alias rgrep ‘find . -name \!:2 | xargs grep -i \!:1 | grep -v .svn | grep -v “~”‘

Example Usage
rgrep “whatever” “*”
rgrep “print” “*.xml” | less

If this is a problem try rsync -e ssh

Could have used something like http://www.dnsstuff.com/ (also recommended) to manually look up each url, but instead i just created a list of the addresses called “siteslist” and used the wonderful dig command from the terminal with a Bash “for loop” :


for i in $(grep . siteslist); do dig $i >> my_looked_up_sites.txt; done


This produces a nice document of information about each site (notably the ip address) and now I can clean my server with a bit more confidence.

this technique is really useful in a wide variety of situations where you have to repeatedly do X for every occurrence of Y.

For example if you needed to copy a robots.txt file into every subdirectory of a parent “/web” directory you could do something like:


for i in $(ls -l |grep '^d' | awk '{print $9}'); do cp /web/robots.txt $i; done


Which in english is:

ls -l

give a long direcotry listing


grep '^d'

display only the lines for directories


awk '{print $9}'

display only the actual directory name (the 9th column)


cp /web/robots.txt

copy this file into that directory.

It seems complicated at first but if you are a web developer dealing with lots of sites then it pays to invest in this kind of administrative swiss army knife.

As chrisb mentioned, find is extremely useful.

awk and cut are also very handy.

man for obvious reasons (man man if it’s not obvious) and because some options for commands like xargs vary from machine to machine.

Some people that don’t use shells very often miss out on some of the builtins like history — every command line you run gets numbered: you can rerun previous command lines by using !num. Example:


% history
1889 find . -name "*.php" | xargs -I% php -l %
1890 grep -rnI needle * | cut -f1 -d: | sort -u | xargs nedit &
1891 history
% !1889
find . -name "*.php" | xargs -I% php -l %
…


Of course you can press the up-arrow in most shells and go through your command history as well and be sure to take advantage of tab-completion for filenames and commands when possible.

Some additional options for the commands discussed in the article:

grep -v #will return lines that DON’T match the search string

rsync -u #update only, don’t overwrite newer files
rsync -n #dry run only, don’t actually do anything but list what files would be transferred - useful because rsync is so powerful and flexible and can let you learn and try additional options (–exclude, –delete, etc) before actually executing them.


find / -type f -size +500000k


will give you all the files on your system that are larger than 500 MB. Simple but undeniably useful when running df shows 95% full.

This list is great because most shell tutorials focus on the *real* geeky sysadmin stuff. Which is useless for the photoshop jockeys.

Personally it has only been in the last year that I realized how much it pays to learn basic command line. I would love to see other commenters post their web-developer-relevant commands.

rename .html .php *.html

substitutes .html with .php in files ending in .html.

easy-peasy.

anyway, does sed work on filenames?

@Dave: That’s a good addition thanks.