@ctshryock

MTL :: posts

2016-02-04T00:00:00-06:00

There are countless articles and guides on helping you get things done. Probably even more books. Talks, sermons, infomercials. Here’s another one.

Over the years I’ve tried several techniques to get things done. A few years back TODOs and GTDs were a huge deal that everyone talked about. I didn’t get into them because of that, but because I needed to get things done and damn, there are tons of distractions, and I’m prone to paying attention to them.

For a while I used various apps to help me here. I even used my own little Basecamp setup, but eventually I think I picked up and used Things.app for the longest. All of these are great tools but none of them really stuck. They are great tools but what I needed was a framework or process.

An internet friend of mine happened to post on Facebook a recommendation of The 5 Choices: The Path to Extraordinary Productivity, right around the time I hadn’t been feeling my most productive lately, so I gave it a spin. It resonated with me. It was less of a “do these 5 things so-and-so does and you’ll be productive!”, and more of a “understand what leads to productivity and train yourself to focus on your life”.

The book is broken into sections:

Decision management / urgent
Extraordinary vs. ordinary
Scheduling
Attention management

The book goes into wonderful detail and good examples on the importance of each of these things. The chapter on Scheduling has the most relevance to this post, as it outlines a Master Task List, which I’ve adopted for my latest get things done approach. It’s really just a simple list:

When something comes up that you might need to do, it goes either on the floor or on the list,
– Kogon, Kory; Merrill, Adam; Rinne, Leena (2014-12-30). The 5 Choices: The Path to Extraordinary Productivity (Kindle Location 1145). Simon & Schuster. Kindle Edition.

The message is simple, but I found it profound. If it needs to be done, write it down. Otherwise just drop it. Your first reaction may be “well, everything needs to be done, so I’m no better off now”. But this partners well with something established in the first chapter re: decision management, which is also profoundly simple:

Is it important?
– Kogon, Kory; Merrill, Adam; Rinne, Leena (2014-12-30). The 5 Choices: The Path to Extraordinary Productivity (Kindle Location 666). Simon & Schuster. Kindle Edition.

“Well, everything is important, right?” No. Changing my way of thinking has helped me here. How important is this? Or a better question, what are the consequences if this does not get done? Take a moment to be calm and think clearly, and ask yourself that question. Does this really matter? If the consequences are pretty trivial, then it doesn’t belong on the list. “Pay my mortgage” is pretty consequential. Making my kids beds is not. I should still do the latter, but it’s not worth cluttering my list of things I need to get done.

My Trello System

So, I’ve developed a system using Trello (a wonderful, free service you should checkout anyway), which helps keep me in-line and productive. It’s not perfect, it’s still a work in progress, but so far it’s been helpful. I have the following lanes in a Trello board, and a Trello card typically progresses from left to right:

MTL: This is where I add things that I need to get done. These represent things that have consequences if they are not done.
This week: These are things things I plan to accomplish this week. This is determined Monday mornings but can shift throughout the week.
Doing: Thing(s) I’m doing right now. This lane should only have 1-3 cards in it at most, otherwise you aren’t focusing on any one of them.
Done: Things I’ve accomplished this week. It’s nice to see this lane fill up. Every Monday I archive all the cards in this list.

This Week is a best-guess effort to plan out what you want to do this week. It requires a calendar and booking time of your week to make sure you can accomplish the big rocks of your to-dos. You will always have things like emails, phone calls, laundry… these things are like gravel. If you spend all your time on sorting the gravel, you’ll never get to the big rocks. At the beginning of the week (Monday), I move 2-3 items in this lane from my MTL, per area of my life I’m tracking (work things, husband things, father things, personal goals things). Putting more than 2 or 3 things in at a time is probably you kidding yourself, but it’s better to need more things in this lane mid-week than to feel overwhelmed by them and get few of them done. The trick is to schedule the big rocks and let the gravel fill in the open spaces. You must focus on getting the big rocks done; you need to schedule time for them and honor that. This is the area I struggle with the most; I’m in the middle of something urgent, and a time block comes up that I’m supposed to work on something else I’ve scheduled, and so of course I bump it because whatever I’m working on is on fire. Scheduling is hard :/

Doing represents what I’m doing at this moment. It sounds silly sometimes but I find it helpful to put myself in that mental state of “this is what I’m working on”. Maybe it’s better to think of it this way; I’m not working on those other things. This Trello lane should not have more than 2 things in it at any time. Sometimes I have to shuffle things in and out of here as priorities change and random things pop up. That’s OK, life can be unpredictable. The key here is to focus on these 1 to 2 things in this lane, until they are done, or for whatever reason(s) you need to change priorities. If that happens, you need to move the card out of the lane.

Done is self explanatory, you’ve done a thing! Hopefully many things!

You can tell that Monday planning plays a key roll here. Every Monday morning I set aside 10-30 minutes for weekly planning. I review my MTL list and decide if all the cards in there still represent things that I need to do, and have consequences if I don’t. Priorities and needs change, so sometimes I get to remove a card because it isn’t important anymore, or has been accomplished by some other means. Occasionally you find a card that’s been in there for a few weeks; at that point I delete the card. If it’s been on my list for weeks and I haven’t made serious effort to do it, then it can’t really be important. This self-reflection is important, otherwise you just end up with a giant, somewhat meaningless list of things, many of which you aren’t doing.

I review my Done list and compare it with what remains in my This Week list. Which has more cards (taking into consideration not every card has the same weight)? Did I over/under schedule last week? Or did I honestly just not accomplish what I could have? These are important questions to ask as you look to planning this coming week.

After some review and honest reflection, I archive all the cards in Done and move all the cards from This Week into MTL, and the Monday process restarts. This is not a perfect process by any means, but for me it’s good enough. So, what are you doing? Does what you’re doing really matter?

Shoulders are weird :: posts

2016-01-12T00:00:00-06:00

Specifically, rotator cuffs, because that’s what I messed up in my right arm.

My wife and I have become avid CrossFitters. I won’t say it’s the best thing for everyone, everywhere, but it’s working great for us. Specifically, it’s engaging and motivating, and we have a great community. But that’s another post…

So, weird shoulders. As it was explained to me, your leg is inside your hip, basically. If you were to remove your muscle and tendons, your leg bone would still be inside your hip, or at least, it would require some force to separate the too. No so for the shoulder. If you remove the muscles and tendons, your arm would just flop away, nothing remotely connecting the two. Your arm is held inside the shoulder “area” by for main muscles plus some tendons and other miscellaneous things. These four main muscles work together to keep your shoulder in one place and working together. If any of these muscles get out of whack, things go bad. Straining or pulling a muscle there can produce pain in a very wide area and variety of places, as muscles are strained and end up pinching nerves in seemingly unrelated areas of your arm. For instance, I experienced a lot of pain in my arm between the elbow and shoulder, but the actual injury was behind/below my shoulder. Shoulders are weird.

Right around memorial day of 2015 I started to learn how to do butterfly pull-ups . I’ll be honest in that I wanted to learn how to do them because they looked awesome, and at the same time sped up pull-ups. I had been doing crossfit for about 1 year by then, and while I didn’t think I was the best in the gym by far, I still thought I was strong enough.

I was half right.

I did in fact learn butterfly pull-ups, but could only do 10 or so unbroken. Unfortunately, while I could do them, I don’t know that I was doing them particularly well. They require a lot of shoulder strength, and when you’re fatigued, they can be particularly rough on your shoulder. I don’t recall a specific “oh crap I just broke/tore/strained something” moment, but I’m 95% sure that butterfly pull-ups lead to my eventual “rotator cuff tendinitis”. Interestingly enough, the pain/discomfort that began to plague me the following 7 months after was in my arms and not my shoulders.

I started feeling very tight in both my arms, particularly between the shoulder and elbow, basically identical spots in both arms. I would describe it as a nerve pain, mostly when my arms were progressing to a full extension above my head. For a month or two, I just used a lacrosse ball to massage my arms before a workout or during small breaks, and it mostly went away. Mostly.

The pain progressed to the point where some movements above my head were nearing impossible. Sharp, sudden pain caused me to drop a barbell a time or two. It even got to the point where my arms would just passively ache while sitting at my desk in the middle of the day. I decided to take 3 or so days off to recover some, take some ibuprofen, and hope my “sore” arms would loosen up and the pain would recede. After 5 days of not exercising and my arms still aching after basically no activity, I decided to see a doctor.

The Dr. suggested I had rotator cuff tendonitis, a rather unpleasant thing but better than having a torn tendon. The treatment was a few weeks of physical therapy and general don’t stress your arm much. Fortunately it was a fairly limited set of movements that triggered the sharp pain, so I still had use of my arm, I just couldn’t do everything.

The recovery process has been long, but I’m happy to say that I’m nearing the end I think. I did 5 weeks of physical therapy, meeting 2-3 times a week with a therapist, and learned weird things about the 4 major muscles and tendons in the shoulder and rotator cuff. Most of the pain has subsided, and I’m back to doing full workouts at CrossFit instead of scaling to avoid using my arms or being too stressful on them. The pain does resurface some, but with better use of a foam roller and stretching, it’s mostly being manageable.

I should be back at 100% by the time the CrossFit Open arrives (end Feb 2016), however, the time off and time scaling has left me with pretty weak endurance and strength. While I’m doing the full work outs, I feel like I’m just miles behind people I used to keep up with. It’s very … infuriating. I’m trying to resist the temptation to just go crazy to keep up, and instead work on pacing and endurance with good form and build up to it, but it’s proving difficult. I know that rushing it will likely just lead back to injury, but it’s incredibly difficult when you’re so frustrated from not being able to do what you used to do, or keep up with the people you used to keep pace with. It’s discouraging.

Of course, it will get better if I maintain that patience to build up the strength and endurance. The road is long but worth it in the end.

Heroes of the Storm :: posts

2016-01-06T00:00:00-06:00

I’m trying to write daily and don’t really have a set of technical topics out yet, so today I’ll write about my game of choice lately, Heroes of the Storm. Heroes is a Multiplayer Online Battle Arena (MOBA), akin to League of Legends or Defense of the Ancients. TL;DR two teams of 5 players pick characters and fight it out, leveling up their characters to unlock abilities and be able to fight more, better.

I won’t go into details on MOBAs or how they are played, you can read that in the links above. While I had heard of League and DOTA, I had never really had much interest in MOBAs. I honestly don’t recall what prompted me to try Heroes, probably my affection for StarCraft and WarCraft, but ever since my first few games I’ve been hooked, pretty bad actually.

I started out with just the free-to-play characters (they change every week), just kind of aimlessly wandering through AI games by myself, clicking talents that sounded cool at the time. Nova was the first hero I purchased with in-game gold and just loved playing. The stealth mechanic and biding my time for the right moment to strike was a ton of fun to play. Of course, I was doing it All Wrong™, like picking talents that help me wave clear… if you don’t know Nova’s play style, just know that wave clearing is not for Nova.

Eventually I got the desire to play better. I began to wonder “What’s the best (or better) way to build Nova?”. Naturally Youtube provided the answer, and the first (good) video I found was from MFPallytime. I’ve been a subscriber ever since. MFPallytime is a professional youtuber and Heroes is a common game he plays. He plays nearly all the heroes and gives general advice and strategy for both the hero he’s chosen in the video, and the game overall. I admit I spend too much time watching matches of Heroes on youtube, but I’m convinced it’s helped me become a better player. I don’t play all the heroes, many don’t appeal to me at all, but knowing just a little bit about their play style and abilities makes me better deal with them when I am playing a hero of my choice.

My current favorite heroes:

Murky – my only master skin at time of writing, just hilarious and fun to play, tilts other players
Thrall – likely my next master, currently at level 9
Tyrande – my most recent favorite, currently at level 9, but feel I need to get Thrall to master first.. it’s also speculated that Blizzard will nerf her soon so I’m already sad

Heroes I like, but I’m not really that good at or are just _fun to play:

Chen (I’m not very good at Chen)
Artanis – I wish I was better, I want to love Artanis but it hasn’t happened
Tassadar – great hero, I’m OK at, fun to play

Heroes is free to play, with options to purchase hero skins and mounts with in-game gold or actual money. At first I thought I’d never spend actual money on this game, but then they had a 50% off sale for the holidays and I went to town…

Hop on the Nexus and give it a shot :)

Daily writing. Again. Maybe. :: posts

2016-01-04T00:00:00-06:00

Trying this daily writing thing again. I attempted this last January, the idea being I’d write a little bit each day. I think it lasted 1 week. Maybe 2. Definitely didn’t make it out of February.

A lot has changed since January 2015. Soon I’ll be a father again, with our third child expected at the end of February. Still don’t feel like I know what I’m doing as a parent, but so far I’m doing OK faking it.

Goals I’d like to accomplish this year:

Author an open source library, maybe people will actually use it
Compete in the CrossFit Open, and do “better” than I did last year
Read a new sci-fi series. Diamond Age, Culture, something…

Goals for the month of January:

make a better list of goals for this year
start an open source library/tool
Finish The Senior Software Engineer
Contribute to Otto
Get better at Golang (how do you quantify this…)
Master Skin Thrall, Tyrande
Be an OK Muradin

The last two for January are kind of sad/funny, as they relate to Heroes of the Storm, a game I’m currently addicted to. But it’s fun, so setting goals is good.

Things I should think about:

Names for third child
OSS contributions and my role @ HashiCorp

Things I can write about

Working on Terraform
Golang.. something

Just a brain dump blog post at this point, trying to kick-start the writing thing…

Managing Go versions :: posts

2015-02-20T00:00:00-06:00

I've managed to limp through the dependency management ordeal¹ in Go with
only minor damage so far. Though my experience isn't extensive, I feel that Go
has so far done a good job on being backwards compatible. Upgrading from, say 1.3
to 1.4, typically just works, enough though I've had mismatched versions between
my local environment and what Godeps installs on Heroku, which went unnoticed
until just recently.

Until I got bit.

More like a scratch, but still…

It came time where I needed to practice what I preach and achieve (better) dev/prod parity. I briefly looked into tools like gvm but wanted to ask around before I installed yet another thing to magically solve my problems. After discussing with colleagues, I've adopted the strategy outlined below with minimal pain/adjusting. If you're willing to live dangerously and occasionally blast away your $GOPATH/pkg folder², supporting multiple versions in Go can be pretty easy.

Prerequisites

If you haven't read Getting Started and How to Write Go Code, then why are you here? Go do that. In fact, I wouldn't necessarily come right back here either, maybe bookmark for months later when you feel comfortable enough to muck up your environment with reckless abandon³. Assuming you followed those guides, you should have Go currently installed in /usr/local/go; We're going to delete that.

$ rm -rf /usr/local/go

Exciting, right?

Environment variables

In order to get started, we need to review important environment variables that you should already be familiar with.

$GOPATH: specifies where your workspace is located, e.g. $GOPATH/bin, $GOPATH/pkg and $GOPATH/src)
$GOROOT: where Go is actually located. Unless specified, Go assumes it lives in /usr/local/go (the thing we just deleted). Fortunately you can install Go anywhere you like so long as set the $GOROOT environment variable and add $GOROOT/bin to your $PATH

You need to setup these environment variables yourself, typically in a ~/.bashrc or a ~/.zshrc file (or whatever shell you're using),
and then make sure the relative bin folders are added to your $PATH:

# in ~/.zshrc  
export GOPATH=$HOME/Go
export GOROOT=/usr/local/go
export PATH=$PATH:$GOPATH/bin:$GOROOT/bin

Your environment should be complete now. Nothing should work though.

Downloading Go versions

The official binary distributions can be found on golang's site here (this example uses the latest for OS X 10.8+). We'll create a folder in /usr/local to store our Go versions, pull some down and extract them to their own folders, and then create a symlink to the one we want to use. For this example, we'll use the latest go 1.3.3 and go 1.4.1:

$ mkdir -p /usr/local/go-versions/go1.3
$ curl https://storage.googleapis.com/golang/go1.3.3.darwin-386-osx10.8.tar.gz | tar -xz -C /usr/local/go-versions/go1.3 --strip-components=1

$ mkdir -p /usr/local/go-versions/go1.4
$ curl https://storage.googleapis.com/golang/go1.4.1.darwin-386-osx10.8.tar.gz | tar -xz -C /usr/local/go-versions/go1.4 --strip-components=1

Now /usr/local/go-versions should have your separate versions in it and look like this:

$ ls /usr/local/go-versions
go1.3 go1.4

Using a symlink to switch versions

To support multiple versions we'll create a symbolic link where your $GOROOT points to. In order to toggle a version, you simply update the symlink at /usr/local/go (your $GOROOT) to point to the desired version. The initial creation of the link and toggling versions can be done with the same command:

$ ln -sfn /usr/local/go-versions/go1.4 /usr/local/go

This command will create a symbolic link at /usr/local/go, pointing to the /usr/local/go-versions/go1.4 directory, replacing the symlink if it already exists. Now you can toggle back and forth:

$ ln -sfn /usr/local/go-versions/go1.4 /usr/local/go
$ go version
go version go1.4.1 darwin/386

$ ln -sfn /usr/local/go-versions/go1.3 /usr/local/go
$ go version
go version go1.3.3 darwin/386

That's it! Mostly.

Caveats

Go installs the packaged objects into your $GOPATH/pkg folder, and compiled binaries in $GOPATH/bin. Not everything you install will place a binary in bin, but it will have something in pkg that mirrors its source directory. As you probably guessed, these things are compiled with the current version of Go. As a result, switching versions could result in undefined behavior.

The binaries in $GOPATH/bin however are statically compiled, meaning they have everything they need to run included in them, and do not depend on anything external at runtime. They should be safe to leave alone.

When you invoke the go tool, it will search the $GOPATH/pkg folder to find existing compiled package objects to avoid recompiling them unnecessarily. Because the objects in $GOPATH/pkg are compiled against the version of Go at that the time of compilation, they may not work if you switch versions.

To get around any potential issues (or resolve them), you can safely destroy your $GOPATH/pkg folder (not your $GOROOT/pkg folder):

$ rm -r $GOPATH/pkg

Future invocations of go will reinstall/compile packages as necessary. You will incur the additional overhead of recompiling these things (downloading from the net if needed), but at least you can be assured that they are compiled against the current version of Go that you're using.

End

Thanks to Ed Muller (@freeformz), Mark Turner (@amerine), Andrew Gwozdziewycz (@apgwoz), and Dane Harrigan (@daneharrigan) for educating me here.

That being there is no one standard for dependency management yet in the Go community ↩
I have no idea how dangerous this is or is not ↩
This shouldn't actually muck anything up ↩

Goodbye Heroku :: posts

2015-01-30T00:00:00-06:00

Today is my last day working at Heroku.

The last day I get to call myself a Herokai.

The last day I log in to the best job I’ve had to date.

I came onboard in October of 2012, a time that seems so long ago but passed too quickly. I am forever grateful for those that helped me get there. I had excellent mentors (or “bosses” as some call them), and so many brilliant and amazing co-workers I don’t have enough words to list them, let alone describe them. I learned so much at Heroku, including how far I still have to go, and leveled up several times.

The people, the culture, and mission, are all amazing.

The week following will be a funemployement week for me, then on February 9th I’ll be joining A New Team¹ and start working on some things that I’m really excited about.

I’m thankful for the opportunity Heroku and the Herokai gave me. I’m excited to see what Heroku ships in the future, and to watch Heroku continue to blaze a trail forward.

But for now, I’m off to new adventures!

| (• ◡•)|/ \(❍ᴥ❍ʋ)

I’ll save the who for another post, mostly because they won’t announce until then anyway. ↩

Get Go-ing :: posts

2015-01-23T09:56:14-06:00

I’ve mentioned previously that I was quite late to the Ruby party. It was still going strong, but the huge wave of excitement had largely peaked. Ruby remains my go-to language and I still love writing Ruby, it still gives me joy, but it’s lost a lot of hype. That’s actually a good thing, really, and Jeff Atwood explains it more succulently than I can. The crux of it is this:

Ruby isn't cool any more. Yeah, you heard me. It's not cool to write Ruby code any more. All the cool people moved on to slinging Scala and Node.js years ago. Our project isn't cool, it's just a bunch of boring old Ruby code. Personally, I'm thrilled that Ruby is now mature enough that the community no longer needs to bother with the pretense of being the coolest kid on the block. That means the rest of us who just like to Get Shit Done can roll up our sleeves and focus on the mission of building stuff with our peers rather than frantically running around trying to suss out the next shiny thing.

(emphasis mine)

Of course, I’m simultaneously in this camp, while also chasing one of the next shiny things, which for me is Go.

Learning Go

I didn’t want to be late to the next party. Maybe driven by some career survival instinct to stay relevant and hirable, or simply just driven by ego to be doing the next cool thing¹, regardless, I wanted to get going with Go.

Below are some resources that I’ve found very helpful in getting up to speed and functional with Go:

Getting oriented on golang.org:

Golang.org: Documentation – an excellent source of information straight from the source. I wouldn’t recommend going top-bottom though, so let’s skip around
Golang.org: Getting Started – for getting setup and introduced
Golang.org: How to write Go code – very important information on organizing Go code (workspaces, packages, etc)

Diving in

Go By Example – An excellent guide from @mmcgrana that will walk you through Go with clear explanations and working code samples
Golang.org: A Tour of Go – a tour of Go from the golang site. I found it less approachable than Go By Example, I but should note that it’s received an update and appears to be restructured

Reference / Books

Golang.org: Effective Go – a required reading, but in my opinion not until you’ve got the basics down and have written some code
Golang.org: Packages – documentation for all the standard library. You will get to know this site very well
Golang-book – a free book online with more good reference / explanations
Go Bootcamp – another free book online
github.com/golang/go/wiki – a wiki on the GitHub page for golang/go with further reading

Ruby jobs are not short in supply and I don’t suspect will be any time soon, so we’ll assume ego. ↩

On Not Taking Leaps :: posts

2015-01-21T14:56:14-06:00

The beginning of my career was using Perl for web development, and then on to PHP. I had great working knowledge of both but never dove too deep into the language. A few years into it I got an itch to pursue a certification for PHP, though in hindsight I wonder if it was mostly due to encouragement from my employer. I never held much stock in certifications like this, especially since the test was all web GUI point-and-click. It certified my knowledge of the language, not necessarily any skill. I failed the practice test(s) by only a few points, so I studied the parts I did poorly on, and then proceeded to the actual test. I failed that too.

As I mentioned, I don't hold much stock in certifications anyway...

I was quite late to the Ruby party. I don't recall the years, but while I was happily doing my PHP and reinventing the wheel for every project, I kept hearing about this RoR thing. A good friend and coworker dove in to RoR some and liked it, so I needed to check it out. This was right around Rails 1.x I think. At first, I made the rookie mistake in learning Rails before I learned Ruby. For whatever reason, Ruby just didn't click with me. My fingers rejected it, it was too magical.

By some chance I stumbled on Python and Django, both of which clicked better for me. Python was more C-like, which appealed to me at the time, C being the first programming language I learned. Django too had this slick admin built-in which was near perfect for most of our use cases at the consultancy we were at.

My coworker and I decided we needed to advance the consultancy into the next level, so we arranged a quick demo of both Rails and Django to our boss. We presented a simple MVC app in both, highlighting how easy CRUD and other simple things that we tend to reinvent were already done for us. "This cuts development time", "more industry standard" things, et. al. Our boss was impressed with both, so he took some time to think on which made more sense for the company, and about the direction we should go.

In the meantime, my friend and I discussed just simply doing an upcoming project in either of the frameworks, without consultation or approval. "Just make the leap" and get rolling. Ultimately though, we did not make any leaps, and waited to hear what our boss thought.

Weeks passed and no decision was made. The topic never came up again. The status quo was kept. We continued to reinvent the wheel with every project in PHP. No leaps were taken, no one leveled up. We stayed the course. Within a year, my friend left the consultancy and went on to bigger things. My career became stagnant in PHP land, reinventing wheels.

I don't blame my then boss; he's a good guy. As a small business owner, I think he felt responsible for a dozen families, and I can only imagine that taking risks like this was unsettling.

I blame me.

I wish I had asked for forgiveness, instead of
asking for permission.

I wish I had taken that leap.

Take your leap

Distributing Reads to a Follower on Heroku :: posts

2013-10-16T14:56:14-05:00

Using the Octopus gem!

TOC

Setup
Designating Replicated Models
Credit
Conclusion

Heroku allows you to easily add horizontal scaling to your application by way of adding additional dynos to meet capacity. Heroku Postgres too allows you to horizontally scale your database by adding followers to your primary database, streaming write-ahead-logs to each follower, keeping it up to date. While these followers are great for analytical purposes and as hot-standby databases, you can also use them as a part of your application for handling read-only queries¹ for your data.

You can do this manually with ActiveRecord, establishing the connection to a follower for a given query, but I recommend the Octopus gem. This gem adds sharding and replication features to ActiveRecord, and in this post I'll add that gem to a Ruby on Rails application to leverage a follower databases, distributing read operations to the follower. Horizontal scaling ftw ☜(ﾟヮﾟ☜).

Prerequisites:

Existing Ruby on Rails application on Heroku (so models are setup etc)²
A Heroku Postgres production database (Crane +) with a Follower (also Crane +)

Setup

In a normal Octopus setup you hard code your shard and replication definitions in a file called config/shards.yml. Because Heroku advocates 12 Factor applications, hardcoding your database information is not going to work. We need a a dynamic shards.yml file, as well as an initializer file, specially catered to the Heroku environment.

Start by adding the Ocotpus gem to your Gemfile and running bundle install

# Gemfile
...
gem 'at-octopus', require: 'octopus'
...

$ bundle install
Fetching gem metadata from https://rubygems.org/....

Next, addd the dynamic config/shards.yml:

This dynamically populates the appropriate configurations for Octopus to use your follower(s) for read operations.

Get the file by running this from your project root:

$ wget https://gist.github.com/catsby/6923840/raw/0aaf94ccc383951118c43b9b794fc62e427c2e51/shards.yml config/

Next, add this initializer to config/initializers:

The initializer adds some convenience methods like Octopus.followers and setups some additional logging.

Get the file by running this from your project root:

$ wget https://gist.github.com/catsby/6923632/raw/87b5abba2e22c3acf8ed35d06e0ab9ca1bd9f0d0/octopus.rb config/initializers/octopus.rb

Commit those changes in git and test out your setup locally. In development mode Octopus will simulate 2 followers for convenience:

$ foreman start
foreman | starting web on port 5000
web     | Puma starting in single mode...
web     | * Version 2.6.0, codename: Pantsuit Party
web     | * Min threads: 5, max threads: 5
web     | * Environment: development
web     | => 2 databases enabled as read-only slaves
web     |   * FOLLOWER 1
web     |   * FOLLOWER 2

Designating Replicated Models

With this setup, your followers will not be marked as 'fully replicated'; a fully replicated application will send all writes to the primary database, and send all reads to followers. This might be A Bad Idea™, depending on your application. Because of this we set fully_replicated: false in the dynamic config/shards.yml. You have to add a replicated_model call to each model you want to be a replicated model, or explicitly query them using methods to have queries sent to your followers.

# app/models/person.rb
class Person < ActiveRecord::Base
  replicated_model
end

This will configure the Person class to be replicated, and read queries will be sent to the follower.

You can explicitly use a follower as well:

$ heroku run rails console
Running `rails console` attached to terminal... up, run.2106
=> 1 database enabled as read-only slave
  * PURPLE follower
Loading production environment (Rails 4.0.0)
irb(main):001:0> Octopus.using(:purple_follower) do
irb(main):002:1* Person.first.name
irb(main):003:1> end
=> "Jessica"
irb(main):005:0> Person.using(:purple_follower).pluck(:name)
=> ["Jessica", "Leto", "Paul"]

(ﾉ^_^)ﾉ Hooray, you're set! You're now distributing reads to a follower database.

But wait, there is more

You can choose the followers you want to use for responding to read request with the environment variables SLAVE_ENABLED_FOLLOWERS and SLAVE_DISABLED_FOLLOWERS. Whitelist followers you want or blacklist the followers you don't want:

heroku config:add SLAVE_ENABLED_FOLLOWERS=PINK, CRIMSON
heroku config:add SLAVE_DISABLED_FOLLOWERS=COBALT

You should do this.

Without these variables set, all followers will be used for reads. This may be undesirable for you. One example of usage is when adding an additional follower to a live application, where the above ENV vars are used to ensure the new follower is excluded until it is sufficiently caught up to master for duty. Other people may have forks or just random databases attached to the app. Without a white/black list, those random possibly databases can be used for reads.

Conclusion

Scaling dynos on Heroku is an easy and great way to scale your application, but the scaling doesn't have to stop there. By horizontally scaling your databases, you distribute the workload placed on the primary database and build in resiliency for your application. By using Octopus to handle the connections, you gain some fault tolerance across several followers, and flexibility with which ones to use.

Credit

Thanks to Evan Prothro for writing the original dynamic shards.yml file and octopus.rb initializer file, as well as the original wiki guide for setting up Octopus on Heroku which the blog post is heavily based on. Thanks also to everyone who has contributed to the Octopus gem.

Postgres followers are read-only and cannot be written to. ↩
I made one for you: catsby/distributed-reads ↩

Dive into Elixir :: posts

2013-08-09T00:00:00-05:00

I've been learning Elixir lately, and so far it's been fascinating.
Elixir is a functional programming language with a very ruby-like
syntax. You get immutable state, actor-based concurrency model, and a
modern syntax, all running on the Erlang VM. It's my
first foray into functional programming, and I'm enjoying learning the
differences between OOP.

This post will get Elixir installed and
introduce you to two important parts of Elixir: Immutable Data and
Pattern Matching.

Getting rolling

Start your dive by installing Elixir. You need at least Erlang R16bB.
Don't worry what that is, just install a precompiled version from
here. Next, install Elixir. There are packages here, but you
should roll with Elixir from the master branch of the repo for now
(Elixir gets a little better every day):

$ brew install --HEAD elixir
==> Cloning https://github.com/elixir-lang/elixir.git
Updating /Library/Caches/Homebrew/elixir--git
==> make
🍺  /Users/clint/Developer/Cellar/elixir/HEAD: 278 files, 3.3M...
$

Now you can use interactive mode and start learning:

$ iex
Interactive Elixir (0.10.2-dev)...
iex(1)> 1 + 12
13
iex(2)> dog_name = "Nero Dog, II"
"Nero Dog, II"
iex(3)> String.length dog_name
12

Press control+c and then a to get out of that. Interactive mode
keeps a count of the commands you've ran in that session (the (1) and
(2)). At 1,000 you earn a badge and get a cookie! ¹

Now that you're up and running, let's look at some Elixir basics.

Immutable Data

Everything in Elixir is immutable. You can't append to a list; you can
only get a new list with those additional items included in it. You cannot increment a
value; you can only get a new value returned. Elixir is all about
transforming data. Here we think about data and what we're doing to
and with it; we're not thinking about instances, inheritance, state, or
encapsulation.

Let's look at variable binding, a thing that looks a lot
like normal assignment. It's not. For example:

# ruby
>> numbers = [1,2,3]
=> [1, 2, 3]

With assignment (in most OOP languages), the variable numbers is an object and
references an array of [1,2,3].

# Elixir
iex(1)> numbers = [1,2,3]
[1, 2, 3]

In Elixir, the variable numbers is now bound to [1,2,3] (which is a List in Elixir, not an array). Outwardly the
same, but a very different thing happened. Binding is more like a
assertion, or a question, "can this statement be made true?". If you think of
it more as an assertion, then this makes perfect sense:

1 # Elixir
2 iex(2)> a = 1
3 1
4 iex(3)> 1 = a
5 1
6 iex(4)> 2 = a
7 ** (MatchError) no match of right hand side value: 1
8     :erl_eval.expr/3

Doing so in Ruby will throw a syntax error. Take note of line 6; this
yields an error because a is bound to 1, and Elixir can't make 2
equal to 1 (Elixir will only change the binding of the variable on the
left side). No match for a could be found, so the assertion fails. Try
thinking "can the left hand side be made to match the
right hand side?", and then the = assignment operator becomes a matching
operator.

Immutable data is great because you never have to worry about someone else changing something
from under your feet. When you get a block of data, that's what you get. Not a shared reference to something,
you never have to worry about someone else (say another thread of execution) changing that thing you have; they
may change something similar, but not your thing. Your thought process and concerns become more local, only worrying about
what you're doing in that particular function with that particular data.

Pattern matching

Pattern matching is the process where Elixir binds values to variables,
but not your conventional "assignment". It's a fundamental part of
programming here. In the above example with a = 1 and 1 = a you saw
the very basics of pattern matching: attempting to make the thing on the left match
the thing on the right.

Some further examples, with lists (not arrays):

 1 iex(6)> animals = ['dog', 'cat', 'chocobo']
 2 ['dog', 'cat', 'chocobo']
 3 iex(7)> [friend, enemy, awesome_bird] = animals
 4 ['dog', 'cat', 'chocobo']
 5 iex(8)> friend
 6 'dog'
 7 iex(9)> enemy
 8 'cat'
 9 iex(10)> awesome_bird
10 'chocobo'
11 iex(11)>

On line 1 we make a list of 3 animals, and on line 3 we match that
list with a new list, [friend, enemy, awesome_bird]. You can then
reference each individually.

Look what happens when we try to use the same variable:

1 iex(11)> [d, c, c] = animals
2 ** (MatchError) no match of right hand side value: ['dog', 'cat', 'chocobo']
3     :erl_eval.expr/3

In something like Ruby c would simply get assigned (bound) twice
and end up as chocobo. In Elixir, we're matching patterns, and the c
variable tries to match both cat and chocobo, but can't. As a
result, the matching fails.

Another example:

1 iex(1)> [dog, name, dog] = ["Nero", "Adelbert Steiner", "Nero"]
2 ["Nero", "Adelbert Steiner", "Nero"]
3 iex(2)> dog
4 "Nero"
5 iex(3)> name
6 "Adelbert Steiner"

Here, dog matches up with Nero in the first and third position,
allowing name to become whatever it needs to
in order for the match to succeed.

Back to the ['dog', 'cat', 'chocobo'] example, maybe you're like me and you don't actually care about cats. Elixir can
help with a wildcard matcher, the underscore:

1 iex(12)> [d,_,c] = ['dog', 'cat', 'chocobo']
2 ['dog', 'cat', 'chocobo']
3 iex(13)> d
4 'dog'
5 iex(14)> c
6 'chocobo'
7 iex(15)> _
8 ** (ErlangError) erlang error: {:unbound_var, :_}
9     :erl_eval.exprs/2

On line 7 we get an error when trying to reference _, that's because
Elixir will match anything there and immediately discard it.

That's all for now. Go install Elixir if you haven't already and try out
some pattern matching to get a feel for it. Immutable data and pattern
matching are important parts of the things I'll cover in my next Dive
Into post covering functions, anonymous functions, and recursive functions.

If you like what you've seen, go pick up Programming Elixir from Dave
Thomas at The Pragmatic Bookshelf.

no, not really ↩

Running Elixir Dynamo on Heroku :: posts

2013-07-02T00:00:00-05:00

"Elixir is a functional meta-programming aware language built on top of the Erlang VM. It is a dynamic language with flexible syntax with macros support that leverages Erlang's abilities to build concurrent, distributed, fault-tolerant applications with hot code upgrades." - from http://elixir-lang.org

Elixir is a new ruby-like language created by Jose Valim, built on
top of the Erlang VM. It's caught my interest for a while now, but until
recently has been too young for me to dive in much.

Dynamo is a web framework written and running on Elixir. Dynamo goals are
performance, robustness and simplicity.

Naturally, I went about trying Dynamo out and wanted to get it running
on Heroku, which proved to be pretty easy. We start by installing
Elixir, checking out the Dynamo rep, and generating a project. From there we use the
Elixir third-party buildpack¹ and things should fall into place nicely.

Installing Elixir

The easiest way to install Elixir is with
Homebrew. With each new Elixir release, Jose submits an updated
formula for Elixir, which makes staying up to date easy.

Assuming you have Homebrew install:

$ brew install elixir
==> Downloading https://github.com/elixir-lang/elixir/archive/v0.9.3.tar.gz 
==> make
/Users/clint/Developer/Cellar/elixir/0.9.3: 273 files, 3.2M, built in 35 seconds
$

Dynamo requires rebar² to compile things, and Elixir makes it easy
to install with the included mix build tool:

$ mix local.rebar
* creating /Users/clint/.mix/rebar
$

Done!

Installing Dynamo

Dynamo, like Elixir, is alpha at this stage, so installation is pretty
manual :/

Start by cloning the repository and getting the dependencies:

$ git clone git@github.com:elixir-lang/dynamo.git
Cloning into 'dynamo'...
remote: Counting objects:
...
Resolving deltas: 100% (2016/2016), done.
$ cd dynamo
$ MIX_ENV=test mix do deps.get, test

Now create a new project:

$ mix dynamo path/to/new/project

Done! You now have a new Dynamo project template at
path/to/new/project. Change to that directory, get your
dependencies, and start the server:

$ cd path/to/new/project
$ mix deps.get
* Getting dynamo [git: "git://github.com/elixir-lang/dynamo.git"]
[...]
* Getting cowboy [git: "git://github.com/extend/cowboy.git"]
[...]
* Compiling dynamo
$ mix server
Running DynamoExample.Dynamo at http://localhost:4000 with Cowboy on dev

Done!. Visit http://localhost:4000 to see your running Dynamo app!
Type cntrl+c and then a to abort the server and get back to your
shell.

Setup the Heroku app

In order to run Elixir on Heroku, you need to utilize Heroku's awesome
third party
buildpacks
feature by setting a custom BUILDPACK_URL environment variable.
From the project directory, create a new Heroku app and specify the buildpack:

$ heroku create --buildpack "https://github.com/goshakkk/heroku-buildpack-elixir.git" [app_name]

If you already have a Heroku app created, you can just set the BUILDPACK_URL config var:

$ heroku config:add BUILDPACK_URL="https://github.com/goshakkk/heroku-buildpack-elixir.git" -a YOUR_APP

Specify Erlang/OTP and Elixir versions

The Elixir buildpack allows you to specify different Erlang versions,
but Elixir requires Erlang/OTP version R15 or greater. Specify a preferred Erlang/OTP version for the buildpack by creating a .preferred_otp_version:

$ echo "OTP_R16B" > .preferred_otp_version

By default, the Elixir buildpack uses the master branch version of Elixir. You can specify a custom branch or tag name from the https://github.com/elixir-lang/elixir repository in the .preferred_elixir_version dotfile:

$ echo "master" > .preferred_elixir_version

Setup a Procfile

Heroku needs a Procfile in order to run your application. Create a Procfile with a web process defined:

$ echo 'web: MIX_ENV=prod mix server -p $PORT' > Procfile

Important Note: Single quotes are important here. $PORT is an environment variable supplied by Heroku. If you use double quotes
in the above echo call, your local shell will try to interpolate the contents, and you'll end up with -p and not -p $PORT.

Deploy

Add and commit your changes, then push to Heroku:

$ git add Procfile .preferred_otp_version .preferred_elixir_version  
$ git commit -m "Setup for Heroku"  
$ git push heroku [[master]-----> Fetching custom git buildpack... done
-----> Elixir app detected
-----> Using Erlang/OTP OTP_R16B
[..]

-----> Installing Rebar from buildpack
-----> Using Elixir master
[...]
   Compiled lib/dynamo_example/dynamo.ex
   Compiled web/routers/application_router.ex
   Generated DynamoExample.Dynamo.CompiledTemplates
   Generated dynamo_example.app
-----> Discovering process types
       Procfile declares types -> web

-----> Compiled slug size: 52.7MB
-----> Launching... done, v6
       http://dynamo-example.herokuapp.com deployed to Heroku

Done!

You now have a working Dynamo app on Heroku!

What's next?

No idea. Elixir and Dynamo are both really young. While a native
database interface is planned, there's nothing out at the time of this
writing. Personal attemps to use
BossDB have failed. If I get it
working, I'll update here.

Until then, get hacking.

https://github.com/goshakkk/heroku-buildpack-elixir ↩
Rebar is a build-tool for Erlang projects, from the folks at Basho. ↩

Clint's Guide to ASCII Issue Status Codes :: posts

2013-01-24T10:43:47-06:00

Status | ASCII
----|-----
Bug: | ౿(ఠ_ఠఎ)

Frustrating bug: | (╯°□°）╯︵ ┻━┻

Infuriating bug: | (ノಠ益ಠ)ノ彡┻━┻

Fixed: | ┬─┬ノ( º _ ºノ)

Won't fix: | (╯°□°）╯︵ sıɥʇ

User error: | (╯°□°）╯︵ ╯(.□.╯)

Feature request: | ☜(ﾟヮﾟ☜)

Resolved: | (ﾉ^_^)ﾉ

TIL: Ruby stuff II :: posts

2012-12-21T08:37:31-06:00

Continuing adventures in The Ruby Programming Language

I see you

Ruby doesn't like when you use variables or methods that aren't defined:

Ruby ❯ irb
irb(main):001:0> a 
NameError: undefined local variable or method `a' for main:Object  
  from (irb):2  
  from /Users/clint/.rbenv/versions/1.9.3-dev/bin/irb:12:in `<main>'

Makes sense – a doesn't exist. However, if the interpreter has at
least seen a, things will be different:

Ruby ❯ irb
irb(main):003:0> if false
irb(main):004:1> a = true
irb(main):005:1> end
=> nil
irb(main):006:0> a
=> nil

Above, a is never assigned, but it's still seen by the interpreter, so
it still exists in some way.

Connected :: posts

2012-11-06T10:48:09-06:00

This was passed on to me today:

It seems that one day a person asked the Buddha “How do you and your monks practice your path?”.

To this, the Buddha replied “We walk, we sit and we eat.”

“But everyone walks, sits and eats!”

“However, when we walk, we know that we are walking. When we sit, we know that we are sitting, and when we eat, we know that we are eating”

Being mindful of what is happening in the here and now is very important. Allowing past emotions, or future fears to rule you is counterproductive. By being fully engaged in the present, we are better parents, employees, and citizens.

Living Buddha, Living Christ

The extent that we are connected in our lives
is not unlike an addiction. It's hard, really hard, to pull yourself
out of the net and take an objective look at just how plugged in you
are. How much you have just passed through, as if on a type of
autopilot, without consciously experiencing.

Ask yourself tonight as you get into bed:

What did I experience today?

What did I do?

Ruby 2 on OS X Mountain Lion :: posts

2012-11-02T08:49:59-05:00

Ruby 2.0.0-preview1 is out “@mametter: ruby 2.0.0-preview1 をリリースしました blade.nagaokaut.ac.jp/cgi-bin/scat.r…”
— Aaron Patterson (@tenderlove) November 2, 2012

[Updated for Ruby 2.0.0-rc1 on 10 January 2013]

Ruby 2.0 release candidate 1 is out! If you're like me you want to try out the new
shiny stuff before it's all done, you can install it now easily with
rbenv and Homebrew.

Warning

My first attempts at installing failed (quietly). Things seemed to
work, but I found this in the configure output:

Noticed this while compiling:"Ignore OpenSSL broken by Apple.Please use another openssl."ಠ_ಠ
— Clint Shryock (@ctshryock) November 2, 2012

OS X ships with an older version of OpenSSL. Even in Mountain Lion.
Homebrew makes it easy to install a new version, but does so cautiously;
after you install OpenSSL, Homebrew displays a message detailing that it
did not automatically link it. I chose to link it, but as a result I
went and recompiled all of my rubies with rbenv and the new OpenSSL. You may or may not need
to do this. It may or may not break other things. Be warned.

Installing

I assuming you're using rbenv, Homebrew, and an updated
OpenSSL.

Download and extract the preview release: ruby-2.0.0-rc1.tar.bz2

With your favorite shell, navigate to the extracted folder and configure -> make -> make install. You need to set the --prefix flag to install it along your other rbenv rubies:

~❯ cd ~/Downloads/ruby-2.0.0-rc1
ruby-2.0.0-rc1 ❯ ./configure --prefix=$HOME/.rbenv/versions/2.0.0-rc1

checking build system type... x86_64-apple-darwin12.2.0
...
...
config.status: creating ruby-2.0.pc
ruby-2.0.0-rc1 ❯ make
...
...
Elapsed: 0.1s
ruby-2.0.0-rc1 ❯ make install
...
...
ruby-2.0.0-rc1 ❯

$\$\$ Profit $$$

You can now use Ruby 2.0 via rbenv:

ruby-2.0.0-rc1 ❯ rbenv local 2.0.0-rc1 
ruby-2.0.0-rc1 ❯ ruby -v
ruby 2.0.0dev (2013-01-07 trunk 38733) [x86_64-darwin12.2.1]

Done. Have fun storming the castle.

TIL: Ruby stuff :: posts

2012-10-22T09:09:11-05:00

. acts as continuation

In Ruby you can chain method calls with .

irb:> "a wild string".gsub(/ /, '+').capitalize
=> "A+wild+string"

But it works for other expressions too

x = 6
y = 4

total = x 
  + y

puts "Total: #{total}" #=> 6

total = x
  .+ y

puts "Total redux: #{total}" #=> 10

Gives you:

$❯ ruby fun.rb
Total: 6
Total redux: 10

You can use _ for Integers and Floats

irb:> int = 1_000  
=> 1000  
irb:> int.class  
=> Fixnum  
irb:> flt = 1_000.00  
=> 1000.0  
irb:> flt.class  
=> Float

Improves readability of your code

Running Rails with Puma and JRuby on Heroku :: posts

2012-07-23T08:22:18-05:00

Last post I demonstrated setting up a Ruby on Rails app on Heroku,
running with the Puma server. As I mentioned before, in order to
get the most out of Puma you should use a Ruby implementation with real
threads like JRuby or Rubinius 2. In this post we'll repeat
most of the previous steps, but this time we'll use a Third-Party Buildpack
and get setup with JRuby.

Create a new Rails 3.x app with JRuby

Note: I'm not covering the specifics other than telling you to use RVM or rbenv to setup JRuby (I'm running 1.6.7.2)

First create a new rails app

~$ jruby -S rails new heroku-jruby-puma
  create 
  ...
~$ cd heroku-jruby-puma
~$ git init

If you check your Gemfile, you should see some key differences:

instead of sqlite3 gem you have activerecord-jdbcsqlite3-adapter
gem jruby-openssl included
assets group includes therubyrhino gem

Configure Rails to work with JRuby

Some adjustments are needed to get this new Rails app to play nice on
Heroku with Puma. First, we can't just use the pg gem; You need to use
the activerecord-jdbcpostgresql-adapter instead. Add puma and set
some configuration variables for Rails and we're set.

Gemfile

...
gem 'activerecord-jdbcpostgresql-adapter'
gem 'puma'
...
group :development do 
  gem 'foreman'
  gem 'activerecord-jdbcsqlite3-adapter'
end

Note: here I left activerecord-jdbcsqlite3-adapter in the
development group just as an example.

config/application.rb

...
config.assets.initialize_on_precompile = false

config/production.rb

...
config.serve_static_assets = true
...
config.threadsafe!

# Heroku log stuff
STDOUT.sync = true
config.logger = Logger.new(STDOUT)

Create Heroku app with JRuby buildpack

Heroku's Cedar stack has no native language or framework; instead your
sever is setup via a set of scripts called Buildpacks. Heroku
provides buildpacks for Ruby, Node.js, Python and a handful of others,
but you can also create or use your own Buildpack by using the --buildpack flag and providing a URL. Here we'll use JRuby's Heroku Buildpack from github.com/jruby/heroku-buildpack-jruby.

~$ heroku create heroku-jruby-puma \ 
  --buildpack https://github.com/jruby/heroku-buildpack-jruby.git  

  Creating heroku-jruby-puma ... done, stack is cedar  
  BUILDPACK_URL=https://github.com/jruby/heroku-buildpack-jruby.git
  ...
~$

Now when you push your code up, Heroku's slug compiler knows how to
package things an build them to use JRuby.

Add a Procfile

As I mentioned last time, a Procfile tells Heroku how to run your app:

web: bin/puma -p $PORT -e $RACK_ENV

Thats it! Commit your changes and push to Heroku like before:

~$ git push
  -----> Heroku receiving push
  -----> Fetching custom buildpack... done
  -----> JRuby app detected
  -----> Vendoring JRuby into slug
  -----> Installing dependencies with Bundler
  ...
  -----> Writing config/database.yml to read from DATABASE_URL
  -----> Precompiling assets
  ...
  -----> Discovering process types
       Procfile declares types -> web
  -----> Compiled slug size is 44.5MB
  -----> Launching... done, v7
         http://heroku-jruby-puma.herokuapp.com deployed to Heroku

  To git@heroku.com:heroku-jruby-puma.git
     6bced4f..383a7cc  master -> master
~$

Within a few moments you should be running on JRuby/Puma on Heroku.
Notice that your slug size is significantly larger than a normal stock
Rails app running on MRI. That's because of 4th item you see above,
"Vendoring JRuby into slug".

Congratulations! You're now running Rails with JRuby and Puma on Heroku.

Running Rails with Puma on Heroku :: posts

2012-07-12T22:26:13-05:00

02 July 2013 Puma 2 Updates:

Update Procfile
Add Configuration
Add Clustered mode
Mention threads

I've been playing around with Heroku a lot lately and noticed they
recommend using Thin for your production server when running Rails. In recent months
I've grown attached to Puma (for no qualified reasons) and wanted to see if
I can get a Rails app running on Heroku's cedar stack with Puma instead. It
turned out to be really, really simple.

It's important to keep in mind that Puma is threaded; if your code is not threadsafe, or your not using config.threadsafe!, you may have a bad time (Rails 4 is threadsafe by default).

Create a new Rails 3.x app

First create a new rails app, heroku app, and initialize a git repo:

$ rails new heroku-puma
  create 
  ...
$ cd heroku-puma
$ heroku create
  Creating high-day-4093... done, stack is cedar
  http://high-day-4093.herokuapp.com/ | git@heroku.com:high-day-4093.git
$ git init
$ git remote add heroku git@heroku.com:high-day-4093.git

Launch your favorite editor and make some standard adjustments in your `Gemfile`, just as a demonstration:

...
gem 'sqlite3'
...

becomes

...
gem 'thin'
...

## Deploy with Thin, just for fun Back in the terminal, add your files to the git repo and make your first commit, then push it up to Heroku:

$ git add .
$ git commit -m "Initial import, using Thin server"
$ git push heroku master
-----> Heroku receiving push
-----> Ruby/Rails app detected
-----> Installing dependencies using Bundler version 1.2.0.pre
Running: bundle install --without development:test --path vendor/bundle --binstubs bin/ --deployment
Fetching gem metadata from https://rubygems.org/.......
...
Installing thin (1.4.1) with native extensions
...
-----> Launching... done, v4
http://high-day-4093.herokuapp.com deployed to Heroku
$

Now that your app is launched, do a quick `curl` call and confirm it's on Thin:

$ curl -I http://high-day-4093.herokuapp.com
HTTP/1.1 200 OK
[...]
Server: thin 1.4.1 codename Chromeo
[..]
$

Hooray, we're running Thin! You can see this in the **Server** header that Thin sets. Except, we want to run Puma... ## Procfile Cedar introduced a new way to think about scaling your app; "the process model". It's a generalized approach to managing processes across Heroku's distributed environment, allowing you to tweak how and what gets run via a [Procfile][5]. Here we'll create a simple `Procfile` and tell our app to run with Puma instead of Thin. First, swap out Thin for Puma in your `Gemfile`:

...
gem 'puma'
...

[^proc]Then add a file named `Procfile` to the root of your project and add the following: web: bundle exec puma -p $PORT -e $RACK_ENV -t 0:5 Those are the instructions Heroku will use to run your `web` process.

Here $PORT and $RACK_ENV reference environment variables provided by Heroku, port being the listen port assigned by the Heroku router, and $RACK_ENV being the mode to run your app, typically 'production'. The -t flag lets you tune the threads Puma will use; here we set it to minimum 0, maximum 5, to line up with the default connection pool count from ActiveRecord.

Making sure to add the new Procfile to your repo, commit your changes
and push to Heroku again:

$ git add Procfile
$ git add Gemfile
$ git commit -m "Use Puma via a Procfile"
$ git push heroku master
  -----> Heroku receiving push
  -----> Ruby/Rails app detected
  -----> Installing dependencies using Bundler version 1.2.0.pre
  Running: bundle install --without development:test --path vendor/bundle --binstubs bin/ --deployment
  Fetching gem metadata from https://rubygems.org/.......
  ...
  Installing puma (1.4.0) with native extensions
  ...
  -----> Launching... done, v5
  http://high-day-4093.herokuapp.com deployed to Heroku

Now we run our `curl` command again:

$ curl -I http://high-day-4093.herokuapp.com
HTTP/1.1 200 OK
[...]
$

Hooray, we're not running Thin! Except it doesn't say we're running Puma either... I suppose Puma doesn't report itself like Thin and WEBrick do. Now if you run `heroku ps` you should see you have 1 web dyno running, and the statement used to run it:

$ heroku ps
=== web: bundle exec puma -p $PORT -e $RACK_ENV -t 0:5
web.1: up for 2m

Congratulations, you're running Puma! Note that Puma is said to really shine with Rubinius and JRuby where it can utilize multiple threads. Still, it should give you some benefits on MRI as well. ## Configuration As you continue to fine tune your Puma setup, you may find yourself with more than a few flags and switches in your invocation (the `web:` part of the `Procfile` in this case). Puma lets you specify a configuration file for all this, using the `-C` flag:

web: bundle exec puma -p $PORT -C ./config/puma.rb

And matching `./config/puma.rb` file:

# config/puma.rb
environment ENV['RACK_ENV']
threads 0,5

**Done!** Checkout the [sample configuration][7] or [configuration.rb][8] in the repository to see all available options.

Note: you can't configure the port in your configuration file. Not sure why, maybe a bug. ¯\_(ツ)_/¯

Clustered Mode

With it's 2.0 release, Puma learned some new tricks, key being the
addition of Clustered
Mode. Puma gains a level of concurrency by spawning workers to handle requests,
each worker with it's own set of threads. You enable clustered mode with
the -w or --workers flag, providing the number of workers you'd like
to spawn. To enable this on Heroku, simply update your Procfile:

web: bundle exec puma -p $PORT -e $RACK_ENV -w 3 -t 0:5

Or add workers to your ./config/puma.rb file:

# config/puma.rb
environment ENV['RACK_ENV']
threads 0,5

workers 3

You can optionally choose to preload your application before spinning up
your workers. Use the --preload flag or call preload_app! in your
config file.

Finally, you can hook into your workers before they boot with the
on_worker_boot method in your config file. An example of something to
do here is if you're preloading your application, you'll want to
establish your ActiveRecord connections here:

# config/puma.rb
environment ENV['RACK_ENV']
threads 0,5

workers 3
preload_app!

on_worker_boot do
  ActiveSupport.on_load(:active_record) do
    ActiveRecord::Base.establish_connection
  end
end

Done!

You've ran the gamut and setup Puma on Heroku, and threw in lots of
optional things a long the way. Checkout puma.io and
github.com/puma/puma for more info.

Now get hacking.

Be helpful; Don't be helpful :: posts

2012-03-13T14:38:52-05:00

You can often find me lurking on IRC in #github. It's taken me some time but
I've built up a lot of good git knowledge and experience and I enjoy
being able to help people out now and again.

It bugs the hell out of me when I see things like this:

[2:42pm] PersonWithQuestion: Why its not pulling? [pastie link]
[2:43pm] PersonWithQuestion: 1) From my laptop its pushed to github.com 2) In RHEL i am now pulling the latest and it gives me that above failure.
[2:45pm] ctshryock: PersonWithQuestion: Theres a change in that file on REHL
[2:46pm] PersonWithQuestion: ctshryock, its old one (git clone only)
[2:48pm] PersonWithQuestion: REHL is only downloader, wont ever commit anything just for upto date purpose.
[2:48pm] ctshryock: run `git status` on REHL
[2:50pm] PersonWithQuestion: [pastie link]    its saying to commit but i do not want it commit anything it will be tricky/risky. Is there a way to tell force pull only?
[2:51pm] ImpatientPerson: you can always go back and remove or update the commit if you need to
[2:51pm] ImpatientPerson: but another alternative is git stash
[2:52pm] OtherImpatientPerson: git stash
[2:52pm] OtherImpatientPerson: git stash pop
[2:52pm] ctshryock: the index.php has been modified by someone/thing on REHL
[2:52pm] PersonWithQuestion: ctshryock, i did that yes
[2:52pm] OtherImpatientPerson: PersonWithQuestion: git stash
[2:52pm] OtherImpatientPerson: git pull
[2:52pm] ImpatientPerson: i end up doing git stash && git pull && git stash pop
[2:52pm] OtherImpatientPerson: git stash pop
[2:52pm] OtherImpatientPerson: yup
[2:52pm] ctshryock: PersonWithQuestion: Do you want to keep those changes?
[2:53pm] ImpatientPerson: behind the scenes stash is creating a commit for the file anyway
[2:53pm] PersonWithQuestion: ctshryock, no 100% ignore and pull whatever is now in github.com i just tried : git stash && git pull && git stash pop  but it saying conflict (content): public/index.php
[2:53pm] PersonWithQuestion: wow stash is cool i think it did all the updates
[2:53pm] ctshryock: git checkout -- public/index.hphp
[2:54pm] ctshryock: That will restore public/index.php to the unmodified state
[2:54pm] PersonWithQuestion: error: path 'public/index.php' is unmerged
[2:54pm] ctshryock: then you can `git pull`
[2:54pm] ctshryock: From the checkout line above?
[2:55pm] ctshryock: oh you're in the middle of a merge conflict from doing all that stash pop nonsense
[2:57pm] ctshryock: `git merge --abort`
[2:57pm] ctshryock: git checkout -- public/index.php
[2:57pm] ctshryock: git pull

Here we have ImpatientPerson and OtherImpatientPerson rattling
things about git stash and what not before even understanding what PersonWithQuestion was even trying to accomplish. Meanwhile
PersonWithQuestion ends up mucking up their working directory
running each command you threw at them because they don't understand the
problem either.

You mean well, but try to understand the problem first before you start shouting answers.

That 80's Guy is running Mozilla :: posts

2011-11-09T08:23:51-06:00

Awhile back Mozilla announced their 2011 roadmap, taking Firefox from
v4 all the way to (atleast) v7 by years end. Today I updated to v8, not
because I use Firefox, but because I heard it was out and thought I'd
see what's new.

I hear it has integrated Twitter search. That's about all I've heard is
new.

It feels odd to jump from this

To this

But hey, whatever works

Buggers :: posts

2011-09-08T07:42:27-05:00

  "When Ender Wiggin holds our fleets in his control, when he must 
make the decisions that bring us victory or destruction, will there 
be military police to come save him if things get out of hand?
  "I fail to see the connection."
  "Obviously. But the connection is there. Ender Wiggin must believe
that no matter what happens, no adult will ever, ever step in to 
help him in any way. He must believe, to the core of his soul, that 
he can only do what he and the other children work out for 
themselves. If he does not believe that, then he will never reach 
the peak of his abilities."

-Enders Game

You can only do what you work out for yourself. No one is going to save you.

#IgniteComo :: posts

2011-08-19T12:52:47-05:00

Last night was Columbia's first Ignite event, hosted by the Social Media Club of Mid-Missouri. I had the
privilege of presenting on my passion, "Hacking".

The Media focuses on one tiny part of hacking, "cracking" as White
Hats like to call it, and labels it Hacking. The general population's
mental image of hackers follows suit.

Hacking isn't about breaking into computer systems, stealing
digital things or causing mayhem on the internet. Hacking is
exploring. Hacking is creating. Hacking is diving deep into the
essence of something in order to understand it. The act of hacking is more
akin to playfulness, cleverness, and exploration, not violating the bits
of a computer or network.

Hacking isn't wrong, bad, or illegal. It shouldn't be frowned upon or mistrusted. It should be encouraged. Hacking is curiosity at play. Hacking
is learning. Hacking is breaking down limitations and going beyond them.

And it certainly isn't isolated to computers and software.

Thanks to Scott over at PenguinSpark for his work organizing the
event.

(Ka)Pow :: posts

2011-08-03T23:07:30-05:00

If you do any Rack or Ruby on Rails development, you should checkout
Pow, the zero configuration Rack server for OS X.

Pow makes getting up and running with these apps dead simple. No running rails server
one app at a time, no Apache/Passenger configs to setup, no mucking with /etc/hosts. Pow doesn't
require special privileges (runs as your user), with no gems or
extensions to intall. Just Pow, and it installs in seconds.

From the manual:

How does it work? A few simple conventions eliminate the need for tedious configuration. Pow runs as your user on an unprivileged port, and includes both an HTTP and a DNS server. The installation process sets up a firewall rule to forward incoming requests on port 80 to Pow. It also sets up a system hook so that all DNS queries for a special top-level domain (.dev) resolve to your local machine.

The good and bad:

Any request to *.dev is sent to Pow
Pow by default hijacks port 80

You can configure that last one, and it's simple to do so, incase you do any other type of development where you already have something like Apache setup to handle those requests. Without configuring otherwise Pow will respond instead and likely not know what to do.

Setup and adding Apps

Installation requires authentication simply because it creates a new
Firewall rule.

$ curl get.pow.cx | sh
*** Installing Pow 0.3.1...
*** Installing local configuration files...
*** Starting the Pow server...
*** Performing self-test...
*** Installed

It's that easy. Uninstallation is just as easy too:

$ curl get.pow.cx/uninstall.sh | sh
Sorry to see you go. Uninstall Pow [y/n]? y
*** Uninstalling Pow...
*** Uninstalled

Pow creates a ~/.pow directory in your home folder. To setup an app
with Pow, cd there and create a symlink:

$ cd ~/.pow
$ ln -s /some/rackapp

Done! Visit http://rackapp.dev and you're rack app will load
before your awe-struck eyes.

But what about my port 80?

Yeah, you probably want that back huh...
Fortunately you have some options there, detailed on this helpful wiki on
the Pow Github project. It outlines two steps, but depending on your
desired setup, you may be fine with just the first one.

Without adjusting Apache

Create a ~/.powconfig file and put the following there:

# ~/.powconfig
export POW_DST_PORT=88

On Boot, Pow will read that file for any configuration options (more
here). This tells Pow not to use port 80, but instead use
port 88.

From there forward, just add :88 to your app urls, a la
http://rackapp.dev:88. Apache will still use port 80 and your other
apps will be fine, while any Pow app now responds to
<app-name>.dev:88.

Adjusting Apache

If you like the idea of not having to specify a port at all, you can
do the additional step of adding an Apache conf file in
/etc/apache2/other/ instructing Apache to act as a reverse proxy for
Pow instead. This allows you to have your Pow apps respond at
<app-name>.dev, while maintaining your current Apache process.

$ sudo curl https://raw.github.com/gist/1058580/zzz_pow.conf -o /etc/apache2/other/zzz_pow.conf
$ sudo apachectl restart

The naming is important because Apache will load it last. You want Apache to behave like normal for all
requests and then proxy anything that doesn't match your other host files to the Pow server.

A little Pow(der)

Pow has already made your life better, what more could you want?

You're right, cd ~/.pow and creating a symlink is asking an awful lot of you just to run multiple Rack apps simultaneously with no hassle...

So you use Powder, a handy little gem that lets you link and unlink Pow apps without
leaving the comfort of your cwd. And hey, it can give you a bit more insight to what Pow's doing too.

$ powder help
Tasks:
  powder applog       # Tails in current app
  powder config       # Shows current pow configuration
  powder down         # Disable pow
  powder help [TASK]  # Describe available tasks or one specific task
  powder install      # Installs pow
  powder link         # Link a pow
  powder list         # List current pows
  powder log          # Tails the Pow log
  powder open         # Open a pow in the browser
  powder remove       # An alias to Unlink (depreciated)
  powder restart      # Restart current pow
  powder status       # Shows current pow status
  powder uninstall    # Uninstalls pow
  powder unlink       # Unlink a pow app
  powder up           # Enable pow
  powder version      # Shows the version

Epilogue

Pow gives you really, really easy setups for your Rack apps. Now you
can work on several at the same time or support interaction between
them with ease. Coupled with one-line install/uninstall, the barrier to entry is
incredibly low too, so give it a shot. It can make your (dev) life
simpler.

Git in practice :: posts

2011-08-02T08:42:04-05:00

Last Wednesday I gave a quick lightning talk on using Git in your
everday workflow. You can find the source here on Github, and the
slideshow on the gh-pages of the CoMO Rich Web Group fork.

Microsoft and Skype :: posts

2011-05-10T00:00:00-05:00

A company whos product I use day in and day out as part of my job is
being acquired by a larger company who I don't really care for but am no doubt
indebted to. I have mixed feelings about this.

Since September 2010 I've been working at CramerDev, and a key to
our communication has been Skype. Per-project group chats as well as
video chat when needed, it's been a great tool.

We're mostly a Mac shop (another key to our success IMO). Needless to
say Skype 5 was seen as a horrible, horrible sin delivered upon us, as others eloquently put. I immediately reverted to Skype 2.8 (and you should too)

I'm not really worried about this. Skype 5 is
pretty bad, and Skype 2.x is still supported. If Microsoft screws this
up, which it may, I'm sure another competitor will shine and we'll
migrate. I've been hearing a lot of great things about Campfire
lately.

Lastly, this GigaOM article sums up my prediction for both
companies in this regard:

It won’t surprise me if Microsoft comes in for major heat on this
decision to buy Skype — and the software company could always
botch this purchase, as it often does when it buys a company. The
Skype team is also full of hired guns who are likely to move on to the
next opportunity rather than dealing with the famed Microsoft
bureaucracy.h

Git Flow at CoMo Rich Web :: posts

2011-04-28T00:00:00-05:00

Last night at the CoMo Rich Web meet-up I gave a quick lighting talk on using Git-Flow in your
projects. Afterwards Scott Davis of ThirstyHead gave a great talk
on HTML5, which I'm trying to find a link to the
slides. who's slides are available here.

You can view my talk at ctshryock.github.com/git-flow-talk, source
here.

For my slides I used ShowOff, a Sinatra app that reads
simple files and creates an HTML presentation. Having used
Keynote for my previous talk on Git, I really like the simplicity of
ShowOff, both in capabilities and actual usage. Not being terrible
graphically gifted, I often felt overwhelmed using Keynote. The
simplicity of ShowOff felt much more at home for me.

Hail the Glorious Cloud :: posts

2011-04-22T00:00:00-05:00

On April 21st 2011 Amazon's AWS experienced a lot of downtime, upwards
of 9 hours for one region. Several companies who rely on Amazon for
their cloud services experienced downtime, and even more companies who
rely on those companies were out as well. Engine Yard, Foursquare,
Hootsuite, Heroku, Quora, and Reddit, all of these had issues or were
simply out. Heroku and rstat.us for example were both completely off line.

alt="rstatus down due to sunny skies" />

"The Cloud" has been trending lately. By and large it's an awesome
thing. "It" makes some people fortunes, while saving others..
fortunes. Of course, "The Cloud" is not so ubiquitous and infallible as
we've come to believe. As more and more people became aware of "The
Cloud" (which is still "The Internet"), more and more people fell to
believing the Fallacies of Distributed Computing.

I believe I first came across this concept when reading Joel on
Software. Essentially, as the internet becomes further entrenched
in our lives, we begin to assume it's just always there. We begin
treating remote resources as local resources. Except they're not.

It's unfortunate that these services had such downtime. I've used many
of them and have always had good experiences. But computers, networks,
servers, and "clouds" are still made by humans. They're still fallible.

On a positive note, the industry as a whole has been far more agile in
recent years; things like this will only spark better fail-overs, in
policy and systems, and probably very quickly. Just imagine if the
federal government had a failure like this. It would takes weeks to
sort it out, and likely such a fuss would be made in the media that
policy and politics would take the stage over whether or not it even got
fixed. Six months would pass before a new strategy would be approved,
and then another 6 before it was in place.

I had a chance to meet some of the bright minds at Engine Yard. I
have a good feeling that they've already got a new plan on the drawing
board, and wouldn't be surprised if some of it's already in motion.

There's good news though, The Cloud isn't all bad, if used cautiously:

alt="Netflix flying high on the clouds" />

#CodeConf :: posts

2011-04-11T00:00:00-05:00

Conference was awesome. There's really not much more that can be said
about it, it was great. Thanks to @luckiestmonkey,
@Github, and all the volunteers for putting it together.

The Changelog has great day by day coverage here:

And Jeff Remer has a great list of the talks and slides available

Code Conf 2011 Notes

Lanyrd has a full lineup

CodeConf 2011

Awesome Neighbors :: posts

2011-03-07T00:00:00-06:00

I'm jealous of my parent's awesome neighbors:

A (yet to be identified) Owl living in a hollowed out old dead tree
across the street

And the Redtail Hawk that apparently likes to hang out in their back yard, on a wooden vine arch _thing_ my Dad and I build when I was ~15.

Gearman (Part 2): Create jobs, add workers, do work :: posts

2011-02-21T00:00:00-06:00

Other posts in this series

Gearman (Part 1): Setup Gearmand
Gearman (Part 2): Create jobs, add workers, do work
Gearman (Part 3): More jobs, other stuff, a better title (soon after Part 2)

Now that Gearmand is setup we need to get some workers going. Gearman is itself language agnostic; it doesn't really
care if your client/worker is in Python, PHP or Ruby, to name a few. There are several libraries available to help
you get going.

PECL::Gearman

In order to get clients and workers setup using PHP you need to have libgearman installed; this came with Gearmand from my
previous post in this series. For this example I'll be using the Gearman PECL extension. I assume you already have PHP installed on your system.

By default, the configure script looks in /usr/local, /usr, and /opt/local to find the libgearman header files. If
you're using Homebrew, MacPorts, or installed from source you should have no problem.

To install using PECL, run

~$ (sudo) pecl install gearman-beta

The -beta is required since there is no full fledged stable release to use at this time.

Side Note:
This should work fine for installations of libgearman I covered previously, but if you're like myself and have your Homebrew
installed elsewhere (I use Cinderella), your header files are in a different place, and you'll need to compile the
extension by downloading the source, extracting it, and running the following

1 ~$ cd Downloads/gearman-0.7.0/gearman-0.7.0
2 ~$ phpize
3 ~$ ./configure --with-gearman=`brew --prefix`
4 ~$ make
5 ~$ (sudo) make install
6 Installing shared extensions:     /usr/lib/php/extensions/no-debug-non-zts-20090626/
7 ~$

When the installation completes, you should receive instructions on activating the extension. They typically involve updating
your php.ini file to include the extension_dir and add extension=gearman.so near the end of the file. After that,
restart your sever.

To confirm everything is OK, create a simple php file that echo's out the gearman version:

//gearman_version.php
<?php
echo "Gearman version: " . gearman_version() . PHP_EOL;
?>

And run the file

~$ php gearman_version.php
Gearman version: 0.15

Creating Jobs

Jobs are created by Clients, who send them to the Gearman server via Tasks
A Task is communication about a job, such as "run this job" or "what is the status of this job"
A Job is something the worker does
Workers continuously waiting on the job server to tell them when to start and with what arguments
Clients submit jobs and ask for status about jobs (both of those things are considered tasks)
Workers actually perform the jobs

Source here

We'll start with a very simple "reverse string" job to get going. First we'll create the client to submit the job. We do so by instantiating a new GearmanClient object,
set the server where Gearmand is running, and calling do with the job name and the workload:

//  reverse_do.php
<?php
# Create our client object.
$client= new GearmanClient();

# Add default server (localhost).
$client->addServer('localhost','4730');

echo "Sending job...\n";
# Send reverse job

$result = $client->do("reverse", "Hello World!");

echo "Result: {$result}" . PHP_EOL;
?>

You can run this on the command line like so:

~$ php reverse_do.php 
Sending job...

The client will wait for the response (there are other types of jobs where the client does not wait, called background jobs).

Now we create a worker in a similar process: instantiating a new GearmanWorker object and set the server where Gearmand is running, except now we register a method of work
that we can do, and define that function:

//  reverse_worker.php
<?php

$worker = new GearmanWorker();
$worker->addServer('localhost','4730');
$worker->addFunction('reverse', 'reverse_function');

while ($worker->work());

function reverse_function($job)
{
    //  For example purpose
    echo "Received job: {$job->handle()}" . PHP_EOL; 
    return strrev($job->workload());
}
?>

Now we run the worker from another terminal session. Because we've put the worker in an endless while loop, the worker
will connect to the server and perform jobs until we kill the process with cntrl+c

~$ php reverse_worker.php 
Received job: H:Waffles.local:11

Back on the original terminal session, where we left the client waiting, you should see an update:

~$ php reverse_do.php 
Sending job...
Result: !dlroW olleH
~$

The Gearmand server held on to the job until a worker arrived that could preform the work. Once our worker fired up, the server handed
the job over, the worker did the work, and told the server it was done and could accept more work. The server in turn passed the result
back to the client, and the client process terminated, printing the reversed string.

Gearman (Part 1): Setup Gearmand :: posts

2011-02-16T00:00:00-06:00

Other posts in this series

Gearman (Part 1): Setup Gearmand
Gearman (Part 2): Create jobs, add workers, do work
Gearman (Part 3): More jobs, other stuff, a better title (soon after Part 2)

Gearman has occupied my thoughts and time lately. Gearman is a "generic application framework to farm out work to other machines or processes that are better suited to do the work".

In short, it's a job-queue system. You send it jobs you want done, it delegates those tasks to workers who know how to handle that work. Workers
connect to the server and announce what jobs it can handle; The server doesn't care about "how" they handle that job, just that they can. The
server passes the job off to the worker and can then (conditionally) pass messages back to the client for messages like status, completions, or failures.

Classic usage examples for websites are image resizing or video conversion. Instead of the user waiting for the server to re-size or convert,
you redirect them elsewhere with a message saying it will be done soon. This lets the user go about their business, and you can delegate the
work to another machine, leaving your web server to serve webpages instead hog memory converting things.

Setup

If you're on a Mac, I highly recommend Homebrew, an excellent package manager for installing all kinds of software with ease. Getting Gearman*d* (the daemon process and client libraries) is easy:

~$ brew install gearman

If you're using MacPorts

~$ (sudo) port install gearmand

If you're on Ubuntu you can use aptitude

~$ sudo aptitude install gearman-job-server (gearman-dev)

Check your favorite Linux Distro's package manager for their equivalent.

Finally, you can always compile from source

~$ tar xzf gearmand-X.Y.tar.gz  
~$ cd gearmand-X.Y  
~$ ./configure  
~$ make  
~$ (sudo) make install

Once that finishes, you should have Gearman all setup

~$ gearmand --version
gearmand 0.15 - https://launchpad.net/gearmand

Running

Gearmand is typically run as a daemon in the background

~$ gearmand -d
~$

This causes Geramand to run to detach from the current shell and run in the background. While developing, it may be helpful to run in debugging mode

~$ gearmand -vvv
 INFO Starting up
 INFO Listening on :::4730 (6)
 INFO Listening on 0.0.0.0:4730 (7)
 INFO Creating wakeup pipe
 INFO Creating IO thread wakeup pipe
 INFO Adding event for listening socket (6)
 INFO Adding event for listening socket (7)
 INFO Adding event for wakeup pipe
 INFO Entering main event loop

By default, Geramand listens on localhost/127.0.0.1 and port 4730. There are many other options you can choose to add.

So now you have Geramand setup. Go forth, and write clients and workers. Other documentation can be found here, including an incomplete list
of libraries (for instance, it's missing gearman-ruby).

Winterpacalypse :: posts

2011-02-01T00:00:00-06:00

North America, right now (image source: NOAA)

My patio at 7am, and again at 11:30am

My deck at 7am, and again at 11:30am

It's still coming down.

GearmanHQ :: posts

2011-01-26T00:00:00-06:00

GearmanHQ is a hosted job queue management system backed by Gearman which is nearing beta. It's a new project from all of us at @CramerDev.

Visit GearmanHQ.com to sign-up for a notice when we launch our private beta.

Importing browser preferences in IE8 :: posts

2011-01-17T00:00:00-06:00

IE8 offered to import my browser settings. Firefox was the only browser available (even though I have Chrome installed).

The import was "successful".

NO disintegrations! :: posts

2011-01-11T00:00:00-06:00

Branching and Databases :: posts

2011-01-10T00:00:00-06:00

We use Git at work and make use of several concurrent topic branches. At the same time, we utilize Doctrine to keep our databases in line and updates made to them sane and manageable.

This creates interesting databases states, however, as seen above. Screenshot is from a hotfix branch where several migrations from topic branches aren't included in the source, but have been ran on my local database. This isn't a problem unique to our workflow, but with Doctrine, I'm now able to visualize it.

Making Super Awesome Oreo Balls :: posts

2010-12-24T00:00:00-06:00

It's Christmas time (technically past at the time of publish). Christmas to me means family, friends, and Super Awesome Oreo Balls.

What's a super Awesome Oreo ball? This is a super awesome Oreo ball:

Composed of Oreo and cream cheese wrapped in a layer of chocolate. Perhaps you just started drooling. This is normal.

Lucky for you making Super Awesome Oreo Balls is super easy.

Steps to reproduce

Clone the source

Box of Oreo cookies
1 24oz bag of "Chocolate coating"
1 24oz bag of "Vanilla flavored coating"
1 box of cream cheese (10 oz?)

Compress objects

Fill your blender to the top with Oreos. If you have Oreos left over, you can add them after you blend the first round down some.
Use a "pulse" blend method by hitting blend for about 1 second; this allows the Oreos that are not blended (on the top) to sift down to the blades and get chopped. Otherwise, the Oreos on bottom would just get finer and finer while chunks remain on top.

Melting: Melt your chocolate covering in the microwave. Be careful not to overcook or melt the chocolate; it needs to be hot enough to be easily stirred with a spoon, but not so melted it's like water.

Merging: Soften the cream cheese in the microwave and mix in the blended Oreos. Stir with a spoon until they are well mixed

Rounding: Grab a chunk of the mix and round it into a ball, approximately the size of a ping pong ball or smaller. The end product is pretty rich so anything bigger and you may struggle to eat the whole thing in one sitting.

Coating

Drop the little Oreo ball into the melted chocolate.
Using spoons, rotate the ball around so that it's entirely wrapped in chocolate.
Scoop the newly covered ball out with the spoons, and place on a sheet of wax paper to harden.

Wait

The hardening process takes 3-5 minutes
Go pet your dogs or something.

Drizzle

Melt the vanilla covering like you did the chocolate covering; again, not too hot.
Using a knife or spoon, dip it into the melted vanilla, and "Drizzle" or streak the vanilla over the hardened Oreo balls like you're Jackson Pollock.
Let the vanilla harden, and you're done.

Arrangement: Remove your now Super Awesome Oreo Balls and place on a nice, festive holiday plate of sorts. This is crucial to feeling better about yourself once you devour 3 in 5 minutes.

"They're holiday treats, that makes it OK"

Enjoy... with caution

Yes, they're awesome.

Yes, it's the holidays.

But seriously, take it easy with the treats.

You're gonna carry that weight

Growl Down :: posts

2010-12-17T00:00:00-06:00

I was working for a client the other day and needed to get a copy of our staging server database down to my development machine. So I ssh'd into the server, piped a mysqldump into bzip2 and then proceeded to download the file via scp. The resulting bzip was ~40MB so the download took a couple of minutes; 3 minutes isn't really a long time, but it's too long for me to just stare at a progress bar and wait while there are other things I could be doing. I thought to myself "I wish this download would send a growl notification when it finished."

Thus, Growl Down was born. Growl Down is a ruby gem that wraps scp/curl on the command line and sends a Growl notification when the transfer is complete.

Growl Down is currently in development; I don't have much ruby gem experience, so I'm learning as I go. I'm hoping for an beta release by years end, but the holidays are always busy so who knows.

Powered by Jekyll :: posts

2010-12-07T00:00:00-06:00

The team took a trip to San Francisco the first weekend of December so I took advantage of the time spent waiting in the airport to migrate my site from Movable Type to Jekyll. Although I was always happy with MT, I had started looking into Jekyll simply to understand how Github Pages works, and became attached to it's simplicity while still covering all of my needs for my personal site. This all seemed to happen at the same time I was re-imagining the purpose of this site.

purpose

So what is the purpose of this site? I started to question why I had a site, at least, why am I using a blog engine that acts as a blog and a site, instead of a site that happens to have a blog. My end goal is to have a presence on the internet; a personal "home" for my doings. I still want to write things, but more and more my writings are small snippets, links, or comments — not full out POSTS. I've started using FeedBurner to see if people even subscribe to it in the first place.

So a blog is not the focus of this site. Information about me and my projects are. I re-organized the homepage to be a simple "about" page. I figured most visitors are coming from Twitter or Github, wanting to learn more about me or a specific project of mine. Landing on the homepage for a series of posts (which were all pretty small on average) is more of an obstruction to that goal.

why Jekyll

Movable Type appealed to me at first because I liked the idea of publishing static files instead of dynamically generating the page each request. MT offers quite a bit of flexibility too, first being open source, second, the admin area lets you tweak and adjust things to your hearts content. I was able to choose from a variety of themes online and tweak it to my likings; this is a common thing among OSS blog engines out there.

So things were good, for a time. But times change, and I wanted to change, so I started tweaking my site some. Originally I had widgets on every page: a tag cloud, archives broken down by month, latest posts, and my pages. I guess at the time I thought those things were "cool" or useful, but as I saw them on other sites I began to despise some of them. I don't care for the idea of tag clouds or archives by month really. So I wanted to rip those out of my site, but having not tinkered with my layout and templates in some time, I couldn't figure out how to accomplish this. That is to say after half an hour I gave up.

I started thinking that this was not "my" site... it's the site that MT gave me based on some input and a lot of things assumed or done for me. I visit the server (via the admin), write a post or setup a page, and ask it kindly to generate some html and place it in the public folder.

I also thought it all overkill. My pages don't change often, and my writings are small and infrequent. For this I had to setup:

cgi on my webserver
a database with user and permissions
MT code in the cgi-bin folder with the correct config information
run the install file via visiting a special url
maintain MT with updates
edit and publish static css on my own
edit and publish templates via the admin interface

"Too much" I said! Jekyll is simpler and for my use cases faster. It lacks much of what Wordpress or MT provide, but much of that I don't care for or need anyway. I now write my site in Textmate and Markdown. I have things laid out in a folder structure, and Jekyll generates html code for me. A little jekyll && rsync script to upload the code to my site and I'm done.

I'll dive into my experience migrating from MT to Jekyll in my next post.

Heros :: posts

2010-11-19T04:44:25-06:00

You can't count on anyone. Especially your heros.

--Syndrome

Disasters :: posts

2010-10-18T09:23:45-05:00

Disasters are opportunities to shine.

Or perish, I suppose

me, just now

The two dollar bill / @aol.com email address relation :: posts

2010-10-15T16:19:19-05:00

I think having an @aol address is like having a 2 dollar bill. They're out there, you've had one before, but they always make you pause when you see one in the wild

me, just now

No Tweet 'till November :: posts

2010-10-11T05:00:00-05:00

I'm taking a break from Twitter, effective today, October 11th 2010, until November 1st. I've removed Twitteriffic and Tweetie from my iPhone and MacBook, as well as the Facebook app for iPhone.

Too much of my free time is spent checking updates and not doing something; paying attention to a conversation, doing something constructive, noticing things around me. An hour wouldn't go by where I didn't check some status somewhere. I've thought about it and I'm simply too connected and not balancing; I need to unplug.

See you in November!

Ghost gem :: posts

2010-09-30T10:02:35-05:00

I typically develop sites on my local machine using custom vhosts in Apache and adding lines to my /etc/hosts file to point custom urls to 127.0.0.1. After some time that file grew to be large and sometimes tedious to update. Enter ghost gem. Simple hostname management.

1 Waffles:~ clint$ ghost modify pup.local 127.0.0.1  
2   [Modifying] pup.local -> 127.0.0.1

Adds a new entry (somewhere, not /etc/hosts though) and you can now use puppy.local as a url to point to your machine with a vhost setup to respond to that scheme.

You can use this for just about any url or IP you want too:

1 Waffles:~ clint$ ghost add puppy google.com  
2   [Adding] puppy -> 74.125.65.105

ghost list shows what you have already:

1 Waffles:~ clint$ ghost list  
2   Listing 2 host(s):  
3     pup.local -> 127.0.0.1  
4     puppy -> 74.125.65.105

Full details and commands are listed on bjeanes github repo.

Getting acquainted with Cinderella :: posts

2010-09-18T13:57:12-05:00

For over 3 years I've used MacPorts to setup my development environment on Mac OS X, but with a recent reformat, I decided to forego port in favor of brew. I've switched to Homebrew for my package management, and I've gone with Cinderella specifically for web development.

Cinderella was appealing for a few reasons, mainly I got quite a few awesome tools such as node, redis, memcached and mondodb, while following the lead of Homebrew and using the existing tools I already had on a stock MacBook Pro, such as apache and php.

I immediately took issue with one thing though; Cinderella starts all of these services as Launchd items in ~/Library/LaunchAgents with the RunAtLoad and KeepAlive options set. I wouldn't mind if this was a workstation, but it's my laptop and I don't always want those things running in the background
UPDATE: I made a ticket on the projects home and the developer responded back with a good point; it's better to have it "just work" than muddy it with configuration options. That said, I no longer take issue with with RunAtLoad option.

Easy enough to change though, just edit the files in ~/Library/LaunchAgents and turn those keys to false and restart your machine and you'll be fine. Note that updating Cinderella will revert these changes, so until that's a configuration option you'll have to reset it each time.

Cinderella sets up PostgreSQL and MySQL for you, but postgres is ran as your normal user, instead of a postgres user like MacPorts. This was actually a relief to me. Maybe that's not "the way it should be", but this is a dev environment not production. I did however miss that this meant the default user was also gone... so psql -U postgres was returning a fatal error because the role postgres didn't exist. Cinderella actually creates a default user of your own username, so psql -U whoami will log you in, and if you're like me, you can then create your default user and feel right at home.

Cinderella is under development and hosted on Github, so head over there and get some hardcore forking action

"Programming" :: posts

2010-09-16T08:40:05-05:00

Faces.app update :: posts

2010-07-26T08:53:53-05:00

I'm pulling the download for Faces.app for a while. Twitter has been having a lot of issues lately with their avatar uploading, both from the API and even the site itself. At several points it has been completely disabled from their end while they "fix" it, however the same issues seem to reoccur.

Until Twitter gets this more stable and I get oauth fully integrated, I can't continue to make Faces.app available.

See these threads for more info

Github Notifier 1.0 :: posts

2010-07-25T13:54:00-05:00

I just released version 1.0 of GithubNotifier, a little menu-bar app I've been tinkering with for the past week or so. It uses your Github username and API token to check your repositories for any activity in your networks.

I built GithubNotifier mostly just to tinker with their API. I leveraged CocoaREST from Steven Degutis and added support for Github's API as I went. You can check those addtions on my fork here.

There are some caveats though. For starters, the network data API from Github doesn't show what branch a given commit was made to, so if I pushed an update to branch development, the API simply details the commit, but doesn't include which branch. When the Growl notifications are made, a click-back is created so that if a user clicks on the message your default browser will open to user/repo/commits, but if the commit was on a separate branch, you wont see it initially. You may have to look around in the branches to see more details on it.

I have a support request in to Github to include branch and owner information in the commit data returned in the network API call, but they may or may not choose to implement it.

You can checkout or fork the project here, patches are welcome.

Good work Adobe installer :: posts

2010-06-28T09:25:43-05:00

At least you tried

Using MGTwitterEngine on the desktop :: posts

2010-06-03T08:46:30-05:00

The main branch of MGTwitterEngine has been updated to support Twitter's OAuth implementation, however it's geared for the iPhone/iPad. With some simple adjustments you can easily use it on the desktop. Instead of using jdg's OAuthConsumer root project, try my fork here, which replaces the custom hmac_sha1 implementation with standard CommonCrypto, which I pulled from roustem's fork.

In a future post I'll do a walkthrough of the basic setup, from cloning the repository, setting up the submodules, and getting a working example.

Introducting Faces.app beta :: posts

2010-04-14T07:58:37-05:00

I'm finally releasing my first app, Faces.app, as a beta. Faces is an OS X desktop application that manages your profile picture across your Twitter account(s). After linking your account, Faces will pull down your current avatar and display it in the grid and from there you can add any image to the grid and upload them to your account(s) with a simple drag and drop.

Faces is a beta application and currently requires OS X 10.6 and an active Twitter account. The app will remain free for the 1.0 cycle and will remain in beta until I get OAuth implemented before the June cutoff for basic auth via Twitter's API.

~~You can download the latest here (ver 1.0b7)~~
Download temporarily removed due to issues at Twitter with profile images, see this update

MGTwitterEngine update :: posts

2010-03-23T17:09:40-05:00

Yesterday I updated the main branch of MGTwitterEngine by combining my fork with tarasis and alexrepty forks. With over 35 some commits in all, I felt lucky there was very little manual merging required.

Since Matt has created and official source of MGTwitterEngine I feel like it's gained a lot of momentum. There are 8 active forks, feature requests and bugs reports starting to accumulate in the issues list, and over 180 watchers.

Updates to CocoaREST :: posts

2010-03-18T20:53:46-05:00

I've been adding support for the Github API to my fork of CocoaREST. Right now I've added:

SDGithubTaskGetRepos: Showing a list of repositories by user
SDGithubTaskGetRepoNetwork: Showing the network for a given user/repository (shows commits by forks of that repository, or commits of forks of it's parent repository)
SDGithubTaskUserShow: show information about a user
SDGithubTaskUserUpdate: Update user information with new name, email, blog, company or location

In addition I've added a tab to the demo app to show some of these. In it you can view user information by name (SDGithubTaskUserShow) and get a list of repositories (SDGithubTaskGetRepos). The tableview is set run a new task to find forks of the selected repository (SDGithubTaskGetRepoNetwork) on selection change.

The GithubDelegate is kind of a cumbersome class in that it handles all of the task results even though the API returns different results for each task. Ideally I would use a specialized class to delegate for each task, but it's just for demonstration.

Extending CocoaRest :: posts

2010-03-08T19:00:00-06:00

My latest tinker project is a Twitterific like application that monitors your Github dashboard, showing both updates to who you follow and updates on any repositories, yours or forks of yours. From what I've seen Github's API supports all of this.

To accomplish this I'm extending Steven Degutis's CocoaREST framework to support the Github API in my fork of the project. So far I've got repository listings by username, and showing forks in the network setup. There's a lot more to integrate but I'm not sure how much I want to demo in the UI.

I also added a tab strictly for Github, with the idea being a tab for each API supported, but I haven't dug into any other API beyond Twitter or Github's I'm not sure when those will come. Patches welcome of course.

Cappuccino web framework :: posts

2010-02-25T06:09:56-06:00

@rentzsch: yup, I'm sold. Awesome tech+community.

I attended C4 2 back in 2008 where the crew from 280 North gave a presentation on the Cappuccino web framework. I have mixed feelings about Cappuccino, the technology itself seems amazing but it's a big leap from "normal" users web app experience. I plan to spend a weekend with it and try it out for myself.

A Long Road :: posts

2010-02-24T08:54:22-06:00

Getting noticed may be the first step. Even if the road is long, you need to make the first step.

Geeks vs. Nerds :: posts

2010-02-16T10:25:13-06:00

Blake : found this extremely helpful: http://szeryf.wordpress.com/2010/02/15/textmate-shortcuts-you-should-be-using/

Clint : nice

Clint : I'll have to bookmark this in case I ever switch back to Textmate

Blake : haha

Blake : still in Vim?

Clint : yes

Clint : LOVE it

Blake : nerd

Blake : ;)

Clint : It's pronounced "geek", du-mass

Blake : no, geek's like star wars and role playing games... nerds do extra efficient commandline nerd crap

Clint : Really? So playing D&D is geeky and not nerdy?

Blake : yeah, nerdy is smart and technical

Blake : geeky is socially awkward and obsessive about the aforementioned geeky things

Clint : I must be a rare hybrid than. All their strengths, none of their weaknesses.

Blake : geeks might even call you "day-walker"

Blake : ;)

Clint : HAH

Clint : nice

Blake : yeah, i'm REALLY proud of that one

Blake : might print this convo out and frame it

Clint : I'm blogging this

Blake : haha, awesome

The Five Minute Management Course :: posts

2010-02-10T08:27:51-06:00

(1) Not everyone who shits on you is your enemy.

(2) Not everyone who gets you out of shit is your friend.

Selected excerpts from the Five Minute Management Course

How to be Awesome :: posts

2010-01-15T08:49:07-06:00

Every now and again I revisit this post on The Art of Non-Conformity. I almost never have the time to read it straight through, but I usually pick one section, and I always read THE FIRST THING:

"The question isn't who is going to let me; it's who is going to stop me." ~ Ayn Rand

A positive note on Android :: posts

2010-01-14T07:42:21-06:00

Last night we had some friends over for dinner and drinks, and I got my first hands on experience with the Motorola Droid and I must say it was the most impressive one yet.

First, the Droid is not near a bulky and ugly as the pictures I had seen lead me to believe. For whatever reasons I thought it was larger and thicker, but it was really quite nice.

Secondly, I've never seen such a responsive Android phone; in my experiences with the G1 and a recent HTC I was not impressed with the lag/sluggishness, but the Droid seemed spot on.

It’s great to see a good implementation of Android. I haven’t heard how the Google Nexus One is, but I have high hopes for Android being a good competitor to the iPhone and iPhone OS. Competition will (hopefully) make both stronger.

Thoughts on Nexus One :: posts

2010-01-12T08:25:56-06:00

This Boy Genius Report has some thoughts on the Nexus One that I've always held (although the author seems to feel more strongly about them)

there's practically no human emotion with Google when it comes to technology. Everything is statistical and analytical. While you could argue that being this way is way superior to "feeling" and "emotion"- it might be 95% of the time -- you still will almost always lose that charm and that amazing feeling of connecting to something.

This sums up my feelings about a lot of products when compared to Apple. If you've met me, you may know I have a thing for Apple products, and it's not because I'm a blind fan-boy, instead it's because their products are easy to make a emotional connection to. There's a lot of thought, hard work and love put into most of them, and it's easy for me to make a human connection to them.

Great Quote from @rands :: posts

2009-11-20T16:24:03-06:00

An organization that wins by exercising power starts to lose the ability to win by doing better work.

Apple's Mistake

Update to ctshryock/MGTwitterEngine :: posts

2009-10-15T16:44:30-05:00

Today I posted my update to ctshryock/MGTwitterEngine which allows you to use the API's update_profile_image and update_profile_background_image. I added to the demo app a window that lets you select an image for either profile image or background image to see it in action.

I had to modify some of the internals to get this to work... originally the _sendRequestWithMethod:path:queryParameters:body:requestType:responseType: method built the NSMutableRequest object internally, but to minimize code duplication I split the initial creation of the request into a new method _baseRequestWithMethod:path:queryParameters: to do that, and added a new method _sendDataRequestWithMethod:path:queryParameters:filePath:body:requestType:responseType: . Both _sendRequest... and _sendDataRequest... call _baseRequest... to start, with the latter adding the needed form/multipart data to the request.

Handcrafted :: posts

2009-10-15T07:41:26-05:00

A comrade of mine has made a very nice custom Textmate theme. If you're using Textmate (which I use and love) you can check it out here: TextMate theme - Handcrafted

Using Zend Tool :: posts

2009-10-12T18:41:15-05:00

Zend Tool is a powerful command-line script to help generate Zend Framework "units" as I call them, from the complete initial application structure to controllers, or actions in controllers that already exist.

Though released a the incubator a few versions back, I've found it to be more stable and easier to setup just recently. Fortunately I just did a clean install and lost my setup of Zend Tool, giving me the chance to blog about it here.

I first learned how to setup Zend Tool with this guide. It details at least two ways to get Zend Tool running, plus some instructions for Windows users. My instructions will be for Mac OS X (but should work for most Unix-like systems) and for development purposes only.

Getting Zend Framework

First step is getting the framework. Zend Tool requires the Zend Framework be in your php.ini's include_path, so you need to obtain the source. The Zend Framework source is available for public checkout via subversion, and since you're developer, we'll go that route. Personally, I like to checkout a release tag and not a trunk, that way I can incrementally update from tag to tag instead of just updating from the trunk. Either should work. I use my user's Sites folder for all my development work so I'm going to check it out there.

svn co http://framework.zend.com/svn/framework/standard/tags/release-1.9.4 ZendFramework

It's not small, depending on your network, this could take awhile...

Once that's finished cd into the ZendFramework dir and the core library is in library/Zend, with the Zend Tool scripts in bin/.

Add Zend Framework to your Include Path

Second step is to add the the contents of the Zend folder to you include path. I'm using Macports so php.ini file is located in /opt/local/etc/php5/. By default the includes path was commented out and looked like this:
;include_path = ".:/php/includes"

Change that to uncomment it and add the library folder of your checkout

include_path = ".:/php/includes:/Users/clint/Sites/ZendFramework/library"

Setting up the script

The bin folder comes with 3 scripts, a shell script zf.sh, bat script for Windows users zf.bat, and a php script that they both use zf.php. zf.sh is the script I'll be using and we need it on our system PATH, which I'm going to do via symbolic link

sudo ln -s ~/Sites/ZendFramework/bin/zf.sh /usr/bin/zf

Using Zend Tool

Check to see if Zend Tool is working by typing zf into the command-line:

The first thing you'll see is an error, since we didn't supply the correct number of arguments, but following that you'll get usage instructions for all the options of Zend Tool.

Now lets create a project by switching to your local server's document root and running :

zf create project test-app

Now take a look around at what was made by opening test-app

and view it on the web: http://localhost/test-app/public. You should get the default homepage setup of a Zend Framework Application:

Zend Tool can do more than just create the project structure, so lets create a controller and an additional action.

1 cd test-app
2   zf create controller tool

You can see that Zend Tool created a controller named ToolController and added view scripts for you under views/scripts/tool/

You can see the default text for that script here: http://localhost/test-app/public/tool

Now lets create an action. Zend Tool's create action method takes requires a controller name as the second arguement; if you don't specify one, it will assume you want to use the Index Controller.

zf create action example tool

Your results should be like this:

There are other options and methods in Zend Tool; if you've made this far and understand, you can figure them out

EMKeychain update :: posts

2009-10-07T12:35:08-05:00

I originally found EMKeychain a few months back and added it to my project to take over managing interactions between my app and the Keychain. It worked well until I updated to Snow Leopard and started using the LLVM GCC 4.2 compiler. Doing so threw some new errors about non-explicit casts and deprecated methods. I went in search of an update on Extendmac's site and by chance found @sdegutis's fork of the project on GitHub.

I forked the project here and started replacing my current EMKeychain setup with this new one. The newer isn't terribly different but it went from 4 files to 2, and we're not using "proxy" object anymore instead just class methods for getting / creating keychain items.

One thing I did notice is the inconsistent results, but I don't think it's due to this newer version. I created a temp project and ran a loop getting and printing a specific keychain password 10 times and got the actual password, null, or the password with some extra characters at random times. Not encouraging.

Luckily I did track down the issue; the raw char password wasn't being converted to an NSString object correctly, so I was able to write a patch using the C strncpy method. You can check out my *fixed* version with the link above, but I did find in the network what is probably a better fix with user irons fork, which I plan on integrating soon.

iPhone :: posts

2009-08-18T18:02:27-05:00

I finally sprung for an iPhone, picking up a new shiny 3Gs over the weekend. I'm still getting used to it, my previous phone was very "dumb", but honestly I'm having mixed feelings about the constant connection. I know now that I can pretty much always check Twitter, but that doesn't mean I should. Frankly there are many times I enjoy being completely disconnected, but now I can't really do that unless I leave my phone off/behind.

I'm very excited to start developing though. As the internet becomes more and more accessible through devices like this, it becomes more and more important to make our sites compatible and in some situations optimized for small screen size.

Zend Db Table in Zend Framework 1.9 :: posts

2009-08-03T10:37:46-05:00

I'm reading up on Zend Db Table as part of getting to better know the technologies we've adopted at Delta Systems Group and have found that now as of Zend Framework 1.9 you no longer "have to extend a base class and configure it to do simple operations such as selecting, inserting, updating and deleteing on a single table".

That's welcome news from the last time I took a look at Zend (circa release 1.4ish). Zend has really grown since then and it's quickly become an exciting platform for PHP developers such as myself.

Quality over Quantity :: posts

2009-07-31T07:47:39-05:00

From John Gruber at Daring Fireball::

And so Microsoft's official stance regarding Apple's growing domination of the $1,000+ market is that Apple is charging hundreds of extra dollars in pure margin -- $500 in the case Turner cited in his prepared remarks. The computers that Microsoft chooses to brag about on stage at a major conference are the $650 17-inch laptops advertised in Best Buy Sunday circulars.

There's no question that retailers sell tens of millions of cheap Windows laptops every year. But no one with a pair of eyes thinks such machines are of comparable quality to Apple MacBooks. Even without turning the machines on, anyone can see the difference in design and build quality. In fact, you don't even need eyes -- just pick them up and see which one squeaks. Apple is selling more MacBooks every quarter. Microsoft thinks it is sitting pretty because Best Buy has a 17-inch Dell for $650.

Well said.

Paul Graham and Maker's Schedule :: posts

2009-07-23T17:19:12-05:00

Paul Graham has written an informative essay on his blog titled "Maker's Schedule, Manager's Schedule" that had Twitter abuzz this morning. Definitely worth the read.

I echo his statements on how meetings can be devastating for a makers schedule. They can just take the wind out of your sales, or worse, make it hard to even get started.

Lately here at the office we've been trying to use our calendar's (Outlook, Entourage) to mark off large chunks of time for doing work for specific clients or projects that we know we've got a lot to do. It sends a message to you and others that this time is reserved already, and I'm going to be productive and get things done. So far, it's worked out pretty well.

JRFeedbackController + Growl :: posts

2009-07-20T17:08:13-05:00

Just pushed another update to JRFeedbackProvider on github. This update displays a "Thank you" sheet on successful submission of feedback. In addition, I added optional support for using the Growl framework.

To use Growl to display your thank you message, change #define USE_GROWL 0 to 1 in JRFeedbackController.h and add a new entry into your plist dictionary defining your other Growl messages. See here for more on using Growl.

ctshryock / jrfeedbackprovider :: posts

2009-07-19T11:47:26-05:00

I forked the JRFeedbackProvider project on github yesterday and made two commits that have already been pulled and merged into the main branch. They're small, one adds a regular expression check on the email address sent in the php file to check it's format as a valid email address, the other adjusts the NSTextView's NSTextStorage to remove the bold font when using showFeedbackWithBugDetails: to pre-populate the textview with bug details.

JRFeedbackProvider is a 'nonviral cocoa source for implementing an application feedback panel' . In short, it provides your users a standardized in-app method of providing feedback, from bug reports to to feature and support requests. The two main approaches are to use the default details pane, which asks basic questions for each category separated into three tabs/panes, and another option which allows the developer to pre-fil the bug report pane with details supplied from within the app.

The second patch I committed dealt with the latter feature. In the default showFeedback method, the textview was pre-filled with some basic questions in bold; "what happened, what did you expect" etc. The showFeedbackWithBugDetails route was essentially a blank text area by default with the exception of whatever the developer chose to fill it with by supplying an NSString, but this always resulted in the supplied text being in bold like the pre-filled bug report. I initially figured out that if you called [textView setString:@""]; it would 'fix' this and the following text would be a normal font weight, but I stumbled across this thread on cocoadev that sent me in another direction. So I went back and updated the method to first grab the font in use by the textStorage's (which was bold) and creating the same font using the NSFont's familyName method, but without the bold, like so:

1 NSFont *resetFontWeight = [[textView textStorage] font];
2 //  Font name: Helvetica-Bold    
3 //   Font Family Name: Helvetica
4 [[textView textStorage] setFont:  
5     [NSFont fontWithName:[resetFontWeight familyName]  
6     size:[resetFontWeight pointSize]]];
7 [textView setString:details];  
8 [resetFontWeight release];

This effectively resets the font, losing the bold.

JRFeedbackProvider :: posts

2009-07-17T07:55:50-05:00

I've taken a big interest in JRFeedbackProvider, already with two (albeit small) patches ready and I filed a new bug on the latest update. JRFeedbackProvider is a add-on to your cocoa app that will let the user easily submit feedback to you, be it a bug, feature request or support request, and all you need to do is copy the files and make a menu item connection, and use the provided php script on your server (or write your own in your language of choice).

This project has given me a great chance to get in to contributing to open source projects, it's scope is reasonable and there seems to be room for development that I can take swings at. I'd like to start a Zend project to act as a collector of these feedback reports, not as a bug tracker, but as a bug/feature/support filtering system. A place that collects reports from the wild and the developer can then make decisions to discard, reply to, or push over any issue to their bug tracker of choice.

Starting new. Again :: posts

2009-07-14T10:06:34-05:00

I've started a new site again. I enjoyed writing my last one using the Zend Framework but ultimately I don't want to maintain it, I'd rather be writing. I've tried schedules for blogging before but I think this time I'm just going to keep notes through out the day and try to post example code and other stuff I can derive from what I work on.

This site is setup using ~~Movable Type~~ Jekyll, which so far is proving to be very nice. See Powered by Jekyll for details,

@ctshryock

MTL :: posts

My Trello System

Shoulders are weird :: posts

Heroes of the Storm :: posts

Daily writing. Again. Maybe. :: posts

Managing Go versions :: posts

Prerequisites

Environment variables

Downloading Go versions

Using a symlink to switch versions

Caveats

End

Goodbye Heroku :: posts

Get Go-ing :: posts

Learning Go

Getting oriented on golang.org:

Diving in

Reference / Books

On Not Taking Leaps :: posts

Distributing Reads to a Follower on Heroku :: posts

Setup

Designating Replicated Models

But wait, there is more

Conclusion

Credit

Dive into Elixir :: posts

Getting rolling

Immutable Data

Pattern matching

Running Elixir Dynamo on Heroku :: posts

Installing Elixir

Installing Dynamo

Setup the Heroku app

Specify Erlang/OTP and Elixir versions

Setup a Procfile

Deploy

What's next?

Clint's Guide to ASCII Issue Status Codes :: posts

TIL: Ruby stuff II :: posts

I see you

Connected :: posts

Ruby 2 on OS X Mountain Lion :: posts

Warning

Installing

TIL: Ruby stuff :: posts

. acts as continuation

You can use _ for Integers and Floats

Running Rails with Puma and JRuby on Heroku :: posts

Create a new Rails 3.x app with JRuby

Configure Rails to work with JRuby

Create Heroku app with JRuby buildpack

Add a Procfile

See Also:

Running Rails with Puma on Heroku :: posts

Create a new Rails 3.x app

Clustered Mode

Be helpful; Don't be helpful :: posts

That 80's Guy is running Mozilla :: posts

Buggers :: posts

#IgniteComo :: posts

(Ka)Pow :: posts

The good and bad:

Setup and adding Apps

But what about my port 80?

Without adjusting Apache

Adjusting Apache

A little Pow(der)

Epilogue

Git in practice :: posts

Microsoft and Skype :: posts

Git Flow at CoMo Rich Web :: posts

Hail the Glorious Cloud :: posts

#CodeConf :: posts

Awesome Neighbors :: posts

Gearman (Part 2): Create jobs, add workers, do work :: posts

PECL::Gearman

Creating Jobs

Gearman (Part 1): Setup Gearmand :: posts

Setup