Wolfmans Howlings

OpenMoko Freerunner after a few weeks

Jim Morris — Tue, 22 Jul 2008 13:20:08 -0700

Well I have had this thing for a few weeks now, and I have burned a lot of hours playing with it :) (Wish I could bill someone for those hours it would have paid for the phone 4 times over!)

I started with 2007.2 the built in image, and upgraded it initially with dfu-util, then with opkg update && opkg upgrade.

I installed a bunch of things and tweaked it to death (literally).

I got Jalimo java installed and working, and wrote the simple SWT app which worked nicely. The only change to the instructions are you need to add -force-depends to the opkg command.

As reported in my other blog article I got WIFI/WPA2 working, and finally got GPRS working.

I have not tried to pair with a bluetooth headset yet, although got far enough to see that the device could be seen using the scan command.

I played with all the varieties of on screen keyboards available, but really couldn't use any of them as they are so small, and my eyes are not that good anymore. I have seen some promising mentions of works in progress for keyboards, but the ones in the Qtopia distribution are the best, plus they have handwriting recognition which is a little fussy but looks like it can be trained.

The GPS problems were fixed by an amazing community and OpenMoko effort, which I have never seen before. There is a S/W fix, and a H/W fix which I suspect most people will not be able to do, soldering surface mount components is not for the faint of heart!

I spent several hours trying to build the OpenEmbedded and the MokoMakefile development environments, so I could start to contribute, but have still not been able to get them to finish building the native toolchain needed to build any of the apps.

I get several errors, some of which I found workarounds to by googling but eventually hit dead ends in both cases which I could not solve and have not been reported or solved by the community. I am using a stock KUbuntu 8.08 Hardy Heron desktop, which should be pretty mainstream, so I don't know what is wrong. The pre built toolchain OpenMoko provides for building apps does work however, but you can't build soft keyboards, or system components that route (at least I couldn't find a way to do it). I'll continue to hammer away at MokoMakefile and see if I can eventually get it going, of course I'll post my findings to the Wiki or dev mailing list if I succeed.

If someone could provide a VMPlayer image of a working OE or Mokomakefile environment that would help a lot of us. (Trolltech/Qtopia do provide exactly that for their development tools).

So after several opkg upgrades and tweaks, I finally killed my highly customized rootfs, X would no longer boot, sound was dead etc etc.

I tried to backup the image using these instructions, but always got an error halfway through, I suspect my flash image of rootfs may have been corrupted. Anyway I lost all that work :(

I tried flashing ASU, but it seems too much a work in progress more so than 2007.8.

Then I tried Qtopia now this is much more to my liking. The interface is clean, intuitive and seems to work pretty well. There are still a few rough edges which Trolltech seem to be taking care of, but it mostly works pretty well. The on screen input methods (which there are several of) are very good, even the keyboard has a nice touch where it zooms into the keys you are touching. It also has the tiny tiny QWERTY keyboard if you prefer that style and your eyes still work.

The downside is of course you lose access to all the applications currently under development for the GTK based 2007.2, but Qtopia does have a growing number of applications, and of course you can write your own.

I am going to experiment a little with trying to run GTK based apps under Qtopia, I know I can do that on my KDE (aka QT) desktop, so why not under Qtopia? I'll update this if I get it to work.

UPDATE Oh ok the reason you can't do that is that Qtopia is not running X Windows, so running GTK is not an option, bummer. I suppose you could render GTK into a QT canvas or something but that is more work than I am prepared to do at the moment.

The current show stoppers for me, which stops me being able to use this a phone are...

Nasty buzzing noise on both ends of the call
Bluetooth pairing with a headset not easily available.

I'm sure people are working on these issues, and I wait patiently for them to get fixed so I can dump my aging Motorola V600. I'd actually try to pitch in myself but I can't get the development environment to work!

Oh well guess I better get back to my paying job :)

Upgrading Ubuntu Gutsy to Hardy

Jim Morris — Sun, 13 Jul 2008 04:32:03 -0700

I did my duty and upgraded from Gutsy to Hardy, after letting Hardy settle for a while. For the most part it was painless (unlike the last upgrade to Gutsy!).

Unfortunately the sound was now broken I have a HDA-Intel AD198x Analog chip set.

Then I discovered (just when I needed it of course) that my Samsung ML-2010 USB printer didn't print anymore.

I would get this weird error...

Unable to open device hal:///org/freedesktop/Hal/devices/...

After Googling I found this, he was half right, I don't know why but it works, except that I needed to do...

sudo chmod 700 /usr/lib/cups/backend/usb
sudo chmod 700 /usr/lib/cups/backend/hal
sudo killall -HUP cupsd
sudo /etc/init.d/cupsys restart

ie both usb and hal.

That fixed it... Amazing!

Now why won't my audio work???

Ok now it does work, go figure. The only things I changed were added this to /etc/modprobe.d/alsa-base

options snd-hda-intel model=3stack-dig

and made sure that Front was checked in the mixer and turned up.

Not sure which of those fixed it though.

OpenMoko Freerunner first impressions

Jim Morris — Fri, 11 Jul 2008 13:53:11 -0700

OK so I just got my shiny new OpenMoko Freerunner GTA02.

This is an Open source GSM cell phone, running Linux and OpenMoko S/W stack.

This phone is really a developer release although it implies on the website it is usable for the masses.

We will see...

Turning it on you see all this Linux boot up tiny tiny text scroll by, ya the average Joe blo is going to love that ;)

So I followed the startup guide and plugged in the usb cable to my Ubuntu desktop. (If you don't have Linux think again, maybe that $400's can go towards a tank of gas).

I followed the usb networking guide, and ssh'd into the phone, and Holy Crap! it just worked!

First off it doesn't come with much, It can make a phone call, but no other apps seem to be loaded.

Oh one gripe is there is nowhere to stow a stylus, it comes with a huge pen/laser pointer/stylus, but who is going to keep that in their pocket, and those with fat fingers are going to have a hard time doing anything.

I actually had to RTFM to figure out how to close an application! BTW this guide seems to be pretty good. (You click the power button to close an app! it really needs a close button somewhere where people expect one).

After loading most of the needed apps via the opkg app, which is slow via USB 1.1 (whats up with that guys? its 2008! USB 2 is REQUIRED) I found that most of them are pretty much unusable. The default 2007.2 UI is unintuitive, and most of the necessary apps simply don't work well. Adding a contact is virtually impossible, try setting the phone number to home instead of work for instance.

The web browser works but you can't input anything into forms, so don't try logging into gmail.

The calendar makes no sense whatsoever.

OK so I am being harsh here, but the damn thing costs $400! and they imply your average user could use it as a phone, so lets get over that issue, and take it that this is a work in progress, very early stages, and open source. It needs a good UI guide, come on we may not have a super UI design guy, but copy the iphone or any other phone on the market to get a clue how to make an intuitive UI. I am not a UI guy, but even I can see this needs some work. BTW I do like the way the scrolling works, that is very cool, you flick up or down and it scrolls with some friction looks great, and works even better.

The platform itself has plenty of potential, built in GPS (which doesn't work without an external antenna), accelerometers, which no one has figured out how to use yet. A really nice VGA display, although its invisible in bright sunlight, wifi which I was totally unable to get to connect to a WPA/PSK station, and bluetooth, with no S/W support for headsets.

Once all these issues are resolved I think it would make a great primary handset.

I'll do my part and write something for it and release it as open source, but I'm disappointed I can't use this as a phone yet.

I'll continue to RTFM if I can find one that describes the current UI (the WIKI describes an obsolete UI). Maybe some of my issues are simply not knowing how to use the UI, but that is a bad sign as I am a developer and can use most UI's without RTFM'ing.

I still think this is a great (if expensive) toy for now, especially for us Linux Geeks, and one day it may actually be a great phone too.

GPRS

UPDATE I got GPRS working, followed the instructions on the WIKI, it is manual but it works. My tmobile chatscript is...

# File: /etc/ppp/chatscripts/tmobile
TIMEOUT 20
ABORT BUSY
ABORT "NO ANSWER"
ABORT "NO DIALTONE"
ABORT VOICE
ABORT ERROR
ABORT RINGING
SAY 'Starting GPRS connect script\n'
"" +++
OK ATZ
OK ATE1
OK AT+CFUN=1
OK AT+COPS
SAY 'Setting APN\n'
OK AT+CGDCONT=1,"IP","wap.voicestream.com"
ABORT 'NO CARRIER'
SAY 'Dialing...\n'
OK ATD*99***1#
CONNECT /n/d

and the script

# File: /etc/ppp/peers/tmobile
#
/dev/ttySAC0
115200
crtscts
lock
hide-password
defaultroute # set the default route
usepeerdns
holdoff 3
ipcp-accept-local
lcp-echo-failure 12
lcp-echo-interval 3
noauth
noipdefault
novj
novjccomp
replacedefaultroute
persist
debug

connect "/usr/sbin/chat -v -f /etc/ppp/chatscripts/tmobile"
disconnect "/usr/sbin/chat -v -f /etc/ppp/chatscripts/tmobile-disconnect"

and a script to run it all...

/etc/init.d/gsmd stop
echo "1" > /sys/bus/platform/devices/neo1973-pm-gsm.0/power_on
chown uucp.uucp /dev/ttySAC0
stty -F /dev/ttySAC0 crtscts
pppd call tmobile

GPS

I did get a fix with the GPS using the internal antenna, however it takes some concerted effort, see this page for hints. NOTE without the sd card I get a fix within 1 minute.

WIFI

UPDATE Got WIFI and WPA/PSK working, again there are two sets of instructions on the WIKI, the first set doesn't work for me, however the second set does... adding

iface eth0 inet dhcp
wpa-conf /etc/wpa_supplicant/wpa_supplicant.conf

to /etc/network/interfaces and then do ifdown usb0 and ifup eth0 of course you need to setup /etc/wpa_supplicant/wpa_supplicant.conf appropriately. However when I ssh into the phone via the wifi after disconnecting the USB, it looses its connection regularly and usually within a few minutes. Not sure why that is, my WIFI is generally very stable with my laptop. I wonder if the bluetooth is interfering with it?

Finally I would like to say that I love the idea of an open spurce cell phone. I can see that these issues will probably get fixed pretty quickly as it is open source and there are thousands of eyes on the code, and if there is something you don't like you can fix it. Lets see you do that on your I-Phone ;)

XPath matchers for rspec

Jim Morris — Wed, 02 Jan 2008 13:32:50 -0800

UPDATE I found this which is a nicer way of doing it using hpricot, which is faster. I have created a slightly different version that does have_xpath, for XML.

I've been working on a project that is mostly Java for the last many months, so haven't had much Ruby or Rails stuff to share.

However one thing I found when working on my tests in Java was an xpath matcher for JUnit 4.0 using the Hamcrest libraries.

When I dropped back into Ruby to write some cron scripts that process information from the database and generate xml files I wanted to check the script with an rspec, and check the xml files it was generating.

To do this I wanted to use a similar matcher to the Hamcrest ones, but use it in RSpec.

I Googled around and found a simple example, but it wasn't very sophisticated and didn't check what I needed so I upgraded it to do the kind of matches I needed, the results are here.


require 'rexml/document'
require 'rexml/element'

module Spec
  module Matchers

    # check if the xpath exists one or more times
    class HaveXpath
      def initialize(xpath)
        @xpath = xpath
      end

      def matches?(response)
        @response = response
        doc = response.is_a?(REXML::Document) ? response : REXML::Document.new(@response)
        match = REXML::XPath.match(doc, @xpath)
        not match.empty?
      end

      def failure_message
        "Did not find expected xpath #{@xpath}"
      end

      def negative_failure_message
        "Did find unexpected xpath #{@xpath}"
      end

      def description
        "match the xpath expression #{@xpath}"
      end
    end

    def have_xpath(xpath)
      HaveXpath.new(xpath)
    end

    # check if the xpath has the specified value
    # value is a string and there must be a single result to match its
    # equality against
    class MatchXpath
      def initialize(xpath, val)
        @xpath = xpath
        @val= val
      end

      def matches?(response)
        @response = response
        doc = response.is_a?(REXML::Document) ? response : REXML::Document.new(@response)
        ok= true
        REXML::XPath.each(doc, @xpath) do |e|
          @actual_val= case e
          when REXML::Attribute
            e.to_s
          when REXML::Element
            e.text
          else
            e.to_s
          end
          return false unless @val == @actual_val
        end
        return ok
      end

      def failure_message
        "The xpath #{@xpath} did not have the value '#{@val}'\nIt was '#{@actual_val}'"
      end

      def description
        "match the xpath expression #{@xpath} with #{@val}"
      end
    end

    def match_xpath(xpath, val)
      MatchXpath.new(xpath, val)
    end

    # checks if the given xpath occurs num times
    class HaveNodes  #:nodoc:
      def initialize(xpath, num)
        @xpath= xpath
        @num = num
      end

      def matches?(response)
        @response = response
        doc = response.is_a?(REXML::Document) ? response : REXML::Document.new(@response)
        match = REXML::XPath.match(doc, @xpath)
        @num_found= match.size
        @num_found == @num
      end

      def failure_message
        "Did not find expected number of nodes #{@num} in xpath #{@xpath}\nFound #{@num_found}"
      end

      def description
        "match the number of nodes #{@num}"
      end
    end

    def have_nodes(xpath, num)
      HaveNodes.new(xpath, num)
    end

  end
end

The first matcher HaveXPath was pretty much the original I found on the net, it simply checks that an XPath exists, I don't use this one.

The next matcher MatchXPath is more like what I was using in Java, it gets an element from the xpath and checks the value is equal to the one expected string. I will eventually add regex matching and arrays of strings or regexs to check against.

The last one HaveNodes I find handy to make sure a given xpath matches the expected number of nodes.

I even wrote an rspec to check the matchers, and this also is handy to show the way to use them.

Note that you can pass the matchers a String containing the XML or (much faster) a REXML::Document.

require 'matchers'

describe "test matchers" do
  before(:each) do
    @xml= <<-EOFXML
    <?xml version='1.0'?>
      <claims>
        <testnode1>
          <day>
            <rank order='1' value='0' userid='26' alias='user25'/>
            <rank order='2' value='0' userid='93' alias='user92'/>
            <rank order='3' value='0' userid='55' alias='user54'/>
            <sometext>this is text</sometext>
          </day>
        </testnode1>
      </claims>
    EOFXML

    @doc= REXML::Document.new(@xml)

  end

  it "should test xpath" do
    @xml.should have_nodes("/claims/*", 1)
    @doc.should have_nodes("/claims/*", 1)
    @doc.should have_nodes("/claims/testnode1/day/rank", 3)
    @doc.should have_xpath("/claims/testnode1/day/rank[@order='1']")
    @doc.should_not have_xpath("/claims/testnode1/day/rank[@order='10']")
    @doc.should match_xpath("/claims/testnode1/day/rank[1]/@order", "1")
    @doc.should match_xpath("/claims/testnode1/day/rank[2]/@value", "0")
    @doc.should match_xpath("/claims/testnode1/day/rank[3]/@alias", "user54")
    @doc.should match_xpath("/claims/testnode1/day/sometext", "this is text")
  end

end

Bit Vector Preferences

Jim Morris — Tue, 07 Aug 2007 23:53:09 -0700

In my latest web project I potentially have a lot of boolean preferences, which I use for enabling or disabling various email notifications to users.

Rather than having to add a migration everytime I want to add a new preference, I thought I would use the composed_of feature in my model and compose the boolean preferences from a bitvector. That way I can simply modify my model to add new preferences rather than add new columns to the database.

I also wanted this to be easy to add new boolean preferences, so I use some Macros (I guess you could also call it Meta-Programming) to do all the repetitive code.

The result is a little class in my Person model called Preferences, one integer field in my persons database called preferences, and a composed_of :preferences in the Person model, and of course the following class in the person.rb model.

class Person < ActiveRecord::Base

  #...

  # place to store bit vector preferences
  # to add a new preference:-
  #   add symbol of preference to @@bits with bit allocation
  #   update initialize defaults if initial default is true
  class Preferences
    @@bits= {
      :comment_notifications => 1,
      :friendship_notifications => 2,
      :event_notifications => 4,
      :misc_notifications => 8 }

    # create a reader for each preference
    @@bits.each_key do |a|
      attr_reader a
    end

    # Initialize from integer or Hash
    def initialize(prefs)
      if prefs.nil?
        # set the defaults to false if not been set before
        @@bits.each do |a, v|
          instance_variable_set("@#{a}", false)  
        end
        # override default here        
        @comment_notifications= true
        @event_notifications= true
      elsif prefs.is_a?(Hash)
        # initialize from parameter Hash, and default to false if absent from hash
        @@bits.each do |a, v|
          instance_variable_set("@#{a}", false)  
        end

        prefs.each do |k,v|
          raise(ArgumentError, "Unknown preference #{k}") unless @@bits.has_key?(k.to_sym)
          instance_variable_set("@#{k}", true) if v == '1'
        end
      else
        # create from integer bit vector
        @@bits.each do |a, v|
          instance_variable_set("@#{a}", (prefs & v) != 0 ? true : false)  
        end
      end
    end

    # returns bit vector of preferences
    def preferences
      bv= 0
      @@bits.each do |a, v|
        bv |= instance_variable_get("@#{a}") ? v : 0  
      end
      return bv
    end

    # create a predicate for each preference
    @@bits.each_key do |a|
      alias_method((a.to_s + '?').to_sym, a)
    end
  end

  # access preferences as bit vector
  composed_of :preferences

  #...

end

All I need to do to add new preferences is add it to the @@bits class variable, which is a Hash of the preference name as a symbol and the bit it sets in the integer (actually the value of the bit, bit0 is 1, bit1 is 2 etc). The rest of the code is derived from the class variable.

To make things easier I also add a predicate for each preference, so I can access @person.preferences.comment_notifications? to see if any comment notifications are required for instance.

One other thing I do in the initialize method is set up defaults for the preferences. This is only really needed if it is being added as an after thought, and the column preferences is NULL in the database.

Because composed_of classes are immutable you must always create a whole new one to update them, so I also allow initialize to be called with a Hash, which can come straight from the controller. The last case of initialize is being passed the integer from the database, expanding it into the various boolean instance variables. The preferences method does the reverse and converts the boolean instance variables into the bit vector. Calls to these are all taken care of by ActiveRecord.

An example of it being called from the controller is...

@person.preferences= Person::Preferences.new(params[:preferences])

presuming you have a bunch of check boxes in your view which are passed in as part of the preferences hash.

Because I added this later I had one migration to initially add the new column...

add_column :people, :preferences, :integer

Paginating acts_as_taggable with will_paginate

Jim Morris — Mon, 30 Jul 2007 15:04:50 -0700

A question I see asked a lot is how do I paginate acts_as_taggable (on steroids)?

I haven't seen any answers I liked, so I created my own, which I'm sure a few people won't like either ;) But it works for me (tm).

I use will_paginate, but this does not work with custom finds that plugins define themselves, as is the case with acts_as_taggable.

If you do use a custom find_by_sql you have to hit the database twice, once to find the total number of items and then the paginated find.

I have a situation where I generate a tag cloud with every page, and part of that tag cloud has already calculated the number of tags for each classification I use.

I combined this with the paginator gem to get myself pages without too many hits to the database.

The first thing I do is pass the total count I get from the tag cloud to the action that renders the index for all those items matching the cloud... If you look at the article cited above I make this modification, this is dumbed down a bit for the sake of simplicity...

  tags= Post.tag_counts(:order => 'tags.name')
  tags.each do |t|
    link_to(h(t.name), tagged_post_path(:tag => t.name, :cnt => t.count))
  end

This passes the count I have already calculated to the action that will list the paginated results.

In my controller I do this to get the paginated results using will_paginate and the Pagination gem...

    def tagged
    ...
      # page if we can
      size= params[:cnt]
      if size
        per_page= 10

        # use Paginator gem to do the actual paging
        pager = ::Paginator.new(size, per_page) do |offset, per_page|
          Post.find_tagged_with(params[:tag], :limit => per_page, :offset => offset)
        end

        # default to page 1 if not specified
        page= params[:page] || 1

        # gets a paged array of posts
        @posts= returning WillPaginate::Collection.new(page, per_page, size) do |p|
          p.replace pager.page(page).items
        end
      else
        # fall back if we don't know the size
        @posts= Post.find_tagged_with(params[:tag])
      end

      render :action => 'index'
    end

This fits in nicely with the tag cloud I need to calculate, and it uses will_paginate just like the regular index action does.

UPDATE

I refactored this to be more generally useful, I added the following as a protected method in application.rb...

  # paginate a call to find_tagged_with
  # klass is the tagged class
  # tag is the tag to find
  # count is the total number of items with that tag, if nil count_tags is called
  # per_page is numbe rof items per page
  # page is the page we are on
  # order is the order to return the items in
  def tag_paginator(klass, tag, count=nil, per_page=10, page=1, order='updated_at DESC')
    count ||= klass.count_tags(tag)
    pager = ::Paginator.new(count, per_page) do |offset, per_page|
      klass.find_tagged_with(tag, :order => order, :limit => per_page, :offset => offset)
    end

    page ||= 1

    returning WillPaginate::Collection.new(page, per_page, count) do |p|
      p.replace pager.page(page).items
    end
  end

I call it from one of my other actions like this...

@faqs= tag_paginator(Post, 'FAQ', nil, per_page, params[:page], 'updated_at DESC')

Passing in nil as the third parameter causes the tag_paginator method to call Post.count_tags which is not part of the acts_as_taggable methods, I added it to the SingletonMethods module myself...

module ActiveRecord
  module Acts #:nodoc:
    module Taggable #:nodoc:
      module SingletonMethods

      ...

        # Return the count of tag tags in this class
       def count_tags(tag)
         count_by_sql("select count(*) FROM tags, taggings WHERE " + sanitize_sql(['name = ? AND tags.id = taggings.tag_id AND taggable_type = ?', tag, name]))
       end

      ....

If you don't want to hack acts_as_taggable then simply leave that part out and call the count_by_sql yourself.

My refactored tagged action from above now looks like this...

def tagged
  ...
  tag= params[:tag]
  per_page= 10
  size= params[:cnt]
  @posts= tag_paginator(Post, tag, size, 10, params[:page])

RSpec testing all actions of a controller

Jim Morris — Sat, 28 Jul 2007 14:23:07 -0700

A pattern I find very helpful is to find all the actions in a controller and apply a test to all those actions.

For instance this is useful for automatically testing all actions are protected from unauthorized access when using a login system.

One nice feature of this pattern is that if you add an action to a controller it will automatically be tested. This is less helpful if you use

before_filter :login_required, :except => {...}

as it will automatically be protected, but there are other use cases where this is not the situation. Just as in the except clause above you need to explicitly add any action that does not need to be tested to an exception list, which is supported by this pattern.

Here are the methods I use to test for login accessibility.

module MySpecHelper

  # get all actions for specified controller
  def get_all_actions(cont)
    c= Module.const_get(cont.to_s.pluralize.capitalize + "Controller")
    c.public_instance_methods(false).reject{ |action| ['rescue_action'].include?(action) }
  end

  # test actions fail if not logged in
  # opts[:exclude] contains an array of actions to skip
  # opts[:include] contains an array of actions to add to the test in addition
  # to any found by get_all_actions
  def controller_actions_should_fail_if_not_logged_in(cont, opts={})
    except= opts[:except] || []
    actions_to_test= get_all_actions(cont).reject{ |a| except.include?(a) }
    actions_to_test += opts[:include] if opts[:include]
    actions_to_test.each do |a|
      #puts "... #{a}"
      get a
      response.should_not be_success
      response.should redirect_to('http://test.host/login')
      flash[:warning].should == @login_warning
   end
 end
end

I put this in my spec_helper.rb and include it as shown here:

describe "When Logged out" do
  include MySpecHelper
  controller_name :events

  before(:each) do
    controller.stub!(:current_user).and_return(:false)
    @login_warning= "You need to be logged in to do that"
  end

  # test all actions require login except the ones specified
  # add new_comment as it is not seen by the automatic collector
  it "actions should fail" do
    controller_actions_should_fail_if_not_logged_in(:input, 
                              :except => ['index', 'show', 'tagged'], 
                              :include => ['new_comment'])
  end
end

The get_all_actions method collects all the public un-inherited methods in the given controller, these will consist of all the accessible actions in that controller. I explicitly exclude rescue_action as it is created by RSpec itself and should not be tested. Note it will not see any actions that are in application.rb so you need to add those to the list manually of you want them tested. (See the :include option in the example).

The controller_actions_should_fail_if_not_logged_in could be put in the spec itself rather than the spec_helper, but as I call this from all my controller specs it is more DRY to put it here. This method takes the controller name and an option array of actions names to ignore. This method tests all the actions and makes sure I get the expected result of the filter failing due to not being logged in.

I show an example spec that uses this to test my events controller, it mocks the login calls to say I am not logged in, and then tests them with the exceptions of the actions in this controller that do not require one to be logged in.

This pattern can be extended to test all sorts of things, and is especially useful for testing things where you can add an action and forget to do something in a filter to protect it. Make sure the default is on the side of caution though. IE you need to explicitly except actions rather than include actions.

Another example is something I recently stumbled upon in my RESTful controllers. In many cases it is good to use a verify statement to make sure that the RESTful actions actually can only be called with PUT, POST or DELETE and fail if called with GET. I use this statement in my controllers to enforce this...

# GETs should be safe (see http://www.w3.org/2001/tag/doc/whenToUseGet.html)
verify :method => :put, :only => [ :update ], :add_flash => { :error => "Operation Failed" }, :redirect_to => { :action => :index }
verify :method => :post, :only => [ :create, :new_comment ], :add_flash => { :error => "Operation Failed" }, :redirect_to => { :action => :index }
verify :method => :delete, :only => [ :destroy ], :add_flash => { :error => "Operation Failed" }, :redirect_to => { :action => :index }

I test this in my specs using this in the MySpecHelper Module

  def controller_actions_should_fail_with_get(cont, except=[])
    actions_to_test= get_all_actions(cont).reject{ |a| except.include?(a) }
    actions_to_test.each do |a|
      #puts "... #{a}"
      get a
      response.should redirect_to("http://test.host/#{cont.to_s.pluralize}")
      flash[:error].should == 'Operation Failed'
    end
  end

and an example of its use in a spec...

  it "actions should fail if not post or put" do
    controller_actions_should_fail_with_get(:event, ['index', 'show', 'edit', 'new'])
  end

Now whenever I add an action, the default is that it will fail with a GET, unless I add it to the exclude list in the spec, this will remind me to check if the action required PUT, POST or DELETE instead and to add it to the verify if so or add it to the specs exclude list if not.

These automatic tests keep me honest, especially in the last case where you really don't want a GET to be able to delete something.

I hope this pattern is useful to you.

Using RSpec to test HAML helpers

Jim Morris — Sat, 14 Jul 2007 16:40:00 -0700

UPDATED for HAML 2.0 and RSpec 1.1.5 - Changed open to haml_tag, prefix helper. to all rspec calls...

The most recent release of HAML introduced a neat feature that allows you to use HAML-like syntax in your helpers to generate HTML HAML#haml_tag.

A question on the HAML news group asked how to test a helper that uses HAML#haml_tag (used to be open/puts) and thanks to Nathan on that list I finally got RSpec to do it. As shown below.

However a really good point was made that really in RSpec the way to test anything is to use mocks to mock any call to an outside method thus focusing the test on the specific module under test. Generally I agree with that philosophy. But this is way cool so I thought I'd do it anyway, and also as it is a new feature in HAML one may not want to simply trust HAML to generate the correct HTML.

So in my application_helper.rb I have a simple helper...

module ApplicationHelper

 ...

  def display_flash
    for name in [:notice, :warning, :error]
      if flash[name]
        haml_tag :div, flash[name], {:class => name.to_s}
      end
    end
    nil
  end

  ...

end

This is called in my views as...

- display_flash

Notice the - instead of =, this is because the open (and puts) write output directly to the HAML buffer, and so this routine should return nothing. (This is also a very simply case and does not show off the utility of the open/puts methods, I'll show one of those later on).

The RSpec helper test that tests this is as follows...

# File: spec/helpers/application_helper_spec.rb
require File.dirname(__FILE__) + '/../spec_helper'

describe ApplicationHelper do

  before :each do
    helper.extend Haml
    helper.extend Haml::Helpers 
    helper.send :init_haml_helpers
  end

  it "should display flash" do
    for name in [:notice, :warning, :error]
      flash[name]= "flash #{name.to_s} message"
      helper.capture_haml{
        helper.display_flash
      }.should =~ /<div class='#{name.to_s}'>\s*#{flash[name]}\s*<\/div>/
      flash[name]= nil
    end
  end

end

Excellent, a simple test for the HTML generated by my haml helper. NOTE the setup required in the before :each, this sets up the haml helpers in the helpers context

Why would I want to use HAML#open you ask?

Well it makes the helpers look so much tidier IMHO, take this example from my previous post on tag clouds, the re-factored helper now looks like this...

  # display a tag cloud for the given model
  def tag_cloud(model, title= nil)
    m= model.to_s.camelcase.constantize
    plural= model.to_s.capitalize.pluralize
    title ||= plural
    tags= m.tag_counts(:order => 'tags.name')
    return false if tags.empty?
    urlmeth= "tagged_#{model.to_s.pluralize}_path".to_sym
    haml_tag :div, {:class => "tagcloud"} do
      haml_tag :h3, title
      tags.each do |t|
        next if t.name == 'FAQ'
        haml_tag :span, {:style => "font-size:#{calc_size(t.count)}%"} do
          puts link_to(h(t.name), self.send(urlmeth, :tag => t.name))
        end
      end
    end
    return true
  end

So much cleaner, plus I can return a boolean to indicate if there was anything output or not, which tells me if I need to output an <hr/> or not.