taras kalapun

Generating API - teaser

2013-03-01T19:35:00+02:00

Imagine you'r defining how Web-service API should look like in a simple YAML file

baseURL: "http://localhost"
fixtureDirectory: ""

defaults:
  request:
    method: "GET"
  response:
    responseCode: 200
    headers:
      Content-Type: application/json

mocks:
  -
    request:
      url: "/api/v1/catalog/products"
    response:
      fixture: "catalog_products.json"
      timeout: 60.0
  -
    request:
      url: "/api/v1/pickups"
    response:
      fixture: "pickups.json"

Now imagine you make a stubs from it, in runtime, on iOS ans Android.. and Ruby…

Or you generate API documentation…

Or just a few diagrams, to show your boss how it works…

./parser.rb api.yml > seq.diag
seqdiag -Tsvg seq.diag > seq.svg
./svg2png.sh seq.svg

And you get a Sequence diagram!

And after a few additional tricks…

Header 1	Header2
kjfek	kdfjnvkerv

./svg2scruffy.rb seq.svg > seq_scruffy.svg

You can get a fun "hand-drawn" diagram

Stay tuned…

Rails - SQLite and PostgreSQL playing nice together

2013-02-12T16:57:00+02:00

So I was asked to add few features to the web-app, that was in production and running PostgreSQL. Well, I didn't want to install the PostgreSQL on my local machine. So I decided to have development version and test version to run in SQLite, while keeping production version unchanged.

So I came up to few hacks.

database.yml

First, edit your config/database.yml and change it so development and test enviroment will use SQLite

development:
  adapter: sqlite3
  database: db/development.sqlite3
  pool: 5
  timeout: 5000
test: &test
  adapter: sqlite3
  database: db/test.sqlite3
  pool: 5
  timeout: 5000

And don't forget to run rake db:reset and rake db:seed to set your new DB and seed it with data.

Gemfile

Don't forget to edit your Gemfile

group :development, :test do
  gem 'sqlite3'
end

group :production do
  gem 'pg'
end

SQL stuff

PostgreSQL has some SQL addons, that SQLite doesn't. Some of them can be changed to DB-agnostic style, and some of them - can be skipped for development.

As the first example, we have a query like

Thing.where("#{key} ilike ?", "%#{value}%")

You see this iLike keyword, don't you? Well the thing is, putting DB-agnostic query is even faster than native ILIKE. So change it to lower(x) like lower(y):

Thing.where("lower(#{resource_class.table_name}.#{key}) LIKE lower(?)", "%#{value}%")

Next one is INDEX. PostgreSQL has an addition with WHERE, but SQLite doesnt. But for development purposes, we don't care.

  if Rails.env.development? || Rails.env.test?
    execute "CREATE UNIQUE INDEX websites_qa_profile_url ON websites (qa_profile_id, url)"
  else
    execute "CREATE UNIQUE INDEX websites_qa_profile_url ON websites (qa_profile_id, url) WHERE deleted_at IS NULL"
  end

This will use the proper SQL query depending on enviroment it runs.

Dates

Next thing, PostgreSQL has his own dating formats like now() - '5 days'::interval, which SQLite doesn't have. So I've made this class

class AdapterSpecific

  class << self
    # Workaround for "updated_at > now() - '5 days'::interval"
    def datetime(timestring, modifier)
      res = ''
      case ActiveRecord::Base.connection.adapter_name
      when 'PostgreSQL'
        timestring = 'now()' if timestring == 'now'
        if modifier[0] == '-'
          add = '-'
          modifier[0] == ''
          modifier.strip!
        else
          add = '+'
        end
        "#{timestring} #{add} '#{modifier}'::interval"
      else
        "datetime(#{timestring}, #{modifier})"
      end
    end

    def before(modifier)
      case ActiveRecord::Base.connection.adapter_name
      when 'PostgreSQL'
        "now() - '#{modifier}'::interval"
      else
        "datetime('now', '-#{modifier}')"
      end
    end

  end
end

Put it as lib/adapter_specific.rb, edit your config/initializers/application.rb and add on top require 'adapter_specific'. Then change your SQL queries to be like one below:

Thing.where("updated_at > #{AdapterSpecific.before('5 days')}").limit(20)

Hope this short post will help you in playing with RoR apps.`

Reverse-engineering iOS app

2013-02-11T18:02:00+02:00

On one of projects I was involved in, we were implementing Captive Login technology, to automatically log in networks that have Captive Portal via WISPr technology. Everything went quite fine, except that on iOS, you can't do auto-login to WiFi network without user's permission. And the CaptiveLogin support is quite limited.

Then we suddenly spotted an app, that could do all that cool things. That was the start of my small research.

The Beginning Is the End Is the Beginning

For sake of convinience, let's call tha app that we'r interested in iApp. We downloaded the app from AppStore, renamed .ipa to .zip, and started looking inside.

First of all, we opened Info.plist. And found a quite interesting string network-authentication for Required background modes. Google gave us 0 results.

Next thing, I checked iApp.entitlements file. And found there a nice key com.apple.developer.CaptiveNetworkPlugin saying YES. That was something.

Disassembling

I wanted to check if the app uses some privete libraries or methods related to CaptiveNetwork. To find that out, I used Hopper Disassembler. For checking the app disassembly, the Demo version is quite good.

After some digging I found a notice of using some Plugin

That gave me idea of usage of some private CaptiveNetwork library. I found one in /Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS6.1.sdk/System/Library/PrivateFrameworks/CaptiveNetwork.framework (similar path to ARM version)

nm -g CaptiveNetwork gave me an interesting list of public methods (one starts with T):

00001e78 T _CNAccountsAdd
00001e40 T _CNAccountsCopy
00002018 T _CNAccountsResolve
00001cb4 T _CNAccountsUse
000022e4 T _CNCheckControlSettings
00001ad0 T _CNDebugLaunchWebsheet
000022c4 T _CNDumpState
00001bf4 T _CNForgetSSID
00002074 T _CNIAmTheWebsheetApp
00001b60 T _CNLogoff
000049c0 T _CNNetworkCopyPluginNames
00004890 T _CNNetworkGetBSSID
00004858 T _CNNetworkGetSSID
00004820 T _CNNetworkGetSSIDString
00004908 T _CNNetworkGetSignalStrength
00004744 T _CNNetworkGetTypeID
000048c8 T _CNNetworkIsProtected
00004770 T _CNNetworkSetCaptive
000047e0 T _CNNetworkSetConfidence
00004970 T _CNNetworkWasAutoJoined
000050e0 T _CNPluginCommandBindReadStream
00005110 T _CNPluginCommandCopyCurrentNetwork
0000512c T _CNPluginCommandCopyNetworkList
000050d0 T _CNPluginCommandGetInterfaceName
000050cc T _CNPluginCommandGetType
000050a0 T _CNPluginCommandGetTypeID
00004bbc T _CNPluginRegister
00005290 T _CNPluginResponseCreate
00005300 T _CNPluginResponseDeliver
00005264 T _CNPluginResponseGetTypeID
000054b8 T _CNPluginResponseSetNetwork
0000543c T _CNPluginResponseSetNetworkList
00001798 T _CNProberCreate
000019dc T _CNProberCreateRunLoopSource
000016f0 T _CNProberGetTypeID

which looked quite interesting. So not thinking for so long, I opened the library in disassembly:

Hmm, they are getting the info from some NSDictionary ?

Putting stuff together

Well, I tried making a demo app and using that private framework.

But how do we get access to it's C-functions? Her's the answer:

// Get the bundle framework
CFBundleRef bundle = CFBundleGetBundleWithIdentifier(CFSTR("com.apple.CaptiveNetworkSupport"));

// Declare the function

// let's start with some simple one
BOOL (*CNNetworkWasAutoJoined)(CFDictionaryRef) = NULL;
    CNNetworkWasAutoJoined = (void *)CFBundleGetFunctionPointerForName(bundle, CFSTR("CNNetworkWasAutoJoined"));

// register us in networks
CFStringRef ssids[2] = { CFSTR("OurWiFi"), CFSTR("NeighboursWiFi") };
    CFArrayRef arr_ssids = CFArrayCreate(NULL, (const void **)ssids, 2, &kCFTypeArrayCallBacks);

    if( CNSetSupportedSSIDs((CFArrayRef)arr_ssids)) {
        NSLog(@"Successfully registered supported network SSIDs");
    } else {
        NSLog(@"Error: Failed to register supported network SSIDs");
    }
    CFRelease(arr_ssids);

// Find network interfaces by CNCopySupportedInterfaces()
NSString* interface = @"en0";  // or hard code it

// Get network details
NSDictionary* networkDetails = (__bridge NSDictionary*)CNCopyCurrentNetworkInfo((__bridge CFStringRef) interface);

// check it?
NSLog(@"all details: %@", [networkDetails description]);

// Let's use the private method
BOOL x = CNNetworkWasAutoJoined((__bridge CFDictionaryRef)(networkDetails));
NSLog(@"X: %d", x);

Now everything seems working till the last one. If we'll check debugger, we're getting the bundle, we're getting the private function in CNNetworkWasAutoJoined, but we'r getting crash when we try using it. It seems the networkDetails doesn't contain the needed information.

The End Is the Beginning Is the End

Well, idea of putting com.apple.developer.CaptiveNetworkPlugin to Entitlments did'nt work. There need to be a valid Provisioning profile for this thing. And the valid one can be issued only by Apple.

So it brings us to idea that the creators of this iApp got in contact with Apple, and got a special priviledges for their app. Well, we also contacted Apple, but still no response from their side.

iOS Unit testing

2013-02-11T16:10:00+02:00

Let me tell you, dear reader, how I simplify my life by writing small and effective unit tests for iOS projects.

I'm quite lazy person, so if I can simplify my life and work - I do so. Of course you agree, that writing tests is cool. Having tests means you have less bugs in code, and if someone changes something, he can check if he's patch breaks anything in existing implementation. And of course, writing unit tests makes you thinking about better architecture, and making smaller, better methods. You know beforehand, that testing a method that is few hundreds of lines of code is a big pain in the ass…

But why people often avoid writing tests? Well, as I see often - because they think it's a lot of additional code to write.

So I'll try to tell from my experience, how to write less test code :)

Asserts

Let's start with simple thing, thing that you can spot in many test samples:

STAssertNotNil(objects, @"Could not load objects");

Well, if it's one line in code - it's kind of OK. But if you need to write these STAssers-s over 9000 times… you know what I mean. It's getting even worse if you want to compare simple types like int or float.

So my number one pick is a small library called Expecta. Just add to your Podfile (I think you'r using cocoapods, don't you?) pod 'Expecta' and you'll have a simple and nice assert tests like

expect(objects).toNot.beNil();
expect(objects.count).to.equal(3);
expect(evaluated).to.equal(YES);

// or more fancy stuff
NSDictionary *dict2 = @{@"id" : @"articleId", @"name" : @"name", @"title" : @"title"};
expect(dict).to.equal(dict2);

HTTP requests

A lot of developers find it hard to test their API-classes and network-related code. Well, I'll give you some tips on it.

I'm using Nocilla library. (pod 'Nocilla'). It's a small library that can stub some network requests, and return back stuff like Headers, response code, etc. It's in active development, so if you want to refer to latest code, you should put in your Podfile pod 'Nocilla', podspec: 'https://github.com/luisobo/Nocilla/raw/master/Nocilla.podspec'

Some code samples:

NSString *data = @"Some data here";
stubRequest(@"GET", @"http://localhost/articles").
andReturn(200).withHeaders(@{@"Content-Type": @"application/json"}).withBody(data);

Now if you'll make a normal NSURLRequest to http://localhost/articles you'll get a response code 200, Header Content-Type: application/json and, of course, your piece of data :). Oh, and don't forget to initialize and clear the stubs:

- (void)setUp {
    [[LSNocilla sharedInstance] start];
}
- (void)tearDown {
    // Clear Network stubs
    [[LSNocilla sharedInstance] clearStubs];
    [[LSNocilla sharedInstance] stop];
}

Asynchronous testing

Network tests have delays. yep. and you need to deal with them. And your tests need to wait for it. Same for blocks. Methods, where you need to check return values in blocks need asynchronous testing.

In internet, you can find a lot of examples how to do that. The one I've used is

dispatch_semaphore_t semaphore = dispatch_semaphore_create(0);
[Thing doSomethingOnSuccess:^(NSArray *objects) {
        expect(objects).toNot.beNil();
        dispatch_semaphore_signal(semaphore);
    } failure:^(NSError *error) {
        dispatch_semaphore_signal(semaphore);
    }];
while (dispatch_semaphore_wait(semaphore, DISPATCH_TIME_NOW)) {
        [[NSRunLoop currentRunLoop] runMode:NSDefaultRunLoopMode
                                 beforeDate:[NSDate dateWithTimeIntervalSinceNow:2]];
}

Writing this once - kind of OK. But again, we can make it simpler. Personally, I've created a small library for that. TKSenTestAsync. Just add pod 'TKSenTestAsync' and the code above can be re-written as

[self runTestWithBlock:^{
        [Thing doSomethingOnSuccess:^(NSArray *objects) {
        expect(objects).toNot.beNil();
         [self blockTestCompleted]; // required
    } failure:^(NSError *error) {
        [self blockTestCompleted]; // required
    }];
}];

looks much more clear?

CoreData and.. yeah, RestKit stuff

Well, currently I'm using RestKit a lot, and a nice thing about it, that it has few nice helpers for Unit testing.

First of all, you need to add it's testing sub-library. Currently, due to some bug in main cocoapods spec, you need to write in your Podfile:

pod 'RestKit', podspec: 'https://github.com/RestKit/RestKit/raw/development/RestKit.podspec'
pod 'RestKit/Testing', podspec: 'https://github.com/RestKit/RestKit/raw/development/RestKit.podspec'

Now, before starting, check that you have included Your.xcdatamodeld in Unit tests target.

Setting up stuff:

    RKManagedObjectStore *store = [RKTestFactory managedObjectStore];
    [RKManagedObjectStore setDefaultStore:store];

Now I found a small bug in this step… in my case, a directory was missing in iPhone Simulator, so the above code was failing. To fix this, just run following before

+ (void)checkPathForCoreDataFile {
    NSString *path = RKApplicationDataDirectory();
    NSError *error = nil;
    BOOL isDir = YES;
    NSFileManager *fm= [NSFileManager defaultManager];
    if (![fm fileExistsAtPath:path isDirectory:&isDir]) {
        if (![fm createDirectoryAtPath:path withIntermediateDirectories:YES attributes:nil error:&error]) {
            NSLog(@"Error: Create folder failed");
        }
    }
}

Now, let's test it:

- (void)testCoreDataIntegration {
    RKManagedObjectStore *managedObjectStore = [RKManagedObjectStore defaultStore];

    expect(managedObjectStore).toNot.beNil();

    NSManagedObjectContext *moc = managedObjectStore.persistentStoreManagedObjectContext;
    expect(moc).toNot.beNil();

    Article *mo = [RKTestFactory insertManagedObjectForEntityForName:@"Article" inManagedObjectContext:moc withProperties:nil];
    mo.title = @"SomeObject";

    [moc saveToPersistentStore:NULL];

    expect(mo).toNot.beNil();

    NSFetchRequest *fetchRequest = [[NSFetchRequest alloc] initWithEntityName:@"Article"];

    NSError *error = nil;
    NSArray *objects = [moc executeFetchRequest:fetchRequest error:&error];

    expect(objects).toNot.beNil();
}

Putting it all together

I can write a lot of code here, but the easiest thing is to point you to the library I'm currently developing, which includes all this tests: RKInjective

You should check the RKInjectiveTests folder contents.

My next article will be about testing a small game engine using Kiwi test framework.

The end.

Well, I hope you liked the stuff I covered here. I hope it can help you in your work. Please leave your comments, tweets, etc. I'll be happy.

Routes and embedded documents

2012-08-30T14:40:00+03:00

Following my previous post Mongoid custom relation between embedded documents

Interesting how to set-up Routes and MVC for embedded documents? I'll try to tell you how I did it.

As I mentioned in previous post, my main document Event looks like

class Event
  # cutted
  embeds_many :acts
  embeds_many :areas
end

Routing

In routes.rb

  resources :events do
    resources :acts
    resources :areas
  end

Now we can access the Acts and Areas with urls like /events/:event_id/acts/, /events/:event_id/acts/:id, and same for areas

To see the all the routes you can use command rake routes in terminal

Controllers

In all controllers for embedded documents, we need to have access to parent document. We can achieve this by adding before_filter to our controllers.

class ActsController < ApplicationController
  before_filter :check_event!

    def check_event!
    @event = Event.find(params[:event_id]) rescue nil

    if !@event
      redirect_to root_path, :alert => "Event not found!"
    end
  end

And also we need to change the actions to use our parent document

  def index
    @acts = @event.acts.all
    # Respond code
  end

  def show
    @act = @event.acts.find(params[:id])
    # Respond code
  end

Also, we need to change the creation method, so the child document will be added to the parent

  def create
    @act = Act.new(params[:act])

    respond_to do |format|
      if @event.acts << @act
        @event.save
        format.html { redirect_to [@event, @act], notice: 'Act was successfully created.' }
        # Other format code
      else
        # Else code
      end
    end
  end

Notice the adding code @event.acts << @act and the redirect code redirect_to [@event, @act]

Views

For every link_to in your view, you need to add relation to Event document, and also some paths needs changing.

= link_to 'Show', [@event, @act]
= link_to 'Edit', edit_event_act_path(@event, act)
= link_to 'New Act', new_event_act_path

For forms, I'm using simple_form gem, but basicaly the only thing you need to change is to put the @event there

= simple_form_for([@event, @act], html:{class:'form-horizontal'}) do |f|
  # inputs go here

I hope this short post will help you getting started…

Mongoid custom relation between embedded documents

2012-08-29T16:40:00+03:00

Imagine we have a Mongoid document, where we have two embedded documents. And now we want to make some relation between them. What you can think of at first is to use

has_many :acts, :foreign_key => 'event_id'

But that won't work. Mongoid is not capable at this moment to support this between embedded documents :(

But we can try to implement our own, custom relationship.

WARNING: This post is only my attempt to make it happen. It can contain bugs, or there can be other ways how to make it work easier. If you found some - let me know.

Overview

So let's say we have a main document called Event.

class Event
  include Mongoid::Document
  field :name, type: String
  embeds_many :acts
  embeds_many :areas
end

As you see, it embeds documents Act and Area.

class Act
  include Mongoid::Document
  include Mongoid::MultiParameterAttributes

  field :title, type: String
  field :start_at, type: DateTime
  field :end_at, type: DateTime

  default_scope asc(:start_at)

  embedded_in :event

and

class Area
  include Mongoid::Document

  field :name, type: String

  embedded_in :event

Ralationship

Let's add some custom relationship:

In Act we want to relate to one Area, so we need an field :area_id, type: String to track that. In Area document we will need to track array of acts, so we'll add field :act_ids, type: Array

Now, let's add some custom relationship management to Act document

  def area
    self.event.areas.find(self.area_id) rescue nil
  end

  def area_name
    self.area.name rescue nil
  end

  def area_id=(value)

    act_id = self.id.to_s

    area = self.area
    if area.nil?
      area = find_area_by_id(value)
    end

    return super(value) if area.nil?

    if value.blank?
      area.pull(:act_ids, act_id) # remove old one        
    else
      area.pull(:act_ids, act_id)
      area = find_area_by_id(value)
      area.add_to_set(:act_ids, act_id)
    end

    super(value)
  end

  def find_area_by_id(value)
    self.event.areas.find(value) rescue nil
  end

  #around_destroy
  before_destroy do |document|
    #Handle callback here.
    area = self.area
    area.pull(:act_ids, self.id.to_s) if area
  end

And for Area

  def act_ids=(values)
    values.delete('')

    old_values = self.act_ids || []
    old_values = old_values - values
    old_values.each do |act_id|
      act = find_act_with_id(act_id)
      act.unset(:area_id) if act
    end

    new_values = (values - old_values) || []

    new_values.each do |act_id|
      act = find_act_with_id(act_id)
      if act
        if !act.area_id.blank?
          self.event.areas.find(act.area_id).pull(:act_ids, act_id)
        end
        act.set(:area_id, self.id.to_s)
      end
    end
    super(values)
  end

  def find_act_with_id(value)
    self.event.acts.find(value) rescue nil
  end

  def acts
    self.event.acts.find(self.act_ids || []) rescue []
  end

  def acts=(value)
    self.act_ids = value.map(&:id)
  end

  before_destroy do |document|
    #Handle callback here.
    self.act_ids.each do |act_id|
      act = find_act_with_id(act_id)
      act.unset(:area_id) if act
    end
  end

Again, this post is just my attempt to make it work. If you have any comments how to improve this - let me know

Userscripts and Apple dev site

2012-07-13T11:34:00+03:00

If you are a pro developer for Apple stuff (ios, mac), you may have a situation, when you are in different teams, and sometimes, the team name can be same. For example, there can be a team Company for AppStore and team Company for Enterprise. So I was bored gessing which one to choose and decided to make a Greesemonkey script for that.

Usage

If you are using Safari, install small extension NinjaKit, for adding support of greesemonkey scripts

https://github.com/os0x/NinjaKit

Now point your browser to my script on userscripts.org and install it

http://userscripts.org/scripts/show/138339

The Code

Well, if you wanna know the code behind the script, here you go:

// ==UserScript==
// @name        Apple Developer Team select Rename
// @description Changes Team name by adding Team ID
// @author      Taras Kalapun 
// @include     http*://developer.apple.com/devcenter/selectTeam*
// ==/UserScript==
teams = document.getElementById("teams")
for (var i=0; i<teams.length; i++){
  teams.options[i].text = teams.options[i].text+" ("+teams.options[i].value+")"
}

Generating API docs from RSpec tests

2012-06-01T14:32:00+03:00

In this small post I'll show how I use RSpec to test and generate docs for my API.

Full documentation for the gem and it's DSL you can find on it's homepage rspec_api_documentation

First of all, you will need nice gem rspec_api_documentation, so edit your Gemfile and add

gem 'rspec_api_documentation', :group => [:test, :development], :git => "git://github.com/zipmark/rspec_api_documentation.git"

Rake helper task

Before we proceed, Let's create rake task for testing API without generation lib/tasks/spec_api.rake

require 'rspec/core/rake_task'

desc 'Run API specs test in spec/acceptance'
  RSpec::Core::RakeTask.new('spec:api') do |t|
  t.pattern = 'spec/acceptance/**/*_spec.rb'
  #t.rspec_opts = ["--format RspecApiDocumentation::ApiFormatter"]
end

Now if we will run rake spec:api it will run tests in files spec/acceptance/*_spec.rb for testing API without generationg docs. So don't forget to create that folder.

Spec helper & override

Also, I create spec/api_helper.rb that contains some helpers for my tests.

def prefill_db_for_api
  puts "Creating DB data"

  # Some factory_girl DB prefilling

end

# Overrides for customizing API docs generator

module RspecApiDocumentation
  class RackTestClient < ClientBase

    def request_headers
      my_headers = env_to_headers(last_request.env)
      my_headers.reject! {|k,v| v.empty?}
      my_headers.reject! {|k,v| (v == 'example.org')}
    end

    alias_method :old_do_request, :do_request
    def do_request(method, path, params)
      old_do_request(method, "#{$base_url}#{path}", params)
    end
  end

  class Curl
    def url
      "#{host}#{$base_url}#{path}"
    end
  end

end

Spec API test Example

I'll show you a sample test file spec/acceptance/business_spec.rb

require 'spec_helper'
require 'api_helper'
require 'rspec_api_documentation/dsl'

resource "Business v1.1" do

  before do
    prefill_db_for_api
  end

  $base_url = '/api/v1_1'
  let(:client) { RspecApiDocumentation::RackTestClient.new(self, :headers => { "HTTP_ACCEPT" => "application/json" }) }

  post "/login" do
    parameter :email, "Employee email", :required => true
    parameter :password, "Employee password", :required => true

    let(:email) { Employee.first.email }
    let(:password) { 'test22' }

    scope_parameters :employee, :all

    example "Authenticate with login and password" do
      do_request
      #puts 'login resp ' + response_body.inspect
      status.should == 200

      employee = Employee.first
      item = JSON.parse(response_body)
      item["authentication_token"].should_not be_empty
      item["authentication_token"].should eq(employee.authentication_token)
      item["email"] == employee.email
    end

  end

    get "/products" do

    # Explictly pre-set authentication_token
    let(:authentication_token) { Employee.first.authentication_token }
    let(:client) { RspecApiDocumentation::RackTestClient.new(self, :headers => { "HTTP_ACCEPT" => "application/json", 'HTTP_X_RABATME_AUTH_TOKEN' => authentication_token }) }

    example "Load products for current shop" do
      explanation "Products that are only available for this shop. Model: ShopProduct"

      do_request
      status.should == 200

      product = ShopProduct.first
      items = JSON.parse(response_body)
      puts items.inspect
      item = items.first
      item["id"].should_not be_empty
      item["id"].should eq(product._id.to_s)
      item["points"].should eq(product.points)
    end
  end

end

Now when you will run rake docs:generate it will run tests & generate API docs for you. If you wont only to run tests - you do rake spec:api

Fixing TextMate2 TODO bundle

2012-05-25T13:59:00+03:00

Having problems looking your TODO list in TextMate2? Run this script to have a fix.

cd ~/Library/Application\ Support/TextMate/Managed/Bundles/TODO.tmbundle/Support/lib && rm -rf settings.rb && curl -O http://ebundles.googlecode.com/svn-history/r216/trunk/Bundles/TODO.tmbundle/Support/lib/settings.rb

How I RestKit

2012-05-17T13:28:00+03:00

In this short post I'll show how I use RestKit library in my iOS projects.

First of all, I want to say that for managing all my libraries in iOS projects I use CocoaPods (kind of Gemfile bundler, but for iOS), which I strongly recommend using.

So, RestKit… open up your Podfile and add

dependency 'RestKit/Network'
dependency 'RestKit/UI'
dependency 'RestKit/ObjectMapping'
dependency 'RestKit/ObjectMapping/CoreData'
dependency 'RestKit/ObjectMapping/JSON'

I also use Injective dependency a lot, so I add dependency 'Injective' to my Podfile

Next, run pod install which will download latest RestKit (In my case - version 0.10.0) and integrate it to your project.

The code

Examples I'll provide will use RM class prefix, and API I'll refer is from my start-up rabat.me. First I'll write examples without data persistancy, than I'll provide examples how to use CoreData. Also, I add #import and #import "IJContext.h" to my App-Prefix.pch

You can find full source code of the project on GitHub https://github.com/xslim/RabatMe

Initialization

For initializing Injective library, edit your AppDelegate.m

- (BOOL)application:(UIApplication )application didFinishLaunchingWithOptions:(NSDictionary )launchOptions
{
    [IJContext setDefaultContext:[[IJContext alloc] init]];

What I also like doing, is initializing RestKit-s logging, so I can use it in my app and changing verbocity level from enviroment variables.

// This goes to application:didFinishLaunchingWithOptions:
RKLogInitialize();
RKLogConfigureFromEnvironment();

Create class RMApiConnector which we will use for all our RestKit bootstrapping and network connections.

NSString  const RMErrorDomain = @"org.API.ErrorDomain";

@interface RMApiConnector ()
- (void)fireErrorBlock:(RKRequestDidFailLoadWithErrorBlock)failBlock onErrorInResponse:(RKResponse )response;
@end

@implementation RMApiConnector

injective_register_singleton(RMApiConnector)

- (id)init {
    self = [super init];
    if (self) {
        [self setupConnector];
        [self setupMapping];
    }
    return self;
}

injective_register_singleton will add RMApiConnector to singleton collection on DI, so we will be able to access it like

RMApiConnector *connector = [RMApiConnector injectiveInstantiate]

Setting up Object Loader

- (void)setupConnector {
    NSString *baseUrl = @"http://example.org/api/"
    RKObjectManager *manager = [RKObjectManager sharedManager];

    if (!manager) {
        manager = [RKObjectManager objectManagerWithBaseURL:[RKURL URLWithString:baseUrl]];
        manager.client.serviceUnavailableAlertEnabled = YES;
        manager.requestQueue.showsNetworkActivityIndicatorWhenBusy = YES;
    } else {
        manager.client.baseURL = [RKURL URLWithString:baseUrl];
    }
}

Setting up Object Mapping

- (void)setupMapping {
    RKObjectMappingProvider *omp = [RKObjectManager sharedManager].mappingProvider;

    RKObjectMapping *productMapping = [RMProduct mapping];
    [omp addObjectMapping:productMapping];
    [omp setObjectMapping:productMapping forResourcePathPattern:@"/products"];

    RKObjectMapping *shopMapping = [RMShop mapping];
    [omp addObjectMapping:shopMapping];
    [omp setObjectMapping:shopMapping forResourcePathPattern:@"/shops/:id"];
    [omp setObjectMapping:shopMapping forResourcePathPattern:@"/shops/near/:lat/:lng/:radius"];
}

The mapping itself I store with Models like

@class RKObjectMapping;

@interface RMShop : NSObject

@property (nonatomic, strong) NSString itemId;
@property (nonatomic, strong) NSNumber points;
@property (nonatomic, strong) NSString name;
// Some other properties

@property (nonatomic, strong) NSArray products;
@property (nonatomic, strong) NSArray rewards;

+ (RKObjectMapping )mapping;

@end

@implementation RMShop

@synthesize itemId, points; // and others

+ (RKObjectMapping *)mapping {
    RKObjectMapping *objectMapping = [RKObjectMapping mappingForClass:[self class] usingBlock:^(RKObjectMapping *mapping) {
        [mapping mapAttributes:@"name", @"address", @"distance", @"location", @"points", nil];
        [mapping mapKeyPathsToAttributes:
         @"id", @"itemId",
         @"description", @"descriptionText",
         @"distance_unit", @"distanceUnit",
         nil];
    }];
    [objectMapping hasMany:@"products" withMapping:[RMProduct mapping]];
    [objectMapping hasMany:@"rewards" withMapping:[RMReward mapping]];
    return objectMapping;
}

@end

Creating API requests

Let's create few requests

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

< What we can see here, aption>

 class='obj-c'>- (void)loadNearbyShopsForLocation:(CLLocation )location onLoad:(RKObjectLoaderDidLoadObjectsBlock)loadBlock onError:(RKRequestDidFailLoadWithErrorBlock)failBlock class="p">{ NSString params = [NSDictionary dictionaryWithKeysAndObjects: @"lat", [NSNumber numberWithDouble:location.coordinate.latitude], @"lng", [NSNumber numberWithDouble:location.coordinate.longitude], @"radius", [NSNumber numberWithInt:20], // later change this to accuracy nil]; NSString url = [@"/shops/near/:lat/:lng/:radius" interpolateWithObject:params]; RKObjectManager manager = [RKObjectManager sharedManager]; [manager loadObjectsAtResourcePath:url usingBlock: class="p">(RKObjectLoader loader) { loader.onDidLoadObjects = loadBlock; loader.onDidFailWithError = failBlock; loader.onDidFailLoadWithError = failBlock; loader.onDidLoadResponse =  class="p">(RKResponse response) { [self fireErrorBlock:failBlock onErrorInResponse:response]; }; }]; class="p">} /tr>

is that I'm using block-based API of RestKit. loadBlock will get array of loaded objects. failBlock will get NSError object on error. Also I'm using fireErrorBlock:onErrorInResponse: helper to get error message from my API

- (void)fireErrorBlock:(RKRequestDidFailLoadWithErrorBlock)failBlock onErrorInResponse:(RKResponse *)response {

    if (![response isOK]) {
        id parsedResponse = [response parsedBody:NULL];
        NSString *errorText = nil;
        if ([parsedResponse isKindOfClass:[NSDictionary class]]) {
            errorText = [parsedResponse objectForKey:@"error"];
        }
        if (errorText) failBlock([self errorWithMessage:errorText code:[response statusCode]]);
    } else {
        //id parsedResponse = [response parsedBody:NULL];
        //RKLogTrace(@"response: [%@] %@\n%@", [parsedResponse class], parsedResponse, [response bodyAsString]);
    }
}

What if we want to make a POST call? Easy. Just tell the loader loader.method = RKRequestMethodPOST; and pass NSDictionary POST params to loader.params

RKObjectManager *manager = [RKObjectManager sharedManager];
[manager loadObjectsAtResourcePath:@"/stamps" usingBlock:^(RKObjectLoader *loader) {
    loader.method = RKRequestMethodPOST;
    loader.params = [NSDictionary dictionaryWithKeysAndObjects:
                         @"code", clientCode,
                         @"products", productsToSend,
                         nil];

What if we don't need to load any objects? Let's say we want to make client Authentication.

- (void)authenticateWithLogin:(NSString *)login password:(NSString *)password onLoad:(RKRequestDidLoadResponseBlock)loadBlock onFail:(RKRequestDidFailLoadWithErrorBlock)failBlock
{
    [[RKClient sharedClient] post:@"/login" usingBlock:^(RKRequest *request) {
        request.params = [NSDictionary dictionaryWithKeysAndObjects:
                          @"employee[email]", login,
                          @"employee[password]", password,
                          nil];
        request.onDidLoadResponse = ^(RKResponse *response) {

            id parsedResponse = [response parsedBody:NULL];
            NSString *token = [parsedResponse valueForKey:@"authentication_token"];
            //NSLog(@"response: [%@] %@", [parsedResponse class], parsedResponse);

            if (token.length > 0) {
                NSLog(@"response status: %d, token: %@", response.statusCode, token);
                [[RKClient sharedClient] setValue:token forHTTPHeaderField:@"X-Rabatme-Auth-Token"];

                if (loadBlock) loadBlock(response);
            }

            [self fireErrorBlock:failBlock onErrorInResponse:response];
        };
        request.onDidFailLoadWithError = failBlock;
    }];
}

Integrating CoreData

Great. What about CoreData? Change RMApiConnector's init method

if (self) {
    [self setupConnector];
    [self setupDB];
    [self setupMapping];
}

- (void)setupDB {
    RKObjectManager *manager = [RKObjectManager sharedManager];
    manager.objectStore = [RKManagedObjectStore objectStoreWithStoreFilename:@"HotSpots.sqlite"];
}

For mapping Managed Objects you do

- (void)setupMapping {
    RKObjectManager manager = [RKObjectManager sharedManager];
    RKObjectMappingProvider omp = [RKObjectManager sharedManager].mappingProvider;

    RKManagedObjectMapping *hotSpotMapping = [RKManagedObjectMapping mappingForClass:[HotSpot class] inManagedObjectStore:manager.objectStore];
    [hotSpotMapping mapAttributes:@"name", @"city", @"country", @"latitude", @"longitude", @"type", @"zipcode", nil];
    [hotSpotMapping mapKeyPathsToAttributes:
     @"address", @"street",
     @"description", @"details",
     @"id", @"hotSpotID",
     @"openinghours", @"openingHours",
     nil];
    hotSpotMapping.primaryKeyAttribute = @"hotSpotID";

    [omp addObjectMapping:hotSpotMapping];
    [omp setObjectMapping:hotSpotMapping forResourcePathPattern:@"app/location"];

Now if Object Mapper will see that the object he's mapping is a NSManagedObject he will check the DB if the'r already object with same primaryId. If yes - he'll update it with new values, if no - he'll create one for you.

Rails, Maps and ActiveAdmin

2012-05-17T07:16:00+03:00

In this short post I'll tell how I made Google-Maps-for-Rails be able to work with ActiveAdmin.

Let's asume you have a working configured Rails 3.2, MongoDB and ActiveAdmin.

First you need to install gem - gem 'gmaps4rails', then run rails g gmaps4rails:install

Let's add maps for our model object Shop.

class Shop
  include Mongoid::Document
  include Mongoid::Timestamps
  include Mongoid::Spacial::Document

  field :location, :type => Array, :spacial => {:lat => :lat, :lng => :lng, :return_array => true }
  index [[:location, Mongo::GEO2D]], :min => -180, :max => 180, :bits => 24, :background => true
end

We need to change it a bit and add this things:

  include Gmaps4rails::ActsAsGmappable
  acts_as_gmappable :process_geocoding => false

  def latitude
    location[1]
  end

  def longitude
    location[0]
  end

Now let's edit our app/admin/shops.rb

show :title => proc{ "#{shop.business.name if shop.business} / #{shop.name}" } do
attributes_table do
  row :name
  row(:location) {|o| gmaps("markers" => {data: o.to_gmaps4rails}, "map_options" =>  { auto_zoom: false, zoom: 15 }) }
  row :address
  row :description
  row :created_at
  row :updated_at
end
  end

If we will try to test it - we won't see a map, because ActiveAdmin is not auto loading our scripts and styles. So let's try fix that.

First, we need to add gmaps4rails css & javascript files for loading.

app/assets/stylesheets/active_admin.css.scss

@import "gmaps4rails";

app/assets/javascripts/active_admin.js

//= require gmaps4rails/gmaps4rails.base
//= require gmaps4rails/gmaps4rails.googlemaps

Now, we need to make custom view for out maps, that will also load google-map javascript. Create file app/views/gmaps4rails/_gmaps4rails.html.erb With contents:

<% case dom.map_provider %>
<% when "mapquest" %>
 id="<%= dom.map_id %>" style="width:750px; height:475px;">


<% when "bing" %>
   id="<%= dom.map_id %>" class="<%= dom.map_class %>">
<% else %>
 class="<%= dom.container_class %>">
   id="<%= dom.map_id %>" class="<%= dom.map_class %>">

<% end %>

Now if we test showing our model in ActiveAdmin, we will see a nice map. But let's add few more things. Edit your Shop model again and add this code:

  def gmaps4rails_title
    self.name
  end

  def gmaps4rails_infowindow
    "#{self.name}
#{self.description}

#{self.address}
#{self.location.join(', ')}"
  end

Now our markers will be clickable and will show nice info window.

Capistrano helpers

2012-05-16T16:37:00+03:00

While working on new rails project, I've made few lazy helpers for Capistrano cap command that I want to share. All of them are in my deploy.rb file.

Forwarding your id_rsa key for remote pulling the repo

ssh_options[:forward_agent] = true
before 'deploy:update', 'local:sshadd'

namespace :local do
  desc "Localy run ssh-add"
  task :sshadd, :roles => :app do
    system "ssh-add"
  end
end

Running remote bundle install

after 'deploy:update', 'bundle:install'

namespace :bundle do
  desc "Installs the application dependencies"
  task :install, :roles => :app do
    # For production, exclude gems from development & test groups
    run "cd #{current_path} && #{bundle_cmd} --without development test"
  end
end

Restarting Thin server

after 'deploy:update', 'thin:restart'

namespace :thin do
  desc "Restart Thin"
  task :restart do
    invoke_command "cd #{current_path} && bundle exec thin restart -C #{deploy_to}/current/thin.yml"
  end
end

Helpers to syncing local and remote MongoDB

set :mongodbname_prod, 'ximity_me_development'
set :mongodbname_dev, 'ximity_me_development'

namespace :syncdb do

  desc 'Synchronize MongoDB local -> production.'
  task :dev2prod, :hosts => "#{application}" do
    database = mongodbname_dev
    dev_database = database
    filename = "database.#{database}.#{Time.now.strftime '%Y-%m-%d_%H-%M-%S'}.tar.bz2"
    system "/usr/local/bin/mongodump -db #{database}"
    system "tar -cjf #{filename} dump/#{database}"
    upload filename, "#{shared_path}/#{filename}"
    system "rm -rf #{filename} | rm -rf dump"

    database = mongodbname_prod
    run "tar -xjvf #{shared_path}/#{filename}"
    run "/usr/local/bin/mongorestore #{fetch(:db_drop, '')} -db #{database} dump/#{dev_database}"
    run "rm -rf dump"
  end

  desc 'Synchronize MongoDB production -> local.'
  task :prod2dev, :hosts => "#{application}" do
    database = mongodbname_dev
    dev_database = database
    filename = "database.#{database}.#{Time.now.strftime '%Y-%m-%d_%H-%M-%S'}.tar.bz2"
    run "cd #{shared_path} && /usr/local/bin/mongodump -db #{database}"
    run "cd #{shared_path} && tar -cjf #{filename} dump/#{database}"
    download "#{shared_path}/#{filename}", "#{filename}"
    run "cd #{shared_path} && rm -rf #{filename} | rm -rf dump"

    database = mongodbname_prod
    system "tar -xjvf #{filename}"
    system "/usr/local/bin/mongorestore #{fetch(:db_drop, '')} -db #{database} dump/#{dev_database}"
    system "rm -rf #{filename}"
    system "rm -rf dump"
  end

  desc 'Generate indexes.'
  task :reindex, :host => "#{application}" do
    run "cd #{current_path} && bundle exec rake db:mongoid:create_indexes --trace"
  end

end

Do you have any useful helpers for cap ?

Fonts in iOS project

2012-05-16T16:34:00+03:00

Small piece of code that will show you available font names for using in you iOS project

NSArray fonts = [UIFont familyNames];
for (NSString fName in fonts) {
    NSLog(@"---> %@ :", fName);
    NSLog(@"\n%@", [UIFont fontNamesForFamilyName:fName]);
}

Start reading details

2012-04-07T20:31:00+03:00

I wonder when people will start reading some details before asking a lot of questions? Like the questions "Let's meet today in center (Kiev)?" or "Oh, so where do you live now?". Come on, people.

CLANG performSelector leak

2012-04-07T17:16:00+03:00

If you ever get a compiler warning performSelector may cause a leak because its selector is unknown you can suppress it in LLVM 3.0:

#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Warc-performSelector-leaks"
    [self.target performSelector:@selector(someAction:) withObject:self];
#pragma clang diagnostic pop

CocoaPods

2012-04-06T15:16:00+03:00

As CocoaPods is a ruby thing, you install it via gem:

$ sudo gem install cocoapods
$ pod setup

Installing missing libs:

RestKit 0.10.0

PodSpecName="RestKit"; PodSpecVersion="0.10.0"; PodSpecHome="${HOME}/.cocoapods/master"; PodRepoUrl="https://github.com/xslim/Specs" PodSpecFile="${PodSpecHome}/${PodSpecName}/${PodSpecVersion}/${PodSpecName}.podspec"; if [ ! -f  "${PodSpecFile}" ]; then echo "Will download to ${PodSpecFile}"; mkdir -p $(dirname ${PodSpecFile}) ; curl -L "${PodRepoUrl}/Specs/raw/master${PodSpecName}/${PodSpecVersion}/${PodSpecName}.podspec" -o "${PodSpecFile}" ; echo "Done."  ; fi

LibComponentLogging-NSLogger

PodSpecName="LibComponentLogging-NSLogger"; PodSpecVersion="1.0.4"; PodSpecHome="${HOME}/.cocoapods/master"; PodRepoUrl="https://github.com/xslim/Specs" PodSpecFile="${PodSpecHome}/${PodSpecName}/${PodSpecVersion}/${PodSpecName}.podspec"; if [ ! -f  "${PodSpecFile}" ]; then echo "Will download to ${PodSpecFile}"; mkdir -p $(dirname ${PodSpecFile}) ; curl -L "${PodRepoUrl}/Specs/raw/master${PodSpecName}/${PodSpecVersion}/${PodSpecName}.podspec" -o "${PodSpecFile}" ; echo "Done."  ; fi

After installing CocoaPods, update the project: cd project-dir

If it's first time you are installing: pod install MyProject.xcodeproj

Next time, you just update it with pod install

UPDATE: RestKit 0.10.0 and LibComponentLogging-NSLogger are now in CocoaPods specs, no need for manual install

Usefull Octopress links

2012-03-30T20:31:00+03:00

Рецензия на книгу ReWork

2012-01-03T00:00:00+02:00

ReWork - книга о ведении бизнеса от известной в IT кругах компании 37 Signals. Да да, именно тех 37 Signals которые создали Ruby on Rails и Basecamp. Она стала бестселлером в Америке сразу после выхода.

Авторы книги - Джейсон Фрайд и Дэвид Хайнемайер еще до этой книги выпустили бестселлер Getting Real. Так же у них есть замечательный блог - Signal vs Noise, на который подписано более 115 000 человек.

Удивительно что в компании 37 Signals, которая на рынке уже 10 лет, на данный момент работает всего 14 постоянных сотрудников, а программами, созданными этой компанией, пользуется более 3 миллионов человек по всему миру. Самые известные программные продукты компании: Basecamp - система управления проектами, Backpack - система управления знаниями, Highrise - система CRM.

При прочтении книги у меня было ощущение что со мной делятся секретами успеха в бизнесе мои друзья. Делятся легко и непринужденно, живо и искренне. Я постоянно сравниваю данные советы с личным опытом или опытом знакомых бизнесменов. Почти со всеми советами я соглашаюсь сразу, некоторые перечитываю по несколько раз что бы лучше понять. Некоторые так и не понимаю. Но все они заставляют включать серое вещество и смотреть на привычные вещи совсем под другим углом.

Нужны ли огромные инвестиции? Нужен ли вам офис? Нужно ли расти? Нужны ли совещания вообще? Нужен ли вам четкий бизнес план? Или гораздо выгоднее импровизировать?

Авторы на протяжении всей книги доказывают что привычный путь ведения бизнеса не всегда самый лучший. Они дают четкие стратегии и на своем опыте показывают как стоит вести бизнес и какие распространенные ошибки могут быть.

Структура

Книга написана простыми словами, без всяких заумностей, читается на одном дыхании. Каждая глава разбита на 5-6 блока:

заголовок - как краткое правило
обсуждение проблемы, с которой сталкивается большинство бизнесменов или начинающих
немного теории, с объяснениями что и почему
пример как было сделано хорошо или плохо в других компаниях
личный пример
краткий вывод

Оно вам надо?

Стоит ли читать книгу если вы занимаетесь бизнесом или только задумываетесь - однозначно. Я изначально прослушал ее как аудиокнигу на английском языке, потом прочитал на английском и теперь опять на русском.

Вам все еще что то мешает начать свой бизнес? Прочитайте ReWork и начинайте действовать!

Пару заметок из книги

Стоит сказать что делать заметки и цитаты с книги очень сложно, ибо авторы и так выбросили всю воду, оставили только четкий понятный текст мотивирующий к действию.

Помните известную байку о том, что, по мнению инженеров, шмель летать не может: аэродинамические качества не позволяют? Каждый май миллионы шмелей легко опровергают сложные научные расчеты.
Реальный мир – это не место, а всего лишь предлог. Оправдание, чтобы не действовать
Игнорируйте реальный мир
Определите очередную наиболее важную задачу и займитесь ею
Симпатичную рубашку, крем для бритья и зубную щетку можно купить, уже добравшись до места
Вам не нужно иметь степень МВА, модный костюм, портфель или склонность к риску выше обычного. Вам нужна идея, немного уверенности и небольшое усилие, чтобы начать
Если вы собираетесь что-то сделать, фокусируйтесь на том, что будет иметь значение
простой и очевидный способ создать первоклассный продукт или услугу – сделать что-то, что бы вы сами хотели использовать
Значение имеет только то, что вы делаете, а не то, о чем вы думаете, говорите или что планируете
Возьмите камеру, немного пленки и снимите абсолютно любой фильм
Ответственность за достижение ваших целей и осуществление мечтаний лежит полностью на вас
идеальное время никогда не наступит. Вы всегда либо слишком молоды, либо слишком стары, либо слишком заняты, либо слишком устали, либо еще что-нибудь
Нет смысла продавать лишние пару сэндвичей, если хлеб не так хорош
Мы скорее предпочтем потерять часть клиентов, если будем уверены, что это поможет оставшимся сильнее любить наши продукты
Отстаивать свои убеждения – не означает просто записать их. Это означает верить в них и жить ими
Внешние деньги – это план «Z»
Начинайте бизнес, а не стартап
Масштаб растет из-за: долгосрочных договоров; избыточного персонала; долгосрочных решений; встреч; невнятных процедур; запасов (физических или ментальных); лишнего оборудования, программного обеспечения, технологических процессов; долгосрочных планов развития; офисных политик.
Путь к великому начинается с отбрасывания всего, что хорошо лишь в умеренных дозах
Создайте половину продукта, а не недоделанный продукт
Если я это уберу, будет ли все еще существовать то, что я продаю?
если вы хотите открыть палатку по продаже хот-догов, вы можете беспокоиться о приправах, самой палатке, названии, оформлении. Но первое, о чем вам стоит позаботиться, – это о самих хот-догах. Хот-доги – это эпицентр. Все остальное вторично
Игнорируйте детали на ранних стадиях
Длительные проекты пагубно влияют на командный дух. Чем больше времени занимает разработка, тем менее вероятен запуск
Отделяйте все лишнее, пока у вас не останется только самое важное. Затем сделайте это снова
Когда что-то не работает, обычный подход заключается в том, чтобы бросать на решение проблемы больше ресурсов. Больше людей, времени и денег. Все это в итоге приводит к разрастанию проблемы. Правильный подход – действовать в обратном направлении, то есть сокращать
Вы же хотите создать что-то замечательное. Мы говорим о том, что лучший путь достичь этого – через последовательные итерации. Перестаньте думать о том, что будет работать, – выясните это на практике
Отложите все, что вам не нужно для запуска. Создавайте все жизненно необходимое сейчас, а все роскошества оставьте на потом
Запускайтесь сейчас
Перерывы – враги продуктивности
ваш день находится в осаде постоянных перерывов. Ваша задача – отвоевать его обратно
Если вы находите свой способ решения задачи достаточно хорошим, останавливайтесь на нем. Это гораздо лучше, чем тратить ресурсы или, что хуже, ничего не делать из-за того, что вы не придумали более красивое решение
Вместо планирования одного проекта, рассчитанного на 12 недель, разбейте его на 12 однонедельных проектов
существует отличный способ оградить себя от подражателей – самим стать частью своего продукта или услуги
Проблема копирования заключается в следующем: пропускается стадия понимания, осознания глубинных механизмов, с помощью которых растет бизнес. При слепом подражании воспроизводится последний из слоев без того, чтобы понять, какие слои находятся под ним.
вы можете себе позволить обучать других. Большие компании никогда не смогут этого делать, потому что одержимы манией секретности
Обучайте – и вы «привяжете» к себе людей так, как не снилось поклонникам традиционных маркетинговых приемов
Большинство вещей, о которых вы переживаете, так никогда и не произойдут.
Если вы хотите работать с человеком долгое время, не нужно рассчитывать на то, что работа будет полностью занимать его жизнь.
Если вы хотите, чтобы что-то было сделано, поручите выполнение самому занятому человеку
Если вдохновение пришло к вам в пятницу, пожертвуйте выходными и полностью погрузитесь в проект. В состоянии подъема вы можете за 24 часа выполнить объем работы, на который в обычной ситуации пришлось бы потратить пару недель

How to get iBooks Hightlihts

2012-01-03T00:00:00+02:00

So for example you want to get the highligtings from iBooks, what you need is few steps:

Get some app to mount iBooks Folder (like PhoneDisk)
Download AEAnnotation sqlite database from Documents/storeFiles to some local folder
cd in that folder with your terminal application and execute

sqlite3 AEAnnotation_*_local.sqlite 'select `ZANNOTATIONSELECTEDTEXT`, `ZANNOTATIONASSETID` from `ZAEANNOTATION` where (`ZANNOTATIONSELECTEDTEXT` not NULL) order by `ZANNOTATIONASSETID`,`ZPLLOCATIONRANGESTART`'

You will see list of notes and Boor IDs to which they are associated

Find out the Book ID you need (Example: 06CA7575F9AACCAC5027272BA8926BB1)
execute the command

sqlite3 AEAnnotation_*_local.sqlite 'select `ZANNOTATIONSELECTEDTEXT` from `ZAEANNOTATION` where (`ZANNOTATIONASSETID` = "06CA7575F9AACCAC5027272BA8926BB1") AND (`ZANNOTATIONSELECTEDTEXT` not NULL) order by `ZPLLOCATIONRANGESTART`'

And you will get your highlights !

Few dev notes

Create Cocoa app for Exporting iBooks Highlights
Use https://github.com/xslim/mobileDeviceManager to get data from device
Open BKLibrary_database/iBooks_*.sqlite (*BKL*) and storeFiles/AEAnnotation_*.sqlite (*AEA*)
Interconnection by BKL->ZBKBOOKINFO->ZPLUGINASSETID and AEA->ZAEANNOTATION->ZANNOTATIONASSETID

PHP to AD LDAP connection

2011-12-27T16:11:00+02:00

if (!$this->ds) {
  $this->ds = ldap_connect($host);
  if ($this->ds) {
      ldap_set_option($this->ds, LDAP_OPT_PROTOCOL_VERSION, 3);
      ldap_set_option($this->ds, LDAP_OPT_REFERRALS, 0);
      if (!@ldap_bind($this->ds, $login, $passwd)) {
          return false;
      }
      return $this;
  } else {
      return false;
  }
}

public function searchUser($s) {
  $attributes = array('sAMAccountName', 'cn');
  $filter = '(&(objectCategory=user)(displayName=*'.$s.'*))';
  $sr = ldap_search($this->ds, $this->dn, $filter, $attributes);
  $num_enntries = ldap_count_entries($this->ds, $sr);
  $users = array();
  if ($num_enntries > 0) {
      $info = ldap_get_entries($this->ds, $sr);
      foreach($info as $u) {
          if (!empty($u['cn'][0]))
          $users[] = array(
                  'login' => $u['samaccountname'][0],
                  'name'  => $u['cn'][0],
              );
      }
      return $users;
  } else {
      return false;
  }
}


include("pChart/pData.class");
include("pChart/pChart.class");

// Dataset definition 
$DataSet = new pData;
$DataSet->AddPoint($data,"Serie1");
$DataSet->AddPoint($legend,"Serie2"); 
$DataSet->AddAllSeries();
$DataSet->SetAbsciseLabelSerie("Serie2");

// Initialise the graph
$graph = new pChart(720,350);
$graph->loadColorPalette('hard_bright.txt');

// Draw the pie chart
$graph->setFontProperties("Fonts/tahoma.ttf",12);
$graph->drawTitle(50,20,'Top 20 Incominf for '.$date ,0,0,0);  

$graph->setFontProperties("Fonts/tahoma.ttf",9);

$graph->drawPieGraph($DataSet->GetData(),$DataSet->GetDataDescription(),210,160,180,PIE_PERCENTAGE,TRUE,50,20,5);
$graph->drawPieLegend(420,5,$DataSet->GetData(),$DataSet->GetDataDescription(),250,250,250);

$graph->Stroke();