Michael Mahemoff mahemoff

This is like a-b, where a and b are sets, but it effectively uses a case-insensitive comparison function.

The reason for this is that Rails/MySQL find_by is case-insensitive. So if you have a data feed where someone's changed only the case, but don't want that to actually affect anything, you can use this function to see if there's a real difference (beyond just flipping the case).

	class Document

	attr_accessor :body, :headers

	def initialize(text)
	matches = text.match /^(?:([\s\S]+)\n\s*\n)?([\s\S]+)$/
	self.body = matches[2]
	self.headers = {}
	headers_text = matches[1] \|\| ''
	headers_text.split(/\n/).map { \|header_text\|

	.is-left-aligned
	.when-left-aligned
	display: block
	.when-right-aligned
	display: none !important
	.is-right-aligned
	.when-left-aligned
	display: none !important
	.when-right-aligned
	display: block

	module Twitvise

	def search
	return unless twitter_client = client
	twitter_client.search 'groo'
	end

	private

	def client

	I couldn't get jPlayer to play on Firefox.

	At first, I thought it must be the fact that Firefox needs Flash (for MP3s) and Chrome doesn't. So I was trying to check the SWF is loading and then the config.

	Finally, I realised it was visibility of the player element. So yes, this might be something to do with Flash. But the point is, the player must be on displayed (not `display: none`). Which is tricky if you want the player to slide down...means you need to use various tricks instead of just a slideDown() or CSS animation.

	Anyway, it plays now.

	# A word splitting algorithm optimised for laying out arbitrary text into a fixed-width, variable-height, space.
	#
	# Splits the string into lines, where each line is smaller than some MAX characters (related to the width of the space).
	# If the first word is longer than MAX, it is not split, but becomes the entire line.
	# The algorithm isn't (yet) smart enough to try balancing the length of each line. To compensate for this, it tries to avoid
	# the naieve situation where the bottom line ends up being a few remainder characters, simply by reversing the order it
	# processes the string. So if anything's likely to be unbalanced, it's the top line that will be small instead of the bottom line.
	# walks through the string in reverse so as to end up with a longer string at the end.
	#
	# Caveat: Needs more testing

	I often find myself uploading the same video several times to YouTube, because I have to re-do something after I see how YouTube has processed it.

	Screenflow makes this harder work than necessary. It doesn't remember the form details, so I have to re-enter all the settings and content again...and some of them are mandatory.

	And, the old video will still be there. I realise YouTube doesn't let publishers replace a video, but you could provide a checkbox (always defaulting to off) that would automatically delete the previously uploaded video. This way, I could do a quick edit in Screenflow, just hit Publish To YouTube, and done.

	# search by tag
	tire.search(per_page: params[:per_page]\|\|20) {
	query do
	boolean do
	must { term 'keywords', keyword} if keyword.present?
	must { term 'tags', tag} if tag.present?
	must { string q } if q.present?
	end
	end
	highlight 'title', 'description'

	// Async web fetch with JSON parse

	[NSURLConnection sendAsynchronousRequest:[NSURLRequest requestWithURL:[NSURL URLWithString:@"url"]]
	queue:nil completionHandler:^(NSURLResponse r, NSData d, NSError *e) {
	NSDictionary *json = [NSJSONSerialization JSONObjectWithData:d
	options:0
	error:nil];
	}];

	// Using the camera