No need to master sed/awk/grep/wc/... use pure Ruby -- pru

rubygeek

What does this give over "ruby -n -e 'some expression'"?

absinthe718

I really doubt that pru will perform in the same order of magnitude in either speed or memory use when compared with sed/awk/find/grep. Some of us are old enough to have learned sed/awk/find/grep in during the Reagan administration. And we got into arguments about sed/awk/find/grep versus perl in the Clinton years. This is old hat.

eternal_noob

Terrible idea. The examples are really counter-intuitive and often longer than their unix equivalent.

ps -ef | grep foo | grep -v grep
ps -ef | pru 'include?("foo") and not include?("pru")'

Second, sed, awk, et. al are useful on their own. You can't beat the convenience of awk.

cat /var/log/nginx/access_log |awk '$9 ~ 404'

It is more useful to learn the "regular" ruby equivalent instead of an obscure script. See:

ls -al | pru 'gsub(/5/,"five")'
ls -al | ruby -pe 'gsub(/5/, "five")'

Or the line count example:

ls -al | pru --reduce 'size'
ls -al | ruby -ne 'END {puts $.}'

You can't count on rubygems being present either, but short of windows sed, awk and friends is.

Baramin

I wouldn't recommend learning 'pru' instead of the classic tools, as you'll never find pru installed on a unix server anywhere, you'll have to install it yourself, meaning installing ruby first, which can be a big problem on enterprise servers.

est3est

I actually like the idea, if the goal is just to make my work done faster. If you don't like the idea of making things easier and quicker then why would you use languages like Python or Ruby, when there is already C. But to be honest, I found examples a bit confusing at the moments. Still, this have a potential, and I am very interested where it will go...

alterlaszlo999

that' what i dislike in ruby community: it seems they always reinvent the wheel 'cause they have only ONE dogma. and in a way they disvalue the unix under their bottom.

_random_guy_

Great job, kudos for that. But, there are a lot of reason why this is not a good idea - ruby is slow, for instance.

Also, if you look at the examples you've written yourself, they are less concise than the grep equivalent.

projecteternity

Fun fact, ruby regex matching is orders of magnitude slower than sed/awk/grep. The proof is kind of depressing.

rgh

Having used grep/sed/awk/... for the last 15 years or so I actually think this is a nice idea.

As a ruby developer and sysadm and I think there's a place for both. You have to remember that not everyone is going to master sed and awk. As incredibly powerful as they are the syntax is arcane (that's no necessarily a bad thing but if you don't use them regularly they can be hard to remember). So I suspect that this find a sweet spot with ruby devs.

I for one will be adding to my toolkit.

bigfig

Inline perl (perl -pe "s/\w//" etc) often does what sed won't (easily), but I still use sed. Awk never suited me as it seems to fill exactly the same niche as perl, and grep /egrep will remain go-to tools for parsing logs.