Programming on Leif Thoughts

Play Wordle for a Different Day

Wed, 12 Jan 2022 17:55:24 -0800

UPDATE 13 January 2022: A previous version of the article did not include Google Chrome’s Override instructions.

One element of writing my last wordle article was creating screenshots to illustrate the solving algorithm. But I didn’t want to use the word of that day (which was “gorge”), I wanted to use a specific word (“tiger” fom Jan 5).

This is where my digging into the workings behind wordle started. I was curious how the word of the day is chosen. I noticed that the new word is always available at midnight of “your” timezone so I suspected that all of the logic is implemented in the frontend (no secrets that are fetched from a server).

Building a Wordle Solver With Regular Expressions and Elixir

Sun, 09 Jan 2022 16:46:21 -0800

This is one of these articles that is predestined to start with “unless you’ve been living under a rock, you have heard of …”. Unfortunately I don’t know how popular the thing I’m gonna write about is outside my bubble but here we go.

Among my group of friends and colleagues the online game wordle has been hugely popular. It’s a mix of Scrabble and Mastermind where you’ll have to guess a five letter word. Hints about individual characters are conveyed to you via yellow, green and gray indicators. Green means the target word has this character at this location, yellow means the target word has this character but in a different location and gray means the character is not in the target word at all.

Grep Bundler Version From Gemfile.lock

Fri, 10 Dec 2021 17:31:27 -0800

If you’re working on any kind of Ruby project you probably won’t be able to avoid bundler.

This is just a quick tip to install the exact same version of bundler than the Gemfile was originally installed with.

Here you go:

gem install bundler -v $(grep 'BUNDLED WITH' -A1 Gemfile.lock | tail -n 1 )

I used it a couple of times and it can be useful to keep bundler versions in sync when you’re using docker.

Overengineering My Blog for Fun and Definitely not Profit

Tue, 07 Dec 2021 17:08:09 -0800

This publication has been around a while. According to my first article it goes back all the way to January 2010 and was hosted on Wordpress. In the meantime I switched from there to Jekyll until I switched to ghost in late 2013.

This is where the blog spent most of its time, and even thought I had some stories reach the front page of Hacker News it was mostly quiet around here and infrequently updated (I did not write a single post in all of 2016).

Implement a GIT Branch Checkout History

Mon, 22 Feb 2021 16:11:46 -0800

Take this alias

# .gitconfig

[alias]
    hist = "!f() { for i in $(seq 9 $END); do echo "@{-$i}: `git rev-parse --abbrev-ref @{-$i}`"; done }; f "

This will show the last 9 branches you have checked out with the abbreviation to switch to it.

❯ git hist
@{-1}: main
@{-2}: add-healthchecks-io
@{-3}: main
@{-4}: fix-visible-flag
@{-5}: main
@{-6}: implement-join-code
@{-7}: main
@{-8}: fix-visible-flag
@{-9}: main

You can of course use the branch name to check it out, or use the abbreviation in front of it git checkout @{-6} will check out implement-join-code.

My First Event Sourced Application

Tue, 09 Jan 2018 17:53:43 +0700

TL;DR: I built an event sourced application that shows the latest version of common programming languages. Find it here.

For I while I thought about building an event sourced application. The concept is so different from a classical CRUD approach that I was very intrigued.

I tried once with a podcasting transcription app (the details of that can be read here) but I quickly became overwhelmed by all the design decisions I had to make.

Use docker to run the new Elixir Code Formatter

Thu, 12 Oct 2017 17:47:33 +0200

TL;DR: Run this to format your code:

docker run --rm -it -v $(pwd):/app -w /app leifg/elixir:edge mix format

The community has been very excited in the last couple of days about the Elixir 1.6 Code Formatter Announcement.

Of course you can use one of the many Elixir Version Management Tools to try out the formatter. But if you don’t want to switch around between versions or if you (like me) stick to system Elixir, there is an easier way: Docker.

Configure your Elm app with ENV variables

Sun, 01 Oct 2017 23:53:54 +0200

Create Elm App makes it very easy to get started with Elm because it provides a zero configuration application that just works.

However sometimes it is very useful to have some kind of configuration in your app. The example that I recently ran into is making your backend api url configurable. Or do you really want to develop against your production API?

Lucky for you Create Elm App already comes with all necessary tools. It’s kinda tricky to find and that’s why I’m writing this post.

Deploying Elixir with Docker Part 4

Fri, 01 Sep 2017 17:36:28 +0200

This last post is about bringing it all together.

Right now we are at a point where we have random script snippets that executed in the right order will launch a service in the cloud.

But we need to get this in a more structured way on a build server. As a CI provider I use CircleCI and I’ve been waiting to try out their 2.0 version including workflows.

As CircleCI heavily relies on Docker, we can just go ahead and use our base image for building.

Deploying Elixir with Docker Part 3

Fri, 25 Aug 2017 09:51:33 +0200

This is the third part of the Docker Elixir deployment adventure. Read part 1 and part 2 first.

In this part we are going to deploy our Docker container to hyper.sh (Affiliate Link). It’s a relatively new platform that makes it very easy to deploy containers.

You will need the hyper cli to get started. Refer to the official documentation for details.

We will start by starting an instance of our container. By the time of writing this blog post, the latest version of time_tracking is 1.8.3:

Deploying Elixir with Docker Part 2

Thu, 17 Aug 2017 18:20:26 +0200

This is a continuation of the first part of the post where we created a release.

Now that we have a release, let’s pack it into a docker container.

We will use a multi stage container to create the container.

First let’s focus on the build part. A base container will come in handy as we need to run several commands. We need Erlang and Elixir installed (obviously) but I also need a tool to calculate the version number. I use Semantic Release for that. I decided to switch from the JavaScript version to the Go version.

Deploying Elixir with Docker Part 1

Fri, 11 Aug 2017 19:37:45 +0200

This is a multi part article about deployment of an Elixir application with Docker.

This part will be about building a distillery release for a Phoenix Application.

Part 2 will discuss packaging this release into a Docker container

Part 3 will show you how to deploy this container to a container runtime.

Part 4 will wrap it all up and show you how to integrate all if this into a Continuous Delivery Workflow

Try out your code on the latest OTP and Elixir

Sun, 18 Jun 2017 20:06:44 +0200

TL;DR: If you want an easy way to run the latest version of elixir (an the latest version of OTP) and you have docker installed, do this:

docker pull leifg/elixir:edge
docker run --rm -it leifg/elixir:edge iex

Elixir Docker Builds #

A while ago I started building my custom Elixir Docker image using Github and Docker hub automated builds.

I tagged the Elixir versions I wanted to build to create a 1.x and 1.x.y tag on Docker Hub. In addition the master branch would be tagged with latest. I always used the latest stable OTP to reduce complexity.

Never compare dates in Elixir using "<" or ">"

Mon, 29 May 2017 12:23:23 +0200

For the examples in this article I use the data structure Date. The same applies for DateTime or NaiveDateTime

Consider this:

early_june = ~D[2017-06-01]
late_june = ~D[2017-06-30]

How do you find out what’s earlier?

The naive approach is to use the comparison operator <:

early_june < late_june
# => true

That works. But this is tricky. It only works in some cases.

Now consider this:

late_june = ~D[2017-06-30]
early_july = ~D[2017-07-01]

What do you think the output is now?

late_june < early_july
# => false

Turns out: comparison with < (or >) just compares the struct fields (as the current version of the documentation states. My best guess is for same reason the day field has higher priority than the month field so this comparison fail sometimes.

Export your Salesforce Objects to SQLite

Thu, 10 Dec 2015 18:11:22 -0800

Sometimes it can be quite useful to export data you have in Salesforce Org (transferring your data over to a new system is probably the biggest use case).

Saleforce’s suggestion is to use the dataloader but that has a lot of drawbacks:

Flat File Format only
Not (sufficiently) scriptable
Manual process to select the fields you want (instead of just everything)
Only one object at a time

As most of the information you need for a proper export a available via Salesforce’s field description, it was pretty simple to come up with a solution that would let me export a range of objects to a database format (SQLite currently but very easy to extend for other RDBMS).

Redis Tips & Tricks #2 - Connections

Fri, 04 Sep 2015 11:30:44 -0700

As it turns out, connecting to redis may introduce a bit of problem if you’re not aware of the connection handling.

I haven’t tried with every client out there but especially the redis ruby client has this problem.

So imagine this scenario:

(1..10000).each{|num| Redis.connect.get("key_#{num}")}

Every iteration will open a new file handle and will keep it open (I have not been waiting forever but I have seen open files for more than 24 hours).

First steps in Elixir

Fri, 28 Aug 2015 17:53:15 -0700

From time to time I look into new programming languages. But usually I rarely do more than browse the general introduction guide.

However this time I dove a little bit deeper into Elixir and even got myself a pet project.

I tried to find a use case for a little library and I ended up writing a parser for Excel documents. It’s still under heavy development but I gladly accept Pull-Requests.

Executrix is now known as Bulkforce

Tue, 28 Jul 2015 13:36:59 +0200

Today I released a new version of the ruby gem to interact with the Salesforce Bulk API.

I decided to go with a different name and introduce some breaking changes.

And here it is: Bulkforce.

The [old gem](old gem) will still be available as a reference and to provide a backwards compatible option for all users out there. New features however will only be introduced in Bulkforce.

Not only did I rename the gem, I also introduced a new feature. The authentication has been overhauled.

Redis Tips & Tricks #1 - Memory

Fri, 19 Dec 2014 18:15:36 +0100

It’s neither a tip, nor a trick. It’s more a statement:

ALL data is kept in memory

“No shit” some of you might say, but I actually have to confess: I did not know that when I started using redis. I always assumed “there is a file on the disk, that is where my data lives”.

The truth is: all of redis’ data is kept in memory and from time to time it is also written to disk

Why ActiveRecord Enums Have Been Implemented in the Worst Possible Way

Fri, 14 Nov 2014 11:38:04 +0100

UPDATE: I tried out enums in ActiveRecord 5.0 and the biggest pet peeve has been eliminated. Point 3 is no longer valid. Nevertheless decide for yourself if you want to use it.

With Rails 4.1 enums have been introduced to ActiveRecord. I took a good look at it and from the start I didn’t like it. Here are 3 reasons why:

1. It’s all in the application layer #

Ok this might be a bit unfair as Rails is the application layer. Nevertheless the feature could have taken advantage of Postgres’ or MySQL’s native type. An example of how that could look like is in my experimental gem activerecord-real_enums.

Redis Tips Tricks - #0 Introduction

Mon, 10 Feb 2014 15:21:02 +0100

Yai, a series \o/

Here I try to explain in short steps some tricks for beginners in handling redis.

Some of those tricks may be specific to redis, some other might be relevant for all key-value-stores.

So stay tuned.

Morf Hashes Into Hashes

Sat, 21 Dec 2013 16:14:33 +0100

In the last few months I did a lot of work concerning imports.

Using the Sequel Toolokit I was able to treat every query row as a hash and the whole thing has been working out pretty good so far.

However I always needed to do conversions on the hashes (map keys to a different name, concatenate values etc….). Therefore I implemented several mapping classes that helped me with that but was never quite satisfied with the implementation.

Get Yourself a Simple Project for Explaining Stuff

Mon, 16 Dec 2013 00:56:35 +0100

A while ago I spent some time at home, working on a Rails project. I was very passionate about it and at one point I shouted out: “Finally! The tests are running”.

That was the point when my dad got curious about what I was doing and asked me something like “What tests are you talking about?”. I tried to showed him my specs. Unfortunately in order to explain the whole specs I also needed to explain ORM, mocking, factories and a couple of other stuff, because it was a Rails project and gotten pretty far by that point of time. That got me thinking: “If it had been a simple class without any dependencies it would have been a lot easier to explain the tests.”

New Project calcsub.net

Sun, 01 May 2011 17:54:47 +0200

Over the easter holidays I started working on a new project. Today it has finally reached a state where I’m confident to say: I can go public.

It’s just a simple subnet calculator that takes IPv4 and IPv6 addresses and then calculates the start address and the end address.

You can have a look at it here (NON FUNCTIONAL) or see the source on github.

picture created by Leif Gensert. CC BY 2.0

Simplequeue - My First Project

Sun, 24 Oct 2010 16:18:36 +0200

Today I released my first project on github. It’s called simplequeue and is a little tool to queue tasks. In this case everything is a task which can be expressed in a single line of text.

It’s written in ruby and as I’m new to ruby I would really appreciate it if someone took the time and make a code review. Just contact me or fork the project if you have suggestions.

I also added a portfolio page to this site. I will do my best to add more projects to my portfolio.

So You Want to Be a Programmer?

Fri, 23 Jul 2010 23:35:12 +0200

If you’re anything like me, you have always wanted to learn a programming language, do an awesome application and get filthy rich. To achieve this goal you find plenty of resources available online. But all articles on how to become a programmer really boil down to just 2 letters:

#do

That sums it up pretty well. You won’t become a good programmer unless you’ll write code. It’s as simple as that. But as putting out just 2 letters for a guide might seem arrogant, I summed up 3 points which I find very helpful in learning a programming language.

Superfantastic Password Management With Supergenpass

Wed, 14 Jul 2010 10:40:27 +0200

Nowadays, there are a million websites out there which force you to create an account. During this step you are required to provide a username and a password. If this process gets repeated more often, the need of a proper password management increases. The currently most used password management is the simple “use one for all” method. I hope we alle agree that is the worst of all methods. The best available method however (generating long arbitrary passwords and memorizing them) is not that practicable. So you need to find a system between those two extremes. A few months back, I chose a password managing system with the help of SuperGenPass and stuck since then.

Jekyll: Blogging Without Pain

Fri, 18 Jun 2010 02:21:00 +0200

A long time hast passed since my last blog post. But I did a complete redesign of my blog. Not only is the frontend completely different. The underlying backend has changed as well. I completely freed myself from Wordpress and changed to an engine that fits my workflow much better: jekyll.

To call jekyll a blogging-engine is (IMHO) not quite right. It’s just a generator for static contents. The general idea behind it is that you write all your articles in a markup language (e.g. HTML, Markdown or Textile), define the layout via the Liquid Templating engine, run the jekyll generator and end up with a complete directory structure containing HTML/CSS and JavaScript files which can be used as the actual site.