[ILUG] new grep

[kevin lyda]

On Mon, Mar 25, 2002 at 05:05:39PM +0000, Padraig Brady wrote:
> Why?
> 
> sed -n -e "/[0-9]/p" testfile
> 
> is the same as:
> 
> grep "[0-9]" testfile

no it's not if the file args are testfile1 and testfile2.  and that's a
good thing because if i'm editing files i don't want prefixes on the lines
i'm playing with.  however for your example, they're both the same as:

  awk '/[0-9]/{print}' testfile

or

  perl -ne '/[0-9]/ && print' testfile

or

  while read l; do case "$l"; in *[0-9]*) echo "$l" ;; esac; done < testfile

so should we roll the functionality of awk, perl and shell into sed?

and you ignored my comment about complexity.  to do a simple search you
required two flags, two slashes and a p command.  that's insane.

plus i don't see any docs concerning sed's return codes and what they
mean - rather important in scripts that use grep.

> With a common function of finding a pattern.

sometimes.  you can also have sed add a boilerplate license at the top
of each license file.  no searching, just position.  otoh, sometimes
you just want grep to say the pattern exists, not to print it:

    if grep -l foo bar > /dev/null 2>&1; then
	echo get a beer for foo
    else
	echo sing mournfully for foo
    fi

> I think it's as easy for someone to learn the p command
> in sed, as a seperate grep command. That just takes a
> minute or two. The hard bit is understanding the regular
> expressions.

personally, i like anything that reduces typing.

kevin

-- 
kevin at suberic.net          buffy: come on, can't you put your foot down?!
fork()'ed on 37058400      giles: it *is* down.
meatspace place: orbit     buffy: one of these days you're going to have to
http://suberic.net/~kevin         get a grown up car.  --inca mummy girl