Shell Shortcuts



Make sure you 'man' the following functions before you use 'em - follow blindly at your own risk.

Most of these were discovered using RH Linux, though a few are from FreeBSD.

 Good Related Articles 
 
Exploring 'find' 
 


Renaming Files:
	
	rename .htm .html *.htm
	
	will replace the first instance of .htm with .html in all files that match *.htm

Replacing Text:

	all files in current directory:
	
		perl -p -i -e's/str1/str2/g' *

	all files in current directory and sub-directories recursively:

		find . -type f | xargs perl -p -i -e's/str1/str2/gi'

	count occurences of word 'str' in 'file'

		perl -n -le'$c+=/str/g; END{print $c}' file

	replace leading spaces with same number of leading tabs

		perl -p -i -e 's/^( +)/"\t" x length $1/e' file(s)

File Listings:

	count of all the files that are in current path and below, recursively:

		find . | wc -l

	count lines of php code in the current path, recursively:

		find . -name "*.php" | xargs cat | wc -l


Managing Files

	finding the 50 largest files in a directory:

	put this in your ~/.bashrc file, save it, and type "source ~/.bashrc".

	function big()
	{
		find . -printf "%10kk %p\n" 2>/dev/null | sort -rn | head -${1:-50} | more
	}

	now you'll be able to cd to a directory, type "big" and get the 50 largest files and
	their size in kb in that directory and its subdirs. the command take an optional
	argument, so "big 20" will show you the 20 largest files. note: this command can
	take a while from /



	delete all files from current dir that are zero bytes (recursively) (RH Linux, Mandrake):
		better - find . -size 0b |xargs rm
		slow --- find . -size 0c -exec rm {} \;

	delete all files whose names do not contain 'str' (recursively)
		best - grep -R "str" *|grep -Eo '^([^:]+)'|uniq|xargs rm
		slow - find . -type f | perl -ne'`rm $_` if!/str/'

Printing

	replace UNIX linebreaks with Windows linebreaks for printers that need them:
		perl -pe's:\n:\r\n:' filename | lpr -P username@printer_ip
	...but a better way is to get the 'unix2dos' tool and simply:
		dos2unix < filename | lpr -P username@printer_ip


Web Stuff

	figure out how large a single web page is (all files used on page + html)

		wget -r -nv --level=0 --ignore-tags=a --delete-after https://www.domain.com/whatever
	
		wget should then recursively grab everything the page uses in the way of images,
		style sheet, javascript, etc but without following links. you'll get a readout of
		how large all the pieces are and then wget will delete the files

vi[m]

	replace duplicate words throughout the file, confirming each (don't want to screw
	up the file, after all)

	good --- :1,$s/\(\w\+\) \1/\1/gc
	better - :%s/\v(\w+) \1/\1/gc
	best --- :%s/\v(\w+)\s\1/\1/gc

	

















Required Rambling: I did this because I realized a few years ago that *nix cmdline was vastly
more powerful than any other OS (namely Windows) that I'd ever used. But the learning curve is
pretty steep. The cmdline is like a mighty weapon... not just anyone can run on and use it to
it's full potential. Only through experience and training and making mistakes does one really
learn how to wield any tool. It is my firm belief that the commandline is much more powerful than
any GUI.

These are a few snippets I've discovered over the past year or so that have helped me get things
done on the commandline. I hope you find something useful.