If the project team consists of at least 2 members, do a daily standup. It should not take much longer than 15 minutes.

Format

Tell everyone else

• what you did yesterday
• what you intend to do today
• where you might need help or other input
• if there are new developments everyone needs to know about

A "still working on X, will probably be done today" is totally fine. No need to tell a long story.

If you are out of work, find a new story with the others.

If there are new stories in the backlog, look at them and

• make sure ev...

We organize our daily work with issues in our Linear workspace.

Issue format

A good issue needs to be precise. It should be very clear what is part of an issue, and what is not. If there are different expectations between the person who writes and who implements an issue, there will be rejects.

To this end, we use a consistent format for issues that looks like this:

Issue: Autocomplete

As a journalist, I want to have an autocomplete in the search bar, to have a more efficient way to find articles.

Acceptance criteri...

When starting a project we always make a good estimate of all known requirements, and plan budgets and available developers accordingly.

Requirements change. Budgets usually don't.

To make sure a project stays on track, we update our estimates once a month and compare them to the remaining budget. If this doesn't match any more, we have to act.

To update an estimate, do the following:

• Start with the most recent estimate for the project.
• Which stories have been completed? Set their estimate to zero.
• Have any requirements cha...

Ruby has Enumerable.find(&block), which returns the first item in the collection for which the block evaluates to true.

first_post_with_image = posts.find do |post|
  post.image
end

However, sometimes it's not the item you're interested in, but some value depening on it – e.g. the value the block evaluated to. You could first map the collection and then take the first truthy value, but this way you need to process the whole collection twice:

first_image_url = posts.map(&:image).find(&:present?).url

If the mapping ...

On Ruby 1.9+, standard ruby character classes like \w, \d will only match 7-Bit ASCII characters:

"foo" =~ /\w+/   # matches "foo"
"füü" =~ /\w+/   # matches "f", ü is not 7-Bit ASCII

There is a collection of character classes that will match unicode characters. From the documentation:

• /[[:alnum:]]/ Alphabetic and numeric character
• /[[:alpha:]]/ Alphabetic character
• /[[:blank:]]/ Space or tab
• /[[:cntrl:]]/ Control character
• /[[:digit:]]/ Digit
• /[[:graph:]]/ Non-blank character (excludes spaces...

tig is a command line explorer for Git that is just awesome. Install via apt-get or brew.

Handy commands

• t ("tree"): Directory-structure based access. You'll see the current directory annotated with the latest change date and its author. Navigate with arrow keys or vim.
• b ("blame"): Opens the file under the cursor and annotates each line with change date and author.
• d ("diff"): Like ENTER on a commit, but arrow keys will scroll the diff!
• /: Search current view (e.g. commit list, diff). Jump to next hit with n....

Several Rails migration methods accept index: true as an option to create an index. In some cases (like #add_column), this option is silently discarded. Know what you are doing, or use #add_index instead.

Example

Consider the following migration.

class CreateExamples < ActiveRecord::Migration
  def change
    create_table :examples do |t|
      t.references :category, index: true
      t.boolean :positive, index: true
      t.integer :number_of_participants, index: true
    end

    add_reference :examples, :user, index: tr...

Bad

# Given the following models

class Image < ActiveRecord::Base
  has_many :album_images
  has_many :albums, through: :album_images
end

class Album < ActiveRecord::Base
  has_many :album_images
  has_many :images, through: :album_images
end

# Join model
class AlbumImage < ActiveRecord::Base
  belongs_to :album
  belongs_to :image
end

Destroying a record in this setup will only remove the record itself, and leave orphaned join records behind.

image = Image.last
image.destroy # removes only the `image` record,
   ...

When ending a Selenium test Capybara resets the browser state by closing the tab, clearing cookies, localStorage, etc.

It may be a good idea to wait for all in-flight AJAX requests to finish before ending a scenario:

• You may have client-side JavaScript that freaks out when the tab closure kills their pending requests. If that JavaScript opens an error alert or spams errors to the console, your test may fail after the last step.
• With unlucky timing the server may receive an AJAX request as the browser tab closes, causing a connection ...

You know that ActiveRecord caches associations so they are not loaded twice for the same object. You also know that you can reload an association to make Rails load its data from the database again.

user.posts.reload
# discards cache and reloads and returns user.posts right away
# => [...]

If you want to discard the cache but not query the database (only the next time the association is accessed), you can use reset:

user.posts.reset
# discards cache, but does not load anything yet
user.posts
# SQL query happens to ...

Web applications can be used by multiple users at the same time. A typical application server like Passenger has multiple worker processes for a single app. In a distributed deployment setup like we use at makandra you will even have multiple application servers, each with their own worker pool.

This means that your code needs to deal with concurrent data access. The two main tools we use to cope with concurrency are database transactions and distributed locks. These two are not interchangeable. You ca...

This guide shows how to create an AngularJS application that consumes more and more memory until, eventually, the browser process crashes on your users.

Although this guide has been written for Angular 1 originally, most of the advice is relevant for all client-side JavaScript code.

How to observe memory consumption

To inspect the amount of memory consumed by your Javascripts in Chrome:

• Open an incognito window
• Open the page you want to inspect
• Press Shift + ESC to see a list of Chrome processes...

The CSS property z-index is not as global as you might think. Actually, it is scoped to a so-called "stacking context". z-indexes only have meaning within their stacking context, while stacking contexts are treated as a single unit in their parent stacking context. This means indices like 99999 should never actually be needed.

Creating a new stacking context

In order to create a stacking context with the least possible side effects, use the isolation property on an...

One of the many useful features of TextMate is autocompletion of words. If I were in TextMate right now, I could write "au[tab]", and it would complete it to "autocompletion". RubyMine can do this, too. When you write a word (e.g. a variable name), just hit ALT + / repeatedly and it will offer all completions for the letters you typed. This action is called Cyclic Expand Word in RubyMine / IntelliJ IDEA.

This feature keeps you from mistyping variable names, saves you keystrokes and speeds up development. ~10 keystrokes to the price ...

Rails migrations allow you to use a change method whose calls are automatically inverted for the down path. However, if you need to some path-specific logic (like SQL UPDATE statements) you can not define up and down methods at the same time.

If you were to define define all 3 of them, Rails would only run change and ignore up and down. However, Rails 4+ features a helper method called reversible:

class MyMigration < ActiveRecord::Migration

  def cha...

For our production servers we use Passenger as a Ruby application server. While it is possible to use Passenger for development as an Apache module, the installation process is not for the faint of heart.

Luckily Passenger also comes as a standalone binary which requires zero configuration.

You can Passenger Standalone as a replacement for Webrick or Thin if you'd like to:

• Use SSL certificates locally
• Get performance behavior that is closer to ...

In a JavaScript console, type this:

> 9112347935156469760
9112347935156470000

Ooops. And that's not a float!

This occurs because JavaScript uses double precision floats to store numbers.

So according to IEEE floating point definition only numbers between -(2^53 - 1) (-9007199254740991) and 2^53 - 1 (9007199254740991) can safely be represented in JavaScript.

Note that ECMAScript 6 will probably also offer [Number.MAX_SAFE_INTEGER](https://developer.mozilla.org/en-US/docs/W...

Cookies without an expiration timestamp are called "session cookies". [1] They should only be kept until the end of the browsing session.

However, when Chrome or Firefox are configured to reopen tabs from last time upon start, they will keep session cookies when closing the browser. This even applies to tabs that were closed before shutting down the browser.

This is by design in Chrome and [Firefox](https://bugzilla.mozilla.org/buglist.cgi?bug_id=337551,345830,358042,362212,36...

Capybara will fail to find <a> tags that are missing an href attribute. This will probably happen to you every now and then on JavaScript-heavy applications.

An example would be an AngularJS application where the following HTML actually works. [1]

<a ng-click="hello()">Hello</a>

Capybara will fail to find that link, even though looking it up via the DOM shows it:

>> find_link("Hello")
Capybara::ElementNotFound: Unable to find link "Hello"

>> find("a").text
=> "Hello"

To make find_link and click_link work, ...

Due to network or hardware failures, it can happen that one of your cronjobs will not run at the time you specify in the schedule. Your code should be built in a way that it can be re-run at a later time (when the failure is resolved).

For example, if you are synchronizing data with another service once every day, your cronjob should not only synchronize changes from the last 24 hours. If you do this and a network failure will delay the execution of your job by 5 hours, you will only synchronize changes from hour 6-29, but forget change...

Safari on iOS accepts an apple-touch-icon favicon that is used for stuff like desktop bookmarks. Always define a solid background color for them.

If you use PNGs with a transparent background, Safari will use just set a black background on your pretty icon. This is almost never what you want.
You can fix that by applying a white background via ImageMagick like this:

convert a...

When you have a string containing umlauts which don't behave as expected (are not matched with a regexp, can't be found with an SQL query, do not print correctly on LaTeX documents, etc), you may be encountering umlauts which are not actually umlaut characters.

They look, depending on the font, like their "real" umlaut counterpart:

• ä ↔ ä
• ö ↔ ö
• ü ↔ ü

However, they are not the same:

'ä' == 'ä' # false
'ä'.size # 1
'ä'.size # 2

Looking at how those strings are constructed reveals what is going on:

'ä'.unpack('U*...

Your development machine is usually on a very good network connection.
To test how your application behaves on a slow network (e.g. mobile), you can simulate limited bandwidth.

Chrome

• Open the dev tools (Ctrl+Shift+I or F12) and switch to the "Network" tab
• In the row below the dev tool tabs, there's a throttling dropdown which reads "Online" by default.
• Inside the dropdown, you will find a few presets and an option to add your own download/upload/latency settings.

Firefox

• Open the dev tools (Ctrl+Shift+I or F12) and switc...