanted to have a query written to grab all record info with its associations in 1 record in order to be indexed

Enjoy! :)

“Distinct vs Group By” SQL Talk

2010-12-29T01:44:00.001-08:00

I am writing this blog entry to tell people about main differences between these two instructions because any misunderstanding to the true difference between them can cause big problems

That what happened to me actually
I didn’t not know the true difference and as always I have chosen the easy way and rely on Distinct instead of Group By

Let me show an example which I was working on
and that one really helped me understand the difference between these two

Suppose you have a Deal which is associated with several geographies
Each geography has a level representing its level in the tree of geographies. The higher the level, the lower the geography exists in the tree.

The objective was to get all deals in several geographies which I was given their IDS and sort them by depth/level from lowest to highest

I started writing my query to be like that
which is wrong by the way

select Distinct(deals.id) from deals INNER JOIN deals_geographies ON deals.id = deals_geographies.deal_id INNER JOIN geographies ON geographies.id = deals_geographies.geography_id WHERE geographies.id IN (id1, id2, id3) ORDER BY geographies.level DESC

it sounds good at the beginning. If you see it right then you are facing the same problem I was facing before. So please continue reading

The problem here is that inner joining will perform a Cartesian product between deals and geographies and as we have several geographies for each deal, we will get several rows for each deal

When you say DISTINCT without defining the criteria to the DB engine, it will choose any row which could get you the highest geography this deal has or it can get you the lowest geography this deal has

This is an Ambiguous Selection
but it is your fault that you did it that way

Lets see now how group by will solve this

select deals.*, MAX(geographies.level) as max_level from deals INNER JOIN deals_geographies ON deals.id = deals_geographies.deal_id INNER JOIN geographies ON geographies.id = deals_geographies.geography_id WHERE geographies.id IN (id1, id2, id3) ORDER BY max_level DESC group by deals.id

Now I told the DB engine what should the criteria it should use to remove multiple deal rows and which row it should leave which I wanted it to be the one with the largest geography level as this is my desired condition

My last conclusion is that you can use DISTINCT in one of these cases only

you have a 1 to 1 relation between two or several tables and thus no ambiguity will be there
you only care about 1 table fields and you are just using the other table for no more but Filtration which was not the case above as I needed it for ordering

btw, there is a rumor I heard saying that DISTINCT is not standard SQL. Not sure of this info but you can check that yourself and maybe comment and tell me

Hope this blog was useful and enlightening :)

Useful RVM resources

2010-10-05T17:21:00.001-07:00

I think these few resources are useful for any RVM newbie as me :)

Read them in that order

For me, I made use of the first and third links although I see all of them useful but these ones helped me install and use several ruby versions with several gemsets which is why I used RVM at first place

Hope I was useful
and happy RVM journey

Installing RVM System Wide

2010-10-05T17:02:00.001-07:00

Today I got introduced to one very nice solution developed by Wayne E. Seguin. It is called RVM and it is recommended for ruby developers working on Debian machines.

What is RVM ?

Simply RVM allows users to install multiple ruby versions and switch between them easily. It also allows having multiple sets of gems for different projects easily.

Installation Steps

Before getting into details, I would like to say that any steps mentioned here are grabbed from these two links. I only collected parts from them and added them in an easy way

http://rvm.beginrescueend.com/rvm/install/
http://rvm.beginrescueend.com/deployment/system-wide/
http://rvm.beginrescueend.com/rubies/installing/

from your linux console run
bash < <( curl -L http://bit.ly/rvm-install-system-wide )
you now have rvm command installed here at /usr/local/lib/rvm
After installation you should add this line below to your profile
[[ -s "/usr/local/lib/rvm" ]] && . "/usr/local/lib/rvm"
To do so, i added the line above at the end of the /etc/profile file. But, you can add it in other places according to your needs. Read more about profiles here
then run this one
source /usr/local/lib/rvm

Now you are ready to install any versions of Ruby and Ruby Enterprise versions available

For me, I installed Ruby 1.8.7 and its enterprise version this way

rvm install 1.8.7
rvm install ree-1.8.7

That’s it
Hope you enjoy RVM as I hope I enjoy it as well

Forget root password of a Debian Machine & resetting it

2010-08-27T06:54:00.001-07:00

I am writing this small post just because

i forgot the root password of my debian machine
didn’t find a through post that list all steps in one place and had to look at several ones at a time

The steps are as follows

I am quoting these lines from that post

Some Linux distribution, such as Ubuntu for instance, offer a specific boot menu entry where it is stated "Recovery Mode" or "Single-User Mode". If this is your case, selecting this menu entry will boot your machine into single user mode, you can carry on with the next part. If not, you might want to read this part.

Using GRUB, you can manually edit the proposed menu entry at boot time. To do so, when GRUB is presenting the menu list (you might need to press ESC first), follow those instructions:

use the arrows to select the boot entry you want to modify.

press e to edit the entry

use the arrows to go to kernel line

press e to edit this entry

at the end of the line add the word single

press ESC to go back to the parent menu

press b to boot this kernel

The kernel should be booting as usual (except for the graphical splash screen you might be used to), and you will finally get a root prompt (sh#).

Here we are, we have gained root access to the filesystem, let's finally change the password.

According to the above words, we should be ok and we have access to the file system. At this state you run the command “passwd” and enter the new password.

If it worked with you then Thanks to the editor. If you got some problems like me, then keep reading.

Problem 1

After editing grub line and add the “single” keyword at the end of that line. I got it loading well until i was prompted for the root password for maintenance and give the ability to skip but by then i will be leaving runlevel 1 and entering runlevel 2 getting login/password prompt i am trying to skip.

To solve this problem do the following:

Edit the grub line again and leave the keyword “single” there as before but add at the end as well these words “init=/bin/bash”
exit edit mode
press the button “b” while you have this modified grub line highlighted to start booting with this modified grub line

Voila, you have now access to shell as a root. run the command “passwd” and you should be fine entering the new desired password. If you got a problem, then continue reading.

Problem 2

Whenever i run the command “passwd” and re-enter the new password i get this error at the end

authentication token lock busy

If so, know that the problem is that you are accessing the system in read-only mode. In order to access it in read-write mode, do the following.

From the shell run this command

mount -o remount,rw /

after that run “passwd” command and this time you should have the ability to enter the new password

and live happily ever after

Factory Girl in Development

2010-08-15T03:35:00.001-07:00

Some many developers use Factory Girl as a replacement to Fixtures and they depend on this gem in Test cases writing.

But Factory Girl can be very useful also in development and can be used in order to generate a bunch of dummy data with different specs for showing your work and reviewing all its details.

The problem we face is that Factory Girl can generate more records in tables that should have their data unchanged such as Countries table.

For example:

Factory.define :user do |f|
f.association :country
end

Factory.define :job do |f|
…
f.association :user
end

Now, when we run this piece of code that generates 10 jobs

10.times { Factory.create(:job) }

we will get 10 countries auto-generated
violating the rule we wish to maintain which is having the countries table as it is

To workaround this problem without causing any changes in the code written previously, i came up with that solution

class Factory
class << self
    alias_method :create_original, :create

    def create(name, overrides = {})
      if name.to_s == 'country'
        country = Country.first
        return country if country
      end

      create_original(name, overrides)
    end
end
end

The logic introduced in the above script is simply adding a layer before creation that checks if the created object of certain type and if that type is desired to not be generated, we return the first entry we have in the DB else we do the normal creation

You can add this script in a file and load it in the development environment and use Factory Girl safely without fearing of generating data for tables that should remain as lookup tables

Enjoy :)

Factory Girl & Polymorphic Associations

2010-08-09T06:08:00.001-07:00

Suppose you have a case like that

class Address < ActiveRecord::Base
belongs_to :addressable, polymorphic => true
end

class Listing < ActiveRecord::Base
has_one :address, as => :addressable
end

class Customer < ActiveRecord::Base
has_one :address, as => :addressable
end

Now in your factories you will have something like that

Factory.define :listing do |f|
f.association :address
end

since addressable is required, we choose it to be by default related to a customer unless else stated

Factory.define :address do |f|
f.association :addressable, :factory => :customer
end

right now if you tried to use the Listing Factory, you will get an address associated with dummy customer and no effect on your side

How can this be fixed ?

Factory.define :listing do |f|
f.after_build do |listing|
listing.address = Factory.create(:address, :addressable => listing)
end
end

This was the only solution i found after searching for a while
which is a good solution and doesn't need a lot of work to be done

sanitize gem issue

2009-10-25T05:09:00.001-07:00

I have used ‘sanitize’ gem to remove HTML tags from entered text
the version i was using was ‘1.0.8’

After using it for a while, i found an issue in it where calling
Sanitize.clean(“’”) will return “#39;” while quote isn’t HTML character

Looking around for a reason for this issue, i found that this is a defect that was detected in older versions and that it is now fixed in version ‘1.1.0’

So anyone who has this issue can simply remove his installed gem and install the latest one

smtp 555 5.5.2 Syntax error (Net::SMTPFatalError)

2009-08-31T04:00:00.001-07:00

This title was part of an error message i got while dealing with action_mailer

i looked everywhere for a reason for this error message and found so many reasons but none of them matched with my case

So i liked to share why i got this error. Simply because i was adding no recipients to my email. I know it is very weird to send an email without having a recipient but getting this error message i much more weird and strange

/usr/lib/ruby/1.8/net/smtp.rb:930:in `check_response': 555 5.5.2 Syntax error. 24sm123817eyx.21 (Net::SMTPFatalError)
        from /usr/lib/ruby/1.8/net/smtp.rb:899:in `getok'
        from /usr/lib/ruby/1.8/net/smtp.rb:842:in `rcptto'
        from /usr/lib/ruby/1.8/net/smtp.rb:834:in `rcptto_list'
        from /usr/lib/ruby/1.8/net/smtp.rb:833:in `each'
        from /usr/lib/ruby/1.8/net/smtp.rb:833:in `rcptto_list'
        from /usr/lib/ruby/1.8/net/smtp.rb:654:in `sendmail'
        from /usr/lib/ruby/gems/1.8/gems/actionmailer-2.3.2/lib/action_mailer/base.rb:683:in `perform_delivery_smtp'
        from /usr/lib/ruby/1.8/net/smtp.rb:526:in `start'
        from /usr/lib/ruby/gems/1.8/gems/actionmailer-2.3.2/lib/action_mailer/base.rb:681:in `perform_delivery_smtp'
        from /usr/lib/ruby/gems/1.8/gems/actionmailer-2.3.2/lib/action_mailer/base.rb:523:in `__send__'
        from /usr/lib/ruby/gems/1.8/gems/actionmailer-2.3.2/lib/action_mailer/base.rb:523:in `deliver!'
        from /usr/lib/ruby/gems/1.8/gems/actionmailer-2.3.2/lib/action_mailer/base.rb:395:in `method_missing'
        from test_sending_emails_in_ruby.rb:31

this is the error message i got and it was solved when i added a recipient

Memory Leakage while using Mechanize

2009-06-29T02:00:00.001-07:00

I was working on a task that scrape several web pages. After running this task for a while, i found that memory taken by my process is raising forever until it was about to eat all memory available of the server.

after some investigation regarding this matter, i knew that the problem was in my understanding to how mechanize agent works

let me explain with an example

agent = WWW::Mechanize.new
while(true)
page = agent.get(“www.example.com”)
end

in this example, memory will be consumed because mechanize keeps history within the agent, i looked in its documentation and found that there is a parameter which is called “max_history” which when set will fix this issue i think but didn’t try

also a fix to such issue, if you don’t need history is to write your code like that

while(true)
agent = WWW::Mechanize.new
page = agent.get(“www.example.com”)
end

That’s it, maybe this piece of information can be useful for someone facing this issue just like me

Install Mechanize On Debian

2009-06-15T15:09:00.001-07:00

Installing mechanize gem on Debian should be as easy as running this command

gem install mechanize

but this won’t succeed unless you install these packages on your Debian machine

apt-get install libxml-dev libxslt1-dev

Once installed, gem will be installed seamlessly

Update:

if the above apt-get command didn't work with you and you got an error that package doesn't exist

try this new line

apt-get install libxml2-dev libxslt1-dev

as some packages names has changed

[Linux] Mounting windows folders

2009-06-15T03:21:00.001-07:00

Due to my new configuration which is using Windows as my default OS and Debian shell through Virtualbox, my need to have a folder shared between these two environment become a must in order to ease file sharing and exchange.

At first, i depended on “shared folders” feature of VirtualBox which after a while failed as i always get a “Protocol Error” whenever i deal with files IO on that shared folder.

That’s why i looked for another solution that is more stable than that one and reached to using “CIFS” as a way to mount windows shared folder on my Debian machine.

The steps are very easy to be made and gives you a stable robust solution away from VirtualBox problems. i can summarize these steps as follows

install on your Debian machine “smbfs” package which will add this new type of mounting called “CIFS”
suppose your machine IP is “192.168.1.6” and you shared a folder on it called “work” and this folder is secured to be used only with certain group which is administrators and one of these administrators is named “BioNuc” and have a certain password then your command will be mount -t cifs -o username=BioNuc ‘\\192.168.1.6\work’ <linux-path-to-mount-on-it>
After writing this line, you will be asked to insert your password in order to make mount process successful

That’s it, Enjoy sharing folders seamlessly

Giants Alliance

2009-04-13T01:02:00.001-07:00

Two giants are there one called Windows and another called Linux. As i am fan of Windows Interfaces and Usability (by the way i use Windows XP) and also fan of the extreme shell power of Linux, I wished if i can exploit these two powers at the same time.

The solution i found was using Windows XP as my primary OS and adding Debian version on virtual machine. You may be asking why should this solution be said when it is well known and made by some many users but actually, i made some modifications to this solution.

What i did can be summarized in these following points:

installing debian without any interface at all, just a shell no more
using virtual machine shared folder technique to share a folder between my XP machine and my Linux machine
installed ZOC program which is a powerful shell program that allows connecting using openssh and supports multi tabs which is very important feature not available in putty
installed openssh-server on debian
fixed an IP for my debian machine so as to save my openssh configuration one time only without changing it each time i connect to it

This way, i enjoy my XP machine and whenever i need the shell of linux. i connect using openssh on my debian virtualized machine and do what i want

I use this solution while developing ROR applications, i prepared the environment i use as follows

installed ruby, rails, mysql on my debian machine
opened remote connections to mysql database so as to use MySQL Query Browser from XP machine

Now, i open my IDE on XP machine that changes in folders place in the shared folder between two machines and i open my server from virtual machine and test apps on debian machine IP rather than my localhost

This experience is great and i enjoy it so much as i feel i have the power of each environment Windows & Linux at the same time

In order to optimize the performance of my Virtual Machine somehow, i raised the priority of these virtualized processes to ‘High’ or even ‘Realtime’. it is now better but not perfect

I think in order to raise the performance significantly, i need to buy another motherboard supporting Intel VT technology. Using such technology i can move virtualization from software layer (being a process that Windows XP Scheduler deal with) to hardware layer which should improve performance as stated

That’s it, i suggest you try this experience and Enjoy as i do

Export Google Document To WIKI STYLE Document

2009-04-12T06:23:00.001-07:00

the idea is simple and only needed is that you do some simple steps and you are done

follow the following steps:

save google document as HTML file
view html page source and extract the html inside document body tag
use this great tool and add in it the html you copied and you will have the wiki style document generated

That’s it, you now have the document you wrote in WIKI style. Take this text and play with it as you wish

Restore Your Deleted Partition

2009-04-11T23:52:00.001-07:00

Yesterday, i passed through a great panic after wrongly deleting my 3 partitions. Suddenly i felt that i lost all my data forever and there is no way out of such problem.

First how this problem occurred ?? actually this issue aroused from “computer management” program on my Windows XP. I opened it and choose one partition and right click and chosen delete. After confirmation, i found all 3 partitions were deleted. What a stupid application !!! i don’t know why should windows has an application that deal with such sensitive parts when not sure of its performance and reliability

Anyway, after some search on the internet, i found a great application called “TestDisk”. Anyone can download this free application from here

I did download it and followed this nice tutorial who showcase how this application works and how to use it safely. I did exactly as it stated and i restored my deleted partitions back again as before with nothing being lost

Thanks “TestDisk” for the great application and you really deserve a donation.

Gems vs Plugins

2009-04-09T14:46:00.000-07:00

Which one to use GEMS or PLUGINS?
This was a question i always asked and i have searched on the internet for comparisons between each one and the other

i can see that people prefer gems for the sake that it is the new future, it has obligatory versions other than rails plugins

But as for me, i only see gems useful in these two cases

when using code provided as gem outside of Rails, maybe with some ruby application
if you want to share code written in gem among applications

other than that, i see plugins better and much more maintainable for each application on its own

Uploading Images from URL using Paperclip

2009-03-09T05:36:00.000-07:00

I was working on an application that gives variety to user to either upload file locally or add url of this image on the web.

As i was using paperclip, i searched for a method that always such thing but didn't find such functionality available in paperclip although this is a very common task that can be found everywhere nowadays in web applications.

Anyway, i found a great blog post, that makes a workaround for such issue. the idea in this blog was sufficient enough to do all required althought i introduces new field for saving url which isn't needed all the time

I thought that this can be more easily for users by adding such thing in paperclip itself. I looked at the code and find it a very small task

the old code do that

def has_attached_file name, options = {}
.
.

define_method "#{name}=" do |file|
attachment_for(name).assign(file)
end

.
.
end

and patch adds this method, i took it as it is from blog post mentioned above and made this code as follows

def has_attached_file name, options = {}
.
.

define_method "#{name}_url=" do |file_url|
begin
file = open(URI.parse(file_url))
send("#{name}=", file)
rescue
#Not handled yet
end
end

define_method "#{name}=" do |file|
attachment_for(name).assign(file)
end

.
.
end

and also required a file 'open_uri', the new method have same name except that it contains at the end a suffix "_url"

it opens a connection and gets image as IO file and passes it to old method as before

I found this a useful change, that's why i forked paperclip and made the change in it and will commit this above and send pull request to include this feature

until then, the code can be found here

Cascaded HTML SELECT elements using JQuery V2

2009-03-09T01:34:00.000-07:00

This is a small update i made to the code, This code fixes a bug detected and also get rid of some duplicated code

his is an update to the JS function

eSpace.HTML = {
cascadeLists: function(parentId, childId, callbackFunction){
$("body").append('');
var childOptions = $('#' + childId + ' option');
$('#' + parentId + childId).html(childOptions);

$('#' + parentId).change(function(){
var parent = $('#' + parentId)[0];

var selectedOptionValue = '';
if (parent.selectedIndex > -1)
selectedOptionValue = parent.options[parent.selectedIndex].value;

if (selectedOptionValue == '')
$('#' + childId).html($('#' + parentId + childId + ' option[value=""]').clone());
else {
var childs = $('#' + parentId + childId + ' option[parent="'+selectedOptionValue+'"]');
if(childs.size() == 0)
$('#' + childId).html($('#' + parentId + childId + ' option[value=""]').clone());
else
$('#' + childId).html($('#' + parentId + childId + ' option[parent="' + selectedOptionValue + '"]').clone());
}

$('#' + childId).trigger("change");

if(callbackFunction != null) callbackFunction(parent.options[parent.selectedIndex]);
});

$('#' + parentId).trigger("change");
}
}

GMarks vs Foxmarks

2009-03-07T05:12:00.000-08:00

I have found some many friends of mine interested so much in GMarks extension to sync their bookmarks from one pc to another.

As i am also interested in bookmarks sync across computers, and as i also respect google extensions and was a frequent user of Google Browser Sync until they stopped supporting such extension. I said to myself that i have to see what this extension gives

Frankly speaking, using GMarks is very ugly, yes you bookmark and tag your bookmarks but the way of retrieving tagged bookmarks is very old fashioned. It takes us back to the era of Firefox 2 of adding bookmarks in folders and then whenever you look for something, you open your bookmarks and look for what you want. SO UGLY AND OLD FASHIONED way

To understand why GMarks is a bad choice, lets have a look at what Firefox 3 introduced which became a de facto and also is going to be there later on in other browsers even google chrome

With the arrival of Firefox3, we got rid completely of two main stupid things

searching for bookmarks in folders, they introduced tags
searching bookmarks then if not there, google using one of the search engines beside navigation bar

Firefox 3 focused on value, simply we should deal only with one entity whenever we open a new tab, this entity is the location bar

Since you write keywords on you search engine to get you results, and since now you tag your bookmarks with keywords then we all agree that it is only keywords that matter. The new scenario became open a new tab, you are already on the location bar, write the keywords you are searching for, Firefox will list you tagged bookmarks and some other results from history.

Satisified then fine take any like from this collection. if not satisified then simply hit the enter key and you will be taken to search results from the search engine you prefer

Very lovely way, focusing on one position and having everything in hand for you. let's see what this GMarks did for our lovely neat scenario

Now, if you want something, you go to Gmarks tab beside tools and you will find your tags, you search for the tag you want then you find your bookmarks there. Not satisfied then google and write again the same tags you were about to use. SIMPLE WASTING YOUR TIME AND DON'T EXPLOIT MULTIPLE TAGS SEARCH NOR HISTORY SEARCH

simply guys, it is very stupid. Please get rid of this ugly extension right now. you are brainwashed with it. Simply use Foxmarks

it don't interfere in your cycle at all. It just sync on shutdown or from time to time as you configure it and whenever you add it from any other browser on same pc or another pc you get your bookmarks ready for you

it also has this features

compatabile on many browsers, i think safari and opera
has web interface to use from anywhere
sync passwords also encrypted if you wish to
has a new fancy feature added which is suggesting tags from other people to you whenever you tag your bookmark

To be also honest, i must say that Foxmarks is great but you need to disable its sync from time to time and make it on shutdown because it has currently a bug that causes a slowdown to firefox on synchronizarion

i wish this post can help you getting rid of your slavery to GMarks. see the light with Foxmarks instead

Software Update Mess

2009-03-06T06:12:00.000-08:00

Nowadays, every software application has its own update program that is run in the background waiting for any new update to occur and notify you with it.

I see this is leading to more processes being run without any reason and taking time on startup and even bandwidth and cpu clocks

I got rid of all these after using
Filehippo software update program

this program is very nice one, it do this update idea but all these update programs are now in one place only, i stopped firefox update, apple update and so many and whenever something new exists this update program just tell me

I suggest it for anyone who hated such update programs, one for java, one for adobe reader, another for open office and so many others

Hope this may be useful small tip for Windows users like me

Cascaded HTML SELECT elements using JQuery

2009-03-05T08:00:00.000-08:00

Many web applications at least from which i worked in demanded the existance of many select boxes stacked over each other. These HTML select boxes are used for filtering data underneath them.

But the idea is that in some cases, these ones are dependent on each other. One case can be a filter for cars that starts with types, then makes then trims and so on. In some of these filters the amount of overall filters data are not so large that they can all be retrieved with page once and then JS manipulation can be used to propagate between them.

I used to make this myself until i found myself fed up of doing things that should be generic. I looked for a plugin for JQUERY which i use as a JS framework in my application. Lucky me. i found one that does this cascading on two levels and then you can repeat on others until all get cascaded together.

The solution i found was that blog post
the solution proposed here was very nice but i found that i need to modify it somehow to be much more elegent and complete from my opinion. You may disagree with me in some modifications and it's your right to do so. Take the code and modify it as you wish

Before looking at my code, you can have a look at the post in order to be familiar with my solution.

Here is the code


cascadeLists: function(parentId, childId, callbackFunction){
  $("body").append('');
  var childOptions = $('#' + childId + ' option');
  $('#' + parentId + childId).html(childOptions);
   
  var parent = $('#' + parentId)[0];
  if(parent.options[parent.selectedIndex].value == '')
    $('#' + childId).html(childOptions.filter('[value=""]').clone());
   
   $('#' + parentId).change(function(){
      var parent = $('#' + parentId)[0];
           
      var selectedOptionValue = '';
      if (parent.selectedIndex > -1)
        selectedOptionValue = parent.options[parent.selectedIndex].value;
           
      if (selectedOptionValue == '')
        $('#' + childId).html($('#' + parentId + childId + ' option[value=""]').clone());
      else {
        var childs = $('#' + parentId + childId + ' option[parent="'+selectedOptionValue+'"]');
        if(childs.size() == 0)
          $('#' + childId).html($('#' + parentId + childId + ' option[value=""]').clone());
        else
          $('#' + childId).html($('#' + parentId + childId + ' option[parent="' + selectedOptionValue + '"]').clone());
      }
   
      $('#' + childId).trigger("change");
           
      if(callbackFunction != null) callbackFunction(selectedOptionValue);
    });
}

In order to use this function, you will do as this example

<select id="parent"> <option value="1">xxx</option> <option value="2">xxx</option></select>

<select id="child"> <option parent="1" value="1">xxx</option> <option parent="1" value="2">xxx</option> <option parent="2" value="3">xxx</option></select>

just as your normal select boxes, but with additional IDs for each one of them and for the select box that have parent, each option state in its parent attribute the value of its parent

One Question will be: What if i want to add dummy option such as --select-- for both selects and whenever i choose this option in parent, child should be also this dummy one. Don't worry code handles this for you. If available this will be done. As for the dummy one, don't add to it a parent and let its value = ''

Now, just call the function and pass parameters:

parent select id
child select id
callbackFunction (optional) if you wish certain function to be called on any change to parent select, this function is called and pass to this function, the value of the option which was selected. Use this value to make any extra manipulation you want to do in the HTML page

Modification made are:

this method depend on option sent to it in order to show empty option on child, this isn't very nice in all cases, especially when dealing with internationalizations, in this case the data added here will have different text from a language to another. That's why i give you the ability to add empty option at the top of options and use it instead of generating one myself. that way you can choose what to write in this without changing in the JS Code The new logic is if there is an empty value option in child
Another option is needed is to call certain function whenever a change occurs on parent level and this function takes the new value, this way you can do any changes that was waiting such change like loading in Ajax some changes in page
One last change is that i see adding a parent attribute will be better than using sub_ and parent value

MySQL 5.1 Error with Rails 2.2

2009-02-19T09:30:00.000-08:00

if you found yourself stuck with such error

ArgumentError (NULL pointer given):
(eval):3:in `each_hash'
(eval):3:in `all_hashes'

then this might be useful for you.

Actually i don't know why it happened but i reached that this only happened between Rails 2.2 and MySQL 5.1
when i downgraded to MySQL 5.0 everything went fine

of course it will be nice from you to dig and search inside to know WHY THIS HAPPENED but if you have no time AS WAS THE CASE WITH ME, such info can help you survive

Which Stemming Algorithm To use ??

2009-02-17T03:49:00.001-08:00

suppose you want to search for something in a full database of words, using exact matching is a solution but will miss some data you may need

for example if you search for fishes, then document having fish as one of its words will be nice to retrieve them. That's why it is suggested that you use a stemmer so that words in your documents is transferred to its root making retrieval much wider than before

stemmers will take fishes and return it to fish and take your search keyword and stem also then search returning documents you want

there are many types of automatic stemming techniques, i found them to be one of those

Brute Force
Suffix Stripping
Lemmatisation
Stochastic
Hybrid
Affix
Matching

This is a long list of algorithms but the question that i didn't find an answer for it was which one to choose and in which cases and why
i didn't find a fast solution on the internet and i think to get a solution i have to dig inside papers published in that field

anyway, in my two projects i was using Ruby and used in both a stemmer belonging to the Affix class
i found two solutions one called porter and another called snowball

i tried both on a small list of 116 terms that i have in my db and results was the same except in these cases

emotionality -- emotion -- emot
joy -- joi -- joy
negativity -- neg -- negat

the first is the actual word, the second is porter result, the third is snowball result
my opinion is that the result in this small sample is 2 for snowball and 1 for porter
BUT still i can't say which one should i use, that's why i have shown these cases here in that post to see your opinions in that matter

One last point to say is that porter overstemmed negativity and snowball overstemmed emotionality

I am waiting for your comments as i need to take decision and choose one to use in my new project

HTML Scrapping using Javascript ((for google gadgets))

2009-02-11T03:41:00.000-08:00

Some friends at my company were working on doing some google gadgets. A large sum of gadgets were depending on data gathered from other websites which lack of any XML or RSS service providing this data in a direct way.

Since this is a problem we will face every now and then, we started thinking about a more generic solution to use in any gadget depending on such source of data.

The solution we reached was one of these three

using a scrapping service such as Dapper or Yahoo pipes to do the scrapping on behalf of us and returns a well formed XML file to use in any gadget
create a google app engine that we call and it scrape the data and returns XML to us
using JS for scrapping HTML pages

the first and second solutions may seam the same and actually they are except that Dapper isn't that reliable as it sometimes fails due to extra load on it while google app engine was proven to survive under high request rates

Anyway, i liked the third solution and said to myself lets give it a try and see if it will be performant enough or not. I thought scrapping html using JS is an easy matter that can be done easily in any google gadget but i was proven not to be like that at all. I will summerize the trials i made here starting from those who failed to the last solution that worked.

Depending on Google Api method "_IG_FetchXmlContent". This way failed easily because it was expecting XML document and was faced with HTML Page. It gave me parse error on Doctype line. The result is FAILURE
Depending on Google Api method "_IG_FetchContent". This way gave us the html as it is and it was time to parse it using DOM Parsers built already inside browsers. I tried doing so using Firefox browser but also got parse error because this is not a XML document but HTML one and parsers available only expects XML. The result is FAILURE
Repeating step 2 again but after using a regular expression to take only inner HTML of body tag. DOM Parser failed on one of the comments lines present in the HTML page which may appear in
may pages so this isn't a generic solution to be accepted. The result is FAILURE
Using Regular expression to get body inner html and then add this to a hidden div then using normal JS methods for traversing DOM nodes considering this Div as my root. The result is SUCCESS

Since, the fourth trial was successful i made a generic method that anyone can use in his gadget. this simple method will just get html and scrape based on your scrapping function. To understand what i mean, have a look at the function definition first

scrapeHTMLBody = function(url, dataHolderId, scrapeFunction){}

as in this definition we see that the function needs some parameters

url to retrieve html from
dataHolderId the id of the hidden div that the retrieved html will be added to it
scrapeFunction a function that takes the hidden div as a root element and use JS to get data desired "every one should write his according to what he wants to retrieve"

and this is the implementation of it

scrapeHTMLBody = function(url, dataHolderId, scrapeFunction){ _IG_FetchContent(url, function(responseText){ operate(responseText, dataHolderId, scrapeFunction); }); }

operate = function(responseText, dataHolderId, scrapeFunction){ var body = /<body.*?>((.|\n|\r)*)<\/body>/.exec(responseText); var bodyData = body[1]; _gel(dataHolderId).innerHTML = bodyData; scrapeFunction(dataHolderId); }

these two functions are used to get html page then retrieve body inner html then call the scrape function passing to it the if of the hidden div containing the html body
it is your responsibility now to write the scrapping function desired based that this div is the root of your DOM tree

this is an example of a scrapping function i defined

scrape = function(dataHolderId){ var elements = _gel(dataHolderId).getElementsByClassName('main'); var noktas = []; var num = elements.length; for(i=0 ; i<num ; i+=2) noktas.push(elements[i].childNodes[0].innerHTML); for(i=0 ; i<noktas.length ; i++){ var e = document.createElement('p'); e.innerHTML = noktas[i]; document.body.appendChild(e); } }

That's it, i think you are ready now to use these two functions in any gadget whose data source should be scrapped
This method should be better as here all processing is made on client machine rather than any other servers

Ubuntu No Audio Bug After Hibernating

2009-02-04T04:11:00.001-08:00

Well, this was very annoying bug in Ubuntu latest version

but, thanks God it was fixed on 22 Jan

This is just an announcement, you can follow the ticket here

[>30] Mounting my N80 mobile on Ubuntu Machine

2009-02-02T13:06:00.001-08:00

This shall be the first episode in my long series called ">30" which simply means that it took more than 30 minutes. I am collecting here some of the problems that i fought with for more than 30 minutes in order not even solve it but finding a resource on the internet that can relief me from such a pain.

The first episode of this series is talking about how to mount your Bluetooth device on your Ubuntu machine. I will explain such a thing on my N80 mobile and Ubuntu machine but i think anyone can use such points on any other devices and Linux machines.

Before getting into the main point, i would like to state that this post might get you the solution or not. I am writing it while doing such a task and while i am writing such introduction, i actually reach to the steps needed to get out of such misery journey. But, anyone whether we reach a solution together or not, i think this post will put you on a track to continue from so you won't lose your time reading it.

Note: points highlighted in bold style are points related to my case so it should change from one case to another.

The journey starts in this post Mounting a Nokia Phone a Little Bit Easier which is really a nice point to start with. It states these points

Find out your phone’s Bluetooth MAC address if you don’t know it already:
hcitool scan
Find out the OBEX FTP channel it uses
sdptool search FTP
Load the fuse kernel module:
sudo modprobe fuse
Make a suitable mount point for your phone:
mkdir /media/n80
Mount
obexfs -bXX:XX:XX:XX:XX:XX -BYY /media/n80
(where XX:XX:XX:XX:XX:XX is your phone’s MAC and YY is the OBEX channel)
Unmount when you’re done with your file transfers:
fusermount -u ~/Phone

the above steps depends on having some libraries installed on your pc which are obexfs and obexfs. i got them easily using apt-get. If after doing the above steps everything went well then you are done. Yes that easy but if not then continue with me to know what i did after this step.

For me, i was capable of entering the mounted folder and even list files and not only this i removed a file. i thought it is time now to copy some files from my pc to the device. after using the cp command, i took no time at all and then it finished. I was sure something went wrong because i am connecting using bluetooth and things shouldn't end that fast. I used ls command to check that it was added but then it hanged for some time and stated that the folder i am inside isn't a valid one. Simple, mounting failed. I went out and entered again and ran ls command but nothing was listed. I unmounted and remounted but it sometimes shows everything or simply nothing

In some cases, i run "hcitool scan" again and can't find my device even. I disable my bluetooth on the device and enable it to start working again.

Anyway, it was time to search for a solution for such a problem. I found this can't add files to mounted volume, no free space sony Z610i ticket. I digged inside it and read for a while and knew that they have the problem of copying files since the mounted point is stating no free space on the mounted device. They stated this is a bug and pointed to its ticket here. It stated that current version of obexfs don't manage S60 3rd edition mobiles probably and found the ticked closed invalid. Looking at comments i found someone stating that this should be fixed if i used "ObexFTP 0.22 / ObexFS 0.11" instead of the installed ones. These should be compiled as those got by apt-get are one step behind. Ok, no problem let's compile.

prepare an empty folder then wget these files one by one

http://downloads.sourceforge.net/openobex/obexfs-0.11.tar.gz?modtime=1213568386&big_mirror=0
http://downloads.sourceforge.net/openobex/obexftp-0.22.tar.bz2?modtime=1213568417&big_mirror=0
http://downloads.sourceforge.net/openobex/openobex-1.3.tar.gz?modtime=1150294112&big_mirror=0

tar these files using tar -xzf for .gz files and -xjf for .bz2 files. Now lets install them in this order openobex then obexftp and obexfs. This order is important because each one depends on the previous one. But, before doing so, these are some libraries i needed while compiling so make sure you have them installed

python-dev
libfuse-dev
libusb-dev
tcl8.4-dev
tcl-dev

now in each folder. run "./configure" then "make" then make "install" and you should be ready to redo the steps stated above which i wrote them at the beginning of my journey.

I did so but i am sorry to say that all this was useless for me. If everything went ok with you at this step then you are ready to have fun. If not then continue with me, i still have some points to state.

I reached this link now http://www.thinkwiki.org/wiki/How_to_setup_Bluetooth, i noticed that there is a section for symbian mobiles that is different from other mobiles. seems that i have to continue searching but i think enough for me this night. maybe tomorrow. If you still have problems till tomorrow then wait for my next post that i wish will settle things and put an end to this misery

Blogged with the Flock Browser

Try Wine

2008-12-15T08:36:00.000-08:00

Wine is a very nice thing, it is a program you can use in order to run applications dedicated for certain platforms on any other one

for example i use it to run some windows application i get to use them on my linux machine, although Wine isn't fully mature yet but a large number of applications run nice on it

Personal Area Network (PAN)

2008-12-12T03:38:00.001-08:00

Personal Area Network or shortly (PAN) is a communication network between a group of bluetooth-enabled devices (supporting PAN)

this network can be used between these devices in the same manner as local area connection i.e you can use it to transfer files between two devices or even more you can use it in a twisted way to enable one device to use other device internet connection after configuring a bridge between it and WLAN network

anyway it is a nice technique of communication, i liked it so much and tried to use it but i couldn't because my N80 doesn't support it. i suggest everyone to make use of this powerful facility

How to name your command line windows??

2008-10-05T07:20:00.000-07:00

if you are one of the people that use to open so many command line windows while working then you may have faced the problem knowing which one is opened on which task

for example, suppose you have a windows doing task X and another doing Y and another doing Z and so on. all have same default name added by windows when you open a command line window. Due to these similarities you always get lost getting to the window you want in one step

to get rid of this problem, use this small tip and name them with significant names using 'TITLE' command as follows

TITLE {{desired-name}}

that's it, now you can know which windows you have and select the one you desire in a fast way

Rails Console Tips & Tricks II

2008-09-23T03:43:00.000-07:00

This is my second rails console tips & tricks. if you interested in reading previous tips and tricks read my first blog post regarding that matter.

This second part is about one command called "load" that is available in your Rails console. you can use this command to do many things as

1. to reload any file that was loaded when starting console and don't get reloaded when using "reload!" command such as files in your lib folder. to reload any file just write

load "#{RAILS_ROOT}/lib/{{file_name}}"

2. if you faced a problem such as "MySQL has gone away" just reload environment.rb once again as follows

load "#{RAILS_ROOT}/config/environment.rb"

3. also if you want to test a bunch of code in your console not just a single line of code, and want to be able to change part in it and rerun it once again then doing so in console is very time consuming and annoying so rather than doing so, you can write this code in a file and then

load "{{your-ruby-file}}"

Ruby Make Problem

2008-09-08T07:24:00.000-07:00

Still in my early days using Linux and trying to compile ruby 1.9 on my machine in order to move my working environment to Linux

I downloaded ruby 1.9 source from ftp.ruby-lang.org and as stated in readMe file, i ran

./configure --prefix={{any-desired-path}}

successfully then ran

make

it took some time then i was faced with that error

executable host ruby is required. use --with-baseruby option.

What is that error? and what ruby?? i am compiling ruby how come i need it.

Anyway, i began searching for an explanation. i reached to a comment in one of the forums stating that this error can occur in these two cases

1. ruby isn't in the default path
2. ruby isn't installed

More investigation showed that i should have any previous version of ruby installed in order to use it with that make file as follows

./configure --prefix={{any-desired-path}} --with_baseruby={{full-path-to-ruby-executable}}

then run make command once again. And it worked finally

What a nice recursive world, "ruby is needed to compile ruby"

configure: error: C compiler cannot create executables

2008-09-04T03:21:00.000-07:00

In my first start to use ubuntu, i wanted to compile ruby on my machine

i downloaded ruby source and executed configure file to generate make file

./configure

i got this error

C compiler cannot create executables

although i have gcc latest version installed

after investigating this error i found that the solution was

sudo apt-get install build-essential

after doing so, everything worked fine

SVN Branches do HELP

2008-08-19T09:41:00.001-07:00

Q: Have you heard about SVN branches ?
A: yes i have

Q: Have you used it before ?
A: no i didn't

if this scenario applies on you, then let me tell you why you should deal with SVN branches. I will try to prove my concept through a small scenario introducing 3 problems you might face while using SVN and how they can be solved with or without branches

Scenario

suppose you are working on a project and after a while you reached a stable state at which you decided to go live. after going live, you decided to move ahead and work on the next version of your project. while working on this version, a sudden problem occurred on your server and your online version was lost.

HOW WILL YOU RETURN IT BACK LIVE AFTER YOU STARTED DEVELOPMENT ON A NEW UNSTABLE VERSION?? This is problem 1.

hopefully you succeeded to do so, but later on, you found a bug and started fixing it but you can't leave your online version containing this bug until you finish development of the new version.

HOW CAN YOU COPY THIS BUG FIX TO THE ONLINE VERSION?? This is problem 2.

you also succeeded doing so, now suppose that problem 1 happened once more but this time after fixing problem 2.

HOW WILL YOU GET BACK ONLINE WHILE NO REVISION CONTAINS BUG FIXES MADE?? This is problem 3.

Without using SVN Branches

In problem 1, you will start looking at the repository revisions searching for the revision number at which you decided to go live in order to check it out again on your server and return online
accepted solution but considering the time and effort you took to do so i consider this solution a difficult one

In problem 2, you will fix the bug then select from your project the altered files and copy them one by one to your server without selecting any other files related to the new version you are working on. this solution will take a lot of time and effort and some errors might be expected while selecting files

In problem 3, i think you will have to work in your weekend in order to find the stable version and add it to bugs fixes that you have made

After SVN Branches

In problem 1, since you have created a new branch for the new version leaving your stable version as it is. all you need is to check it out again.

In problem 2, when you find a bug you will just commit all files you changed to your branch then start fixing that bug and commit the changed files then finally merge the difference between the last two revisions on branch with trunk

In problem 3, take what is in the trunk directly

I think it is now clear how helpful branches can be especially in a project that is expected to change from time to time which is actually common in a large of projects

Different encoding error

2008-08-18T04:04:00.000-07:00

if you found yourself suffering with this error


ActiveRecord::StatementInvalid: Mysql::Error: #HY000Incorrect string value

don't cry or check your code looking at the data you save thinking that you might have left some bad characters around causing you this misery

all you need to do is to check the used database connection encoding against your db tables encoding, most probably you will find them different

if so then the solution is one of these two options
1) remove encoding field from specification and use default db encoding
2) change the encoding of your db tables to this encoding

once you do so, you will have a good life

acts_as_modified plugin

2008-07-22T12:15:00.000-07:00

is it changed? which attributes? what are the old values? in some cases you may be asking yourself these questions when dealing with some models

When should you care about these questions?

suppose you have a model doing an expensive operation on save. this operation is useful only if one specific attribute changed. so doing it on every save neglecting which attributes changed will degrades performance significantly

So what can you do?

you can load model old data from database and check if that attribute has changed or not and if so do your action

or

you can use acts_as_modified plugin, it helps keeping modified attributes with their old values within model in order to be capable of answering your questions without doing any extra database trips

I am exited, from where can i get it?

it depends on your rails version

1. ((rails 2.1)) do nothing, dirty objects already embedded in this version
2. ((rails r7315 and above)) install from http://svn.viney.net.nz/things/branches/acts_as_modified_edge
3. ((ELSE)) install from http://svn.viney.net.nz/things/rails/plugins/acts_as_modified

That's it, Enjoy the fun

Rails Console Tips & Tricks

2008-07-20T04:39:00.000-07:00

Rails console have many useful tips that you may exploit as

Getting out of a block of code

If you're in an indented block of code (e.g., something that should be terminated with end) and want to escape without finishing, you can use ctrl + c to get back to the top level. The current expression will be ignored and not run.

Accessing the last return value

you can use the magic variable (_) for accessing last value
and you can then assign this value to a variable as x = _

Rollback all database changes on exit

script/console --sandbox

Rails Validations & Unit Tests

2008-06-22T06:47:00.000-07:00

I was doing a new feature in an existing rails project, and after finishing that feature, i ran all tests to check that everything is fine and no failures or errors occurred after my change

i ran the tests and found a failure in one of the unit tests, i opened it to see what is the problem. at this point i found in that file a function that gained my attention.

before stating what is inside this function, lets have a look at the part of the model that is being tested by this function


class Feed < ActiveRecord::Base
  validates_presence_of :name

  ...
end

we can see in this part that name field should be present for the model to be valid, now lets look at the testing function


  f = Feed.new
  assert !f.valid?
  assert f.errors.invalid?(:action)

So, what is the point in this code? the point is that at my first look at this function, i considered it a redundant test function because it checks an already tested rails function which is the validation function.

so i asked a friend of mine about his opinion, and he agreed with me that it is redundant and useless

later on, i asked someone with greater experience about his opinion also in this function as it represents a class of functions found in many models checking such validations.

But his opinion was different. he considered it useful and justified that by stating that this function tests the contract (which is that no feed is accepted if no name is stated) with the client, so that if someone later on changed this part in a wrong way, this test will alert us that the contract has changed. so that we reconsider our action to see if it has been really changed by the contract or by mistake

i found that answer reasonable and added to me a new way of thinking towards tests which is testing the agreements rather than checking whether our code is working or not

Rails: Modules & Routes

2008-06-22T06:13:00.000-07:00

I used to make a module for every set of controllers which are doing certain task or feature because this way, things become organized in a better way

but something weird happened to me while trying to test a controller inside one of the modules, to see what happened i represent my case here by an example

i added in routes file this snippet of code


 map.resources :items do |item|
   item.resource :rate, :controller => 'ItemActions::Rate'
 end

and in my controllers folder, made a folder called 'item_actions' and added inside it a file called 'rate_controller.rb', inside this file i wrote


  module ItemActions
    class RateController < ApplicationController
      ...
    end
  end

restarted the server and run my work, everything fine and i can call the controller actions as any other controller

then i opened the controller test file and inside one of its test functions, i added this code


  post :create, {:item_id => 1, :rating => 50}

i got that error


 No route to match {:controller => 'item_actions/rate', :action => 'create', :item_id => 1}

i changed the controller to be


  class ItemActions::RateController < ApplicationController
    ...
  end

but nothing changed, i am still getting that error once again, searching for a long time, i reached to a working solution which was to open routes file and change it to


 map.resources :items do |item|
   item.resource :rate, :controller => 'item_actions/rate'
 end

That's it, it worked and i didn't get the error once again

Remember User Login Via Cookies

2008-06-18T01:22:00.000-07:00

How to persist user login via cookies if session has expired?, this is a question almost answered in most of nowadays web applications, So how can this be done?

The idea is simple, and can be summarized in these simple steps

1. add a new column in your users table called login_key or remember_token or any name you prefer of type string

2. whenever user tries to login and his login is successful, add in his cookies a cookie called anything ex: remember_token and let its value be a generated random string of length around 40 characters for better security which is our token, and save this token in the db with that user

3. now add, in any action you need to check in it whether this user already in system or guest, a snippet of code that checks if the person isn't logged in whether the request has a cookie with the name choosed before and if so, check the db looking for a user having that token, if found then this user is logged in as if he did that in the normal ordinary way

That's it, now you can use cookie to log user in if his session has expired. There are also some points you may need to take care of when using this technique

1. Make cookie expires after some time so that it gets created once again with new token

2. In order to provide extra security, add with that token another info about the user

ActiveRecord Association Weird Length Method

2008-06-16T05:15:00.000-07:00

The method 'length' defined for ActiveRecord associations works in a very weird manner.

consider this example


class User < ActiveRecord::Base
  has_many :feeds
end

i was using the class and writing this snippet of code to get the count of feeds in the db


@user = User.find :first


puts @user.feeds.length

looking at the log, i found that the added query was


SELECT * FROM `feeds` WHERE (feeds.user_id = 1)

and not


SELECT count(*) AS count_all FROM `feeds` WHERE (feeds.user_id = 1)

so in order to get away from such weird sql statement which will exhaust a very large bandwidth getting all records to count them, use this instead


puts @user.feeds.count


puts @user.feeds.size

but don't use 'length' method

SVN Branches

2008-05-15T07:37:00.000-07:00

Imagine working on a deployed project, and you want to add some new features and changes on the running copy.

Since making new features may cause problems in the deployed project, you need to have another copy from the application and work on it freely then merge it with the running copy to apply your changes.

Subversion makes this very easy, all you need to do is to create a new branch.

So, what is a branch?

in order to understand that, lets look at the repository location hierarichy, you will find that it is as follows

--
-- trunk
-- branches

trunk is the folder containing the version of the website currently deployed

branches is another folder containing some other versions, as a new edition of the application

Now, in order to do that, just run this command


svn copy ((repository-path))/((project-name))/trunk ((repository-path))/((project-name))/branches/((branch-name)) -m "Creating a private branch in order to ... bla bla bla"

now you are done with your branch, checkout the project from that new path, and do any changes you want freely then at the end merge with your trunk to apply your changes

acts_as_changed plugin farewell

2008-05-05T11:08:00.001-07:00

i used acts_as_changed plugin since a while and i have been enjoying the idea of being capable to know which attributes has been changed in my model and to take action upon certain attributes changes

and today i only tried to look at it again to know if there are any changes made in it or not but, i was shocked when i read a comment from the owner stating that it isn't plugin anymore

it is now an alternative base.rb file that you can update your activerecord with if you want to enjoy the plugin functionality once again

and he stated that he did so because he was facing problem in callbacks that he couldn't solve till now except through this way

anyway, i hope he finds a solution for it so that we all benefit from this plugin functionality

Memcache Trick

2008-04-23T08:23:00.000-07:00

i was faced with a very weird behaviour while working with memcache-client with two memcache servers

first lets look of my configuration to the client in order to explain on it
On my first server

memcache_options = {
((some options)
}
CACHE = MemCache.new memcache_options
CACHE.servers = [((SERVER1)), ((SERVER2))]

On my second server

memcache_options = {
((some options)
}
CACHE = MemCache.new memcache_options
CACHE.servers = [((SERVER2)), ((SERVER1))]

after running both servers, i noticed that what i enter in memcache on that server isn't available to the other server

Where is the problem??

here it is, a very strange reason

CACHE.servers = [((SERVER2)), ((SERVER1))]
CACHE.servers = [((SERVER1)), ((SERVER2))]

this is the problem, one uses SERVER1 then SERVER2
other uses SERVER2 then SERVER1

i thought that memcache client when checking for a key, he looks at the first, if not found then looks at the second then say it is a miss
but i realized that if it isn't on the first then it doesn't look at the second

That's it, just take care, cuz small things can waste your time and energy looking after their reasons

Adding Data Through Migrations

2008-04-23T06:52:00.000-07:00

one of the points i am always faced with while adding some migrations, is my need to add some system data as a static group for Administrators or any other.

so simply i create a new migration and add the following
ex:

Group.create((some_attributes>))

and it simply works fine, until someone later one adds a new column
in another new migration and adds validation on that column in the Model

and here the problem occurs
my old migration throws exception because of validation added on a new column added later in another migration but Model carries code validating it

how to solve these problems?

i suggest not using create at all
but rather use new and set our attributes
then save without validations

Thick Models Thin Controllers

2008-04-03T07:17:00.000-07:00

i have worked for about six months with Rails and through this time i was coding a lot in my controller which has shown to me lots of problem later on

for example,

i was having a Topic model and Post model, where every topic has many posts
and in each of these two models there is a field called archive

i was having a controller that changes the archive field to a certain value then pass on every post it owns and change its attribute also

later on, i made another controller and in it i also changed that field in it and passed on all posts to do the same change in it

this led to redundancy and duplication of code

but the problem became more complicated when i wanted to change the logic of this topic and its posts replication to a new way

here i have to move on my controllers and change this code in them

but i have added the code that observe this change in topic attribute and apply it on all its posts inside the model, i would haven't faced this problem anymore
and this will give me flexibility to do changes easily also

so this led to the idea of having my controller thin containing the least amount of code and my model thick containing large amount of code

script/console reload models

2008-03-26T04:35:00.000-07:00

if you are one of the Rails applications developer, then you would have used this command before

ruby script/console

it is very nice and gives you the ability to load your environment and try it as you wish, but what if you changed anything in the application models, you will find that the console isn't aware of these changes at all

so what can you do to make him aware of your changes?

here it is, just run this command in your console

Dispatcher.reset_application!

Now you can try the new changes you made in your models

Unobtrusive JavaScript

2008-02-27T09:53:00.000-08:00

While serving the internet looking for something, i found someone talking about something called 'Unobtrusive JavaScript', it was the first time for me to hear that expression so i looked for it and found that

Unobtrusive JavaScript is one of the emerged paradigms in JavaScript, its main idea is to achieve more separation while working on web applications

Since now, we almost reached to an agreement to separate presentation from content/structure through separation of CSS from HTML, this paradigm is greedy and wants extra separation of behaviour from structure

lets look at this example to clarify things out

<#input type="text" id="date" onchange="validateDate(this);" />

here we can see that our structure is that simple input field but it is mixed by this onchange function which shows to us what this input does on change

Now, look at that one

<#input type="text" id="date" />

here we have only structure and no behaviour, so WHERE SHOULD WE ADD OUR BEHAVIOUR?, we should do that with javascript away from our presentation as follows

window.onload = function(){ //Wait for the page to load.
var input = document.getElementById('input');
input.onchange = function(){
//Do something changed.
}
};

This is the simplest illustration that i can give to this paradigm, it makes you live in a world of layers (structure, presentation, behaviour)

actually, i am admired with this paradigm so much because it layered our application and thus allowed for better separation but i think one of its problems is that looking at structure won't give you a quick idea about what elements behave until you visit the behaviour layer

finally, i would like to end with this nice quote from one of the articles i read about this topic

The first rule of the unobtrusive Javascript club is don't talk about the unobtrusive Javascript club.

this post isn't done yet until i hear from you what do you think about this emerged paradigm and whether it should be the dominant in the following years or not

Ip2Country Mapping

2008-02-27T02:34:00.000-08:00

I was working on a web application that required mapping every request to its corresponding country in order to use this data later in site statistics

i searched over the internet for a solution that can do this mapping for me, i found a lot of solutions that are great to the extent that they can also get the ISP and geographical location of the request but MOST OF THEM WERE NOT FREE

Since i was only looking for a simple solution that simply maps ip to corresponding country, i searched for free versions of these solutions which are limited but free

i found this Website that offered a nice & easy solution for this issue. Although, this website has a very nice tutorial but i will explain what i did because i modified the steps provided by the website somehow

Just do as follows

1. Download the CSV file that contain all mapping details

2. Create a new table in your database that should carries all data needed to do
this mapping, you will needs these columns
* begin_num - FLOAT
* end_num - FLOAT

These are two magical numbers that will be used later in
our mapping

* ISO - VARCHAR(2)
* Name - VARCHAR(255)

2. Import CSV file into that table but, remember to choose only 3rd, 4th, 5th, and 6th values while importing - neglect 1st and 2nd one because they are useless in our mapping

3. On every request, apply this formula on the IP as follows

ipnum = 16777216*w + 65536*x + 256*y + z

where

IP Address = w.x.y.z

4. Use ipnum to retrieve corresponding ISO or name of country through this simple query

SELECT ((required-columns)) FROM ((table-name)) WHERE ((ipnum)) BETWEEN begin_num AND end_num

That's it, you can now map any request to the corresponding country with an accuracy reaching 98% as stated by this website

Note:
1. This query may return nothing if the IP can't be mapped and this will happen especially when using localhost
2. Always index begin_num and end_num for better performance
3. Always update your CSV file because it is updated every now and then on the website so keep an eye on the website to get updates

Subversion for Apache 2.2.x On Windows Platform

2007-12-19T14:54:00.000-08:00

In this post i will talk about how to add Subversion modules on Apache 2.2.x, the problem that i faced while doing this configuration is that SVN executables exist for previous Apache versions only, so i thought about writing this post to give fast summarized steps to prepare SVN for this version of Apache

Enough talking, time to start configuring our Apache, i will do that step by step

1. Choose the latest version of Apache 2.2.x

2. Choose the latest version of SVN(on the time of writing this post, no SVN executable was there to be used with Apache 2.2.x)

3. Install Apache (procedures aren't included in this post) then install SVN

4. Create a repository for your SVN as in that example

mkdir c:\svn
svnadmin create c:\svn\repos

5. Configure Apache server configuration file, you can reach this file through that path

Start/All Programs
/Apache HTTP Server 2.0.50
/Configure Apache Server
/Edit the Apache httpd.conf Configuration File

6. Look for the part of the file that is showing loaded and unloaded Apache modules, and make sure that these two modules are loaded

LoadModule dav_module modules/mod_dav.so
LoadModule dav_fs_module modules/mod_dav_fs.so

7. Download the binary built against Apache 2.2.x from this link

8. Extract the archive you downloaded to a folder and then replace files in your SVN directory folder with these extracted files

9. Go to (SVN directory)/bin and take the two modules

mod_dav_svn.so
mod_authz_svn.so

found there and copy them to (Apache directory)/modules

10. Open Apache configuration file again and add these two lines

LoadModule dav_svn_module modules/mod_dav_svn.so
LoadModule authz_svn_module modules/mod_authz_svn.so

in modules part after the two lines that we added previously in step 6

11. Add at the end of the configuration file these lines

where the path beside the location tag is the path that you will access the SVN through it and the path beside SVNPath is the path where you installed your repository

These are the required steps needed to prepare your SVN with this version of Apache, i hope i was helpful in that article and up to the point

For more details on how to add authentication for SVN and other details follow that link, where you will find other steps related to SVN preparation on Apache Server but for old versions of Apache, skip first part as we already did it and read from the part called "Securing Your Server"