chovy.myopenid.com

[Ruby] dual object creating in Rails

2009-08-03T06:50:01-07:00

This code works, but could use some more love.

The idea is when a user creates a new "Topic" -- they automatically create a new "Post" as well ( handled by the textarea in the "new topic" form).

The code from the book I'm reviewing is outdated and you can no longer pass more than one argument to Foo.new() in rails 2.3

The commented out params[:post].merge(...) command didn't to work.

  # POST /topics
  # POST /topics.xml
  def create
    @topic = Topic.new(params[:topic])
    @topic.forum_id = params[:forum_id]
    @topic.user_id = logged_in_user.id

    #params[:post].merge({:topic_id => @topic.id, :user_id => logged_in_user.id})
    @post = Post.new(params[:post])
    @post.topic = @topic
    @post.user_id = logged_in_user.id

    respond_to do |format|
      if @topic.save && @post.save
        flash[:notice] = 'Topic was successfully created.'
        format.html { redirect_to forum_topic_posts_path(:topic_id => @topic, :forum_id => @topic.forum.id) }
        format.xml  { render :xml => @topic, :status => :created, :location => forum_topic_posts_path(:id => @topic, :form_id => @topic.forum.id) }
      else
        format.html { render :action => "new" }
        format.xml  { render :xml => @topic.errors, :status => :unprocessable_entity }
      end
    end
  end

[Ruby] xml rss feed

2008-11-05T07:13:41-08:00

Looking for a better way to define the xml.description node...which needs html embedded as CDATA, but I think it is ugly concating this way with "+".

Generates a description and a link at the bottom.

    for directory in @directories
      xml.item do
        xml.title "#{directory.name} (PR #{directory.pagerank})"

        //cleaner proc here?
        xml.description { xml.cdata! simple_format(directory.description) + simple_format(link_to directory.software.name, software_url(directory.software)) }

        xml.pubDate directory.created_at.to_s(:rfc822)
        xml.link directory_url(directory)
        xml.guid directory_url(directory)
      end
    end

[PHP] On Getter and Setter

2008-10-08T18:41:57-07:00

I tend to do special handling when I need special handling. Like data validation being different depending on the variable in question.

$foo->email_addr = 'http://www.';

You would want different validation there then say:

$foo->web_addr = 'foo@bar.com';

[JavaScript] Toggle "Check all / Uncheck all" with jQuery

2008-10-06T20:16:20-07:00

I have a simple toggler to mass select (unselect) all input checkboxes with the "id" class in a form. It should be triggered from an anchor tag and have its text updated.

The things I don't like are the hard codings of the "mode" and the anchor text labels.

Q = jQuery.noConflict();

Q(document).ready(function(){
    Q("#toggler").click(function() {
        var mode = this.mode == 'off' ? 'on' : 'off';
        this.mode = mode;

        Q("form#the_form input.id").each(function(){
            mode == 'off' ? this.checked = true : this.checked = false;
        }); 

        Q(this).text() == 'check all' ? Q(this).text('uncheck all') : Q(this).text('check all');
        return false;
    });
};

// called from a link...
// <a href="#" id="toggler">check all</a>

[JavaScript] On Small Javascript Code

2008-10-06T19:46:44-07:00

Some would say you should only have one return point from a function. And there's no need to pass the location object each invocation.

var getAnchor = function (default) {
  default ||= 'news';
  return document.location.toString().split( '#' )[1] || default;
};

loadURL(getAnchor()); // default is used if no anchor
loadURL(getAnchor("help")); // default is set to 'help' and used if no anchor

[JavaScript] On Small Javascript Code

2008-10-06T02:05:25-07:00

Get position of "#" in the location.
Check if we have a "#" at all (-1 is "none")
Get substring from starting position to end of location length.

Uses a two ternary statements to set has_hash boolean, and defining the page.

/*

var biz;
if ( some_test ) { biz = 'foo' } else { biz = 'bar' }

can be written as a one-line ternary:

var biz = some_test ? 'foo' : 'bar';

*/

var pos_hash = document.location.indexOf("#");
var has_hash = pos_hash > 0 ? true : false;
var page = has_hash ? document.location.substring(pos_hash + 1, document.location.length) : 'news';

loadURL(page);

[Ruby] On render partial layout

2008-10-05T02:18:11-07:00

I think you just need to call it as a function call, instead of a symbol.

Try this...

layout :set_layout # might be just set_layout



render partial 'new_alert',  :layout => set_layout


def set_layout
     logged_in? ? "application" : "home"
end

[Ruby] On link_to with an image_tag and text

2008-10-05T01:58:51-07:00

The image itself would be a 32px x 96px image with the 3 states (link, active, hover) shifted up 32px for each state. This allows the image to preload and there will be no delay when doing rollovers.

The text is indented offscreen -9999px so it is not visable when the browser supports CSS. If the browser has CSS disabled, and/or images disabled, the user would simply see the text.

The title attribute is added to the link tag to give users a tooltip and assistive devices.

## view

<%= link_to "Go back", path_to_go_back, { :class => "prev", :title => "Previous page" } %>

## css
<style type="text/css" media="screen">
// normal button icon
a.prev:link, a.prev:visited, a.prev:hover, a.prev:active {
    display: block; width: 32px; height: 32px; text-decoration: none; text-indent: -9999px;
    background: transparent url('/images/icons_prev.png') no-repeat 0 0;
}

// highlighted button
a.prev:hover {
    background-position: 0 -32px;
}

// depressed button icon
a.prev:active {
    background-position: 0 -64px;
}
</style>