Gl.ib.ly

Git is an awesome tool for managing your code. However, it does take a while to get used to doing things in git. One question I get asked a bit too often is "how do I resolve binary merge conflicts"? Thankfully this is is pretty easy. I usually run the git mergetool command to allow me to resolve merge conflicts using a graphical interface and then manipulate the resultant files on the command line.

You can specify which editor to use in your ~/.gitconfig file. Favourites are kdiff3, vimdiff, xxdiff, and opendiff. If your on Mac OSX then you may be used to FileMerge, in which case you should be using opendiff in your .gitconfig; see below.

...
[merge]
tool = opendiff
...


So, when you are using mergetool option you will something like:


I just hit return so the editor opens, it will probably warn you the file is binary. In a separate terminal window navigate to the directory where the file is in and do an ls; you'll see something like what is below:


In this case editor-sprite.gif.REMOTE.41930.gif is the new file and editor-sprite.gif is the original file. To resolve the conflict just copy the REMOTE file over the original file. In my case it would be cp editor-sprite.gif.REMOTE.41930.gif editor-sprite.gif. However, I use the following one liner to resolve these conflicts more generically.



I have it in shell script in my path so I just run it when I need to resolve these types of conflicts. It just finds files with REMOTE in them and then issues a cp command to copy that file over the original.

Hope this helps!

Outside of git mergetool you can just add the binary files using git add file1 file2 ....

Earlier today I got a massive SCORM object that contained lots of mp3 files. They were all high quality files, so I wanted to cut them down in size for web use. For this purpose I am using ffmpeg which you can easily get for Windows, Linux or Mac. I tried the following on a Mac, but it should also work on Linux (paths permitting) and on Windows if you are using something like cygwin.

I highly recommend you get familiar with ffmpeg as it is amazingly useful. Just check out the man pages; swf to avi, avi to something that works on ipods, all a breeze!

Continue reading "Encoding audio and removing file extensions from files"

A few weeks ago I came across a blog post which showed you how to corrupt your own pdf files. That post was in response to the interweb ate my homework post from Bruce Schneier.

Anyhows, Didier's method is fairly simple -- there are many other ways todo this like truncating the file or inserting some randomness using dd and /dev/urandom. So lets automate it.

Continue reading "Auto corrupting your PDF files"

Once upon a time I was comparing some Java files. Now one way to find plagiarism among students who share program code is to look at the white space traces in a file. Students typically try and edit the source code: changing variable names and function names; however, they usually fail to conceal the original author's white space patterns. White space consists of characters which don't appear on screen such as spaces, tabs, and new line characters. Anyway you can catch lots of cheaters by looking at the white space patterns in files ... maybe that's another blog entry, so back to this one.

Continue reading "Creating an HTML file with white space traces"