Brad Dunbar

Dave Winer succintly states my feelings on collaborative software development in a quote from Ars Technica.

Eventually you come around to his way of thinking, or he comes around to yours. These are the essential ingredients in good technology. We’re looking for the right answer, not to be proven right, or to prove the other guy wrong.

Looking for the right answer, together.

Today I learned (after a long and arduous debugging session) that innerHTML can have serious side effects when used in Internet Explorer. Consider the following:

var p = document.createElement('p');
p.innerHTML = '<a>test</a>';
a = p.childNodes[0];
p.innerHTML = '';
console.log(a.innerHTML);

Here’s a fiddle in case you want to play around with it. In most modern browsers this snippet will happily log 'test' before continuing about it’s business. Internet Explorer, on the other hand, will remove the contents of a when setting p.innerHTML and log the empty string to the console. Obviously this can lead to some interesting results if you’re not expecting it. Hopefully I’ve saved you the debugging time I’ve lost today. ;)

I filed a bug concerning this behavior on the jQuery issue tracker, but it can’t be fixed due to performance considerations (more details in the ticket). That said, working around the issue by using .empty().append(…) is easy to do.

Sundays are for bear suits.

It’s official. I’m a father now. My daughter was born on Thursday, December 22^nd 2011 weighing in at 7 lbs 6 oz and measuring 19.25 inches long. I’ve never been so happy or exhausted before in my life. My wife was amazing throughout the pregnancy and birth and our family has been incredibly loving and supportive in every way imaginable. Life will never be the same.

Recently, I caused myself a giant headache because I was uninformed about echo and how it differs from printf (not to mention how it affected my code). Hopefully I can spare you the same pain. For the Software Tools Philosophy to work, one must have a working knowledge of the tools.

Alright, review time. According to man echo, it’s used to “display a line of text”. And so it does! It’s quite handy when you need to dump something to stdout. Especially when that something is literal text.

$ echo cheese and crumpets
cheese and crumpets

I often use echo for this reason, and normally it works in exactly the fashion I expect. However, sometimes what I expect is not reasonable or even logical! I tried to use echo to stitch together a screen stylesheet with a print stylesheet. (I’ve heard that many browser’s download print stylesheets even when they’re not using them. As such, it may as well come in a single http request.) Specifically, I tried to do the following:

echo -n "$(cat screen.css)\n@media print {\n$(cat print.css)\n}\n"

On the surface this looks like a perfectly reasonable approach and, given the correct circumstances, it is. echo dutifully renders the print stylesheet with a wrapping @media declaration. The problem is that the following was included in screen.css:

.clearfix:after, .container:after {
  content:"\0020"; ...
}

You’ll notice the unicode escape sequence for space (\0200). To echo this looks like an escape sequence (and so it is, just not one we’d like to replace) and so it replaces it. At this point I can hear you saying “But wait! echo has the -E switch for that! It disables interpretation of backslash escape sequences!”. You’re right. It does. It cannot, however, be used granularly. This renders it useless for my task. I need to escape the newlines surrounding the @media declaration, and yet I need to leave the unicode escapes in the stylesheet untouched.

This is where printf saves the day. It works almost identically to the version in stdio.h. It will always replace backslash escapes in the format string and never replace backslash escapes in arguments. In the end, I used the following command:

printf "%s\n@media print {\n%s\n}\n" "$(cat screen.css)" "$(cat print.css)"

This worked like a charm and saved my day. \o/

Hopefully, you’ll remember this next time you start echoing things with escape sequences in them. I have no idea how quotes within the arguments are dealt with but that’s a mystery for another day.

This excerpt from the redis manifesto is yet another reason to love redis.

Code is like a poem; it’s not just something we write to reach some practical result. Sometimes people that are far from the Redis philosophy suggest using other code written by other authors (frequently in other languages) in order to implement something Redis currently lacks. But to us this is like if Shakespeare decided to end Enrico IV using the Paradiso from the Divina Commedia. Is using any external code a bad idea? Not at all. Like in “One Thousand and One Nights” smaller self contained stories are embedded in a bigger story, we’ll be happy to use beautiful self contained libraries when needed. At the same time, when writing the Redis story we’re trying to write smaller stories that will fit in to other code.

I’m somewhat worried that my productivity will decline because redis doesn’t do the things an RDBMS does but I’m having so much fun that it hardly matters.

Honestly, I’ve found precious few instances in which I need to use javascript’s bitwise not operator. However, I stumbled across a particularly nice usage of it this morning as I was perusing the express source.

When I use indexOf to check for a substring I often feel the statement is rather clumsy, especially if the variable name is particularly long.

if (someLongVariableName.indexOf(someSubstring) == -1) {
  // do stuff
}

The == -1 part seems to get lost at the end of the expression and it takes me an extra moment to parse the code. However, I found this brilliant little nugget that makes it much easier to read.

res.contentType = function(type){
  if (!~type.indexOf('.')) type = '.' + type;
  return this.header('Content-Type', mime.type(type));
};

Taking advantage of the fact that ~(-1) == 0 (and it is the only number with this property), our statement becomes truthy or falsey and we can change our previous example to

if (!~someLongVariableName.indexOf(someSubstring)) {
  // do stuff
}

Hey! My name is brad. I’m a programmer and I hacked together this site in the hopes that two or three people would eventually check it out. Lately I’ve been thinking to myself

holy cow, node.js is suh-weet! I need to learn more about it post haste and then make a peanut butter sandwich

Strictly speaking, the peanut butter sandwich part is not relevant. It is, however, a good idea and I intend to follow up on it.

node.js is easily the shiniest thing I’ve seen in the past few days, and all the cool kids are doing it so I figured I’d join in the fun. So far I love it although it’s not without a few hurdles. Flow control seemed a bit clumsy at first, but as it turns out there are various libraries that make it much cleaner. I’ll get into details later when I’ve really dug into the platform. I’m also very interested in the v8 engine (and dynamic language engines in general) so I’m going to be making a serious attempt to learn c++ and hack on it.

So if you want to read about javscript, algorithms, computer science or peanut butter sandwiches, this is the place for you. Stop back in sometime soon and I’ll tell you all about them.