Grep'ping on Unix without perl-compatible gnu grep - can you do it better?

[Tommy]

Are you forced to use a system grep without perl-compatible regex support? (i.e.- no "-P")

Let's say you want to grep a file for "foo", but not if it's proceeded by a comment "#" (with possible, legal leading whitespace). How would you grep for that on Unix with it's lame grep?

I turn to Perl. Can it be done without perl in one grep? Can this be done in Perl, but more simply? Show me!

Meanwhile, at the shell...

[[ "$( cat filename | perl -e 'print for grep { /^[[:space:]]{0,}#.+?f
+oo/ } <>' )" != "" ]] && found_but_commented=1;

I guess I'm looking for a re-usable perl-construct to use as a grep substitute. Something along those lines above. Is there anything more easy than | perl -e 'print for grep { /regex/ } <>

[roboticus]

Tommy:

I'd do it like this:

$ grep -E '^[^#]*foo' boom
now is the foo for all

$ cat boom
now is the foo for all
now # foo is the
blarg

...roboticus

When your only tool is a hammer, all problems look like your thumb.

[Anonymous Monk]

Indeed, the -E switch (aka egrep) does activate a modern regexp syntax, and is found in a boatload more systems than where -P is. It does not come with all the Perl features, but it is way more comfortable than the basic 1980s dialect grep & co default to.

[parv]

You could (have) try(ied) awk on systems that has(had) a lacking grep.

[kielstirling]

perl -ne 'print if !/^#/ and /foo/' /tmp/foo.txt

is that what you want? Or using the regex for the above post

perl -ne 'print if /^[^#]*foo/' /tmp/foo.txt

[Tommy]

Very nice...except you still have to account for potential preceeding whitespace before the hashmark /^[[:space:]]{0,}#/

[Tux]

Install App::Ack ?