A couple of days ago, we pushed the latest version of our app into the UAT environment. It wasn’t too long after that when I started noticing that we were getting an increased number of errors on the box. Odd errors that didn’t seem to relate to one another. On top of that, we were getting complaints that response time had moved into the unbearably slow range.

I took a look at the server and pulled up top. Right at the top was a host of Rack processes that were vying for the top slot. An instance would launch, it would jump up to 100% cpu, and then it would keep on going. Killing the processes made things run smoothly, but they would just come back. And I wasn’t getting any information.

A bit of googling and manual reading later:

1

$ kill -s SIGABRT

This caused the process to stop (which is happy making in itself) and also placed a backtrace into the log. This exposed the infinite loop and allowed us to deploy a solution.

It’s even in the manual. Funny, that.

I can’t shake the feeling that I should probably have figured this out sooner, but it’s nice to find a new technique for information gathering.

When I’m coding and I need to remove a file, I rarely remember to use git rm instead of rm. It’s just ingrained in my fingers I guess. Unfortunately, this means that when I create my commit I end up with a bunch of files that I need to use git rm on.

It’s not exactly a hardship, but here’s a simple method for doing that:

Just put that into your .bashrc and you’re ready to go.

I found a great way to test my protected attributes while looking around online. like to code my model tests using the stock unit testing libraries. It’s clean and simple and it does everything that I need. So I took the existing rspec code and I converted it into a standard test helper.

Put this in your test_helper.rb file and you’re good to go.

Here’s what a standard unit test looks like now.

1
2
3

 test "protected attributes" do
  assert_protected_attribute User, :is_admin
 end

Note that you don’t have to instantiate anything now. No mucking about with fixtures or factories.

They only really have the beginners Git tutorials in place right now, but I can see this becoming a very useful link for me. I particularly like the fact that the information has been segregated into different levels and into different “tracks”.

I wish that more tutorial sites used a layout like this.

Cheap Branches Ease Development

Aug 23rd, 2009

Not getting data from your oracle data source? Just move your mouse and the data might show up!

Since my last post I’ve been using git more and more. It really suits the way I work a lot better than subversion. Because of this, I wanted to convert one of my projects from svn to Git. It’s a private project, and while I could pony up the cash and get a paid account on GitHub, I’d rather keep everything on something under my control.

I’ve already moved the main project over to slicehost, and since I’m already paying for the space I thought it would be a great idea to host my repositories there.

It was a hell of a lot easier than I thought it would be.

One of the requirements of the install is that I connect to the git repository over ssh. I’ve got it set up so that my ssh server is listening on port 30000 and it authenticates using a public key. No faffing about with passwords here!

Here’s what I did.

Install Git

The first thing we need to do is install git. I’m using aptitude, so this would be done like this:

1

sudo aptidude install git-core

Now you’re ready to create the actual repository.

Create Your Remote Repository

While still connected to your slice, enter the following. Change anything you need to, of course.

1

mkdir ~/repos/example.git cd ~/repos/example.git git init

It doesn’t really get much easier than that.

1

### Create Your Local Repository

Now we go to work on making the local repsitory. All of this is on your local machine:

1
2
3
4
5
6
7

mkdir example
cd example
git init
touch README
git add README
git commit -m "initial repository import"
git remote add origin ssh://username@yourdomain/example.git

In my case, the ssh section would be:

1

ssh://trevor@66.453.44.xxx:30000/home/trevor/repos/example.git

Push!

Now all you have to do is push your local changes back to the remote server. On your local machine: @ git push origin master @ That’s it. You’re done.

This is mostly based on the commonthread Quick Edit in WordPress

Apr 18th, 2009

I still work more often in Subversion than I do in git, and the commands aren’t quite under my fingers yet. So, more for my memory than anything else, here’s my current git workflow for when I’m working on my own stuff.

First, I create a local version of the repository to work on.

1

git clone my-great-repository @

Next, I’m going to create a branch for whatever it is I’m working on at the moment. I tend to be a bit of a splitter when it comes to branches. One change, feature or fix, per branch. Git is great for that.

1

git checkout -b "stuff_Im_working_on"

After that, I do a small change at a time. Lots and lots of small commits. Loop over this as many times as you need.

1

git commit -a -m "Changed the text as requested."

When I’ve finished everything and it’s time to check my code back in, I checkout my master branch and make sure that I’ve got the latest commit.

1

git checkout master git pull

At that point, I’m ready to merge my new code into the master branch, run all of my tests, and push everything back up to the origin.

1

git merge "stuff_im_working_on" git push

When I’m done, I usually just trash the branch I’ve been working on.

1

git branch -d "stuff_im_working_on"

Please note, I generally work by myself and I’m not too worried about conflicting merges and the like. This workflow would obviously have to be modified to handle that.

Recently, I started going back to the gym. Hours of working on web applications have added to a puffier physique than I would like. I find endless hours of cardio to be boring as hell, and the body builders staring into the mirror while trying to “pump” make me chuckle.

So I’m doing some powerlifting. I’m not at all interested in the competitive nature of it, but I love the fact that everything is quantifiable. And the goals are easy - increase your maximum lift in the squat, the deadlift, and the bench press.

Anyway.

As a good little programmer and emacs addict, I’ve added a simple function to my .emacs to help me keep track of what I’m lifting.

1
2
3
4
5
6
7
8
9
10

  (defun log-workout (exercise sets reps weight)
  (interactive "sExercise: \nNSets: \nNReps: \nNWeight(kg):")
  (find-file "~/Documents/Exercise/exercise-log.org")
  (end-of-buffer)
  (insert (format "| %s | %s | %d | %d | %d | %d | %d |\n"
                  (date-string)
                  exercise sets reps weight
                  (+ weight 20) ; weight plus bar (Kg)
                  (* sets reps (+ weight 20)))) ; total weight lifted (Kg)
  (save-buffer))

And while I’m at it, here’s one for my weight:

1
2
3
4
5
6
7
8
9
10

  (defun log-weight (weight)
  (interactive "NWeight: ")
  (find-file "~/Documents/Exercise/weight-log.org")
  (end-of-buffer)
  (insert
   (format "| %s | %d | %.2f |\n"
           (date-string)
           weight
           (* weight 0.453))) ; weight in Kg
  (save-buffer))

A lot of the heavy lifting is being done by org-mode. The table mode makes everything line up nicely. You’ll probably want to add a header for each column, but that’s not completely necessary.

Just pop these into your .emacs file, change the path on your files, and start logging your progress.