Friday, November 28, 2008

Converting ^M

I keep running into files that have weird encoding for returns.  I suspect it is a combination of optimization and obfuscation.  I particularly see this with javascript files.

Here is the fix…
First, to get the ^M which is a control M, type ctrl v followed by ctrl m.
To get ^@ type ctrl v followed by ctrl shift 2 (or ctrl v followed by ).

Second, the conversion…

In vi (or vim) use the following:

:%s/^M/\n/g

or with perl on the command line:

$ perl -pi.bak -e ’s/^M/\n/g’

^M is ASCII 13 (Ctrl M), which is the carriage return.

Different operating systems use different symbols to set the end of a line/new line.
Unix uses newline (\n)
Mac uses carriage return (\r)
And Windows/DOS use both (\n\r)

The reason I put the ^@ one on there is because some times when I convert the ^M to \n, it becomes ^@ instead, which doesn’t help.  Well, in reality it does.  You just convert the ^@ back to ^M.  When you convert the ^@ back to ^M, then the code looks all pretty.

In vi (or vim) use the following:

:%s/^@/^M/g

or with perl on the command line:

$ perl -pi.bak -e ’s/^@/^M/g’

As with all things nerdy… your mileage may vary.