Jonathan Gnagy's Blog

I’ve had some time to burn since I’m not working for a couple weeks while I’m on paternity leave. During the night, my wife and I take shifts watching Marissa sleep (yeah, we’re paranoid like that). During my last shift (this one), I’ve accomplished quite a bit. I’ve managed to:

• Utilize Google’s Translate javascript API to provide site-wide translations
• Create a fully-functional, albeit simple, Todo / Checklist system and integrate into my blog
  • Uses Scriptaculous effects
  • I serialized the list items into an Array, so I only need a single column for an arbitrary number of list items. Also, no need for a separate “ListItems” model / table.
  • It allows me to mark lists as public, so everyone can see them. By default lists are private.
  • All adding and removing from the list is done via Ajax (with a non-javascript HTTP/1.0 POST fall-through), so it looks pretty snazzy.
  • I’ve started an experimental option to apply my tagging system to lists as well, but I haven’t figured out how I will make it all work, or how to display it.
• Polish up some sections of my site that weren’t quite passing XHTML 1.0 Strict compliance tests. I’m sure there are more out there, so feel free to comment if you happen to come upon a page this isn’t. Come to think of it, I’m sure the list stuff isn’t yet, since I didn’t check it.
• Make more of the site RESTful and support more pure XML operations (for scripting and such).

I’m sure I did more stuff, but I can’t really remember it right now. My eyes are awfully blurry… I’ll try to figure out some code to post tomorrow.

I added some minor improvements to my blog. Some of these feature include:

• All tags now offer feeds, even if they aren’t in the “Popular Tags”. Just click on the tag, and the grey banner contains a feed link.
• All successful searches (i.e., those that return results) allow for subscriptions via Atom feeds. Similar to tag feeds, they exist in the grey banner that details the search term.
• All uploaded media is available and searchable from the newly added “Media” link under the also newly added “Extras” section on the right.
• Several small improvements to admin areas…

Here’s some code:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

def feeds
    case params[:feed]
      when "all"
        @entries = Entry.find(
          :all, 
          :limit => 25, 
          :order => "updated_at DESC"
        )
      when "by_tag"
        @entries = Tag.find_by_name(params[:tag]).entries.find(
          :all, 
          :limit => 25, 
          :order => 'updated_at DESC'
        )
      when "by_search"
        query = "%#{params[:q]}%"
        @entries = Entry.find(
          :all, 
          :conditions => ["body LIKE ?", query], 
          :limit => 25, 
          :order => 'updated_at DESC'
        )
    end # case params[:feed]
    respond_to do |format|
      format.atom  { render :layout => nil }
    end
end

I do side jobs every so often, and I have a small list of regular customers. They’re pretty regular customers, and quite often I need to supply an invoice while at the client location. Rather than supply an actual invoice, I would just use an excel spreadsheet. Although this works, it doesn’t really come across as a professional business should.

Now I’ve added a customer tracker and invoice generator section to my company’s website. Of course, I can’t link to an actual invoice, but I’ve set them up so I can send direct links to my customers. So now, I can create and email an invoice right in front of a customer.

My blog got its first spam bot attack yesterday. I cleaned it up, of course, but I’d like for it to not happen again. So i’m trying something I found on another Ruby on Rails dev’s blog. We’ll see if this does the trick for a little while…

In the beginning, my blog didn’t have any way to notify me when someone comments on an entry. After a while, I added an ActionMailer class (called Notifier) to email me when a comment is saved. Realizing that this might open up the door for people to spam me, I added some very basic spam checking, looking for certain inappropriate key words through regular expression matching. Maybe some day I’ll actually try and tie-in Spamassassin or something.

Today, I added a couple more advancements in regards to comments. Now, it also sends me an Instant Message through Jabber (XMPP) notifying me of a new comment. I also decided to move the sending of the email and the instant message to its own thread because I noticed some slowness when testing new comments. Now things are fast and smooth again.

Here’s the code I used to make the IM and email happen:

Add this to the Comment model:

1
2
3
4
5
6
7

...
  protected
  def validate
    SPAM_REGEXES.each do |r|
      errors.add("body", "fails SPAM checks") if body.match(r)
    end
  end

Add this to the Comment controller:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

...
    if @comment.save
      t = Thread.new do
        Notifier.deliver_comment_notification(@comment.entry.author, @comment)
        begin
          # Connect to the server and set presence to available for chat.
          client = Jabber::Simple.new('jlg_blog_bot@jabber.org/blog', JABBER_PASSWORD.to_s)
          client.deliver(
            @comment.entry.author.email, 
            "You have a new comment on your blog entry titled '#{@comment.entry.title}'",
            :normal
          )
          sleep(1)
          client.disconnect
          logger.info "Sent XMPP message about: '#{@comment.entry.title}'"
        rescue
          logger.warn "Sending of XMPP message failed..."
        end
      end # new thread
    else
      # If the comment wasn't saved, it was probably marked as spam
      flash[:warn] = "Your comment was not saved. Perhaps it was flagged as spam."
    end
...

You still need to modify the ‘config/environment.rb’ file to include definitions for JABBER_PASSWORD and SPAM_REGEXES.

This is my first entry in my Ruby on Rails weblog. This uses some very neat features that were sorely lacking in pretty much all the blog software I could find. Some of these are:

• Textile support (using RedCloth) for wiki-style syntax
• Coderay syntax highlighting
• Full support for tags
• Full-text search of all entries
• Live preview of entry addition via AJAX
• All pages are, as far I can tell, fully compliant with XHTML 1.0 Strict

That said, I think this is going to be exciting!

Here’s an example of the syntax highlighting, displaying the code that makes syntax highlighting work, no less:

1
2
3
4

# Markup parsing (both coderay and textile)
def unmarkup(text)
  return parse_coderay(textilize(text))
end

Let me know what you think!