Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
branch: master

Merge pull request #2 from santagada/master

I did a very small modification on how to install ack
latest commit e94f1d4cc6
Damian Janowski authored
README.pod

    sudo cpan App::Ack
    brew install ack
    sudo port install p5-app-ack
    sudo apt-get install ack-grep; sudo ln -s $(which ack-grep) /usr/local/bin/ack

Say we're working on an application. For this example we'll use the Redis website (http://redis.io), which is open source and available at (https://github.com/antirez/redis-io).

So -- I know there's a helper method called related_topics_for, because it's throwing an exception. I'd like to see where it's defined.

    $ ack related_topics_for

That was a short and simple result. Things to note:

  • Ack is ignoring everything in my .git directory, because it knows I don't care about what's in there.
  • Only 'known' file types are searched. We'll come back to this later.
  • Ack uses colored output by default. It will highlight the file name, the line matched and the match itself (this is as long as I'm not piping its output to another program -- in that case, Ack's output looks a lot like grep's). Usability wins.

Okay. I found where the method is defined, and I'm thinking the issue may be related to some library I'm using. I'd like to see everything that my code is loading. In Ruby, this means searching for the keyword require.

    $ ack require

You'll see the result is much longer than our previous match. It goes way beyond the height of my terminal. Fortunately, we can tell Ack to use a pager:

    $ echo '--pager=less' > ~/.ackrc

If you're using less(1), you'll need to tell it to allow color to be displayed. In short, if you're using Bash:

    $ echo export LESS=-RFX >> ~/.bashrc
    $ source ~/.bashrc

Check out man less to know exactly what -R, -F and -X do, but believe me it's a good general configuration for less.

Try our previous search again:

    $ ack require

It's much more usable now.

However, those search results do have some noise: the word "requires" also appears in some template files. One easy way to filter those out would be to use word boundaries. Remember: Ack is Perl, and its regular expressions are too.

    $ ack '\brequire\b'

Now we see only those cases where "require" appears by itself, surrounded either by spaces, commas, etc. But never "requires", "required", etc. Since this use case is so common, Ack has a switch for it:

    $ ack -w require

Now that we have more or less what we were looking for, it's a common need to open those files. For this use case, there's a very handy switch to only print the filenames that matched our search.

    $ ack -l -w require

A tip: it's common to run Ack a few times to filter what you really need and then wanting to open the files that matched your very last search. Assuming you're using Bash, you can do this:

    $ vim -p $(!! -l)

Hopefully we fixed the bug.

Now, say we would like to check if we're doing any browser sniffing (we all know it's bad, so we'd like to see if we can remove that). In jQuery it's common to do these checks using the object $.browser.

    $ ack $.browser

That will return no results, because in regular expressions $ is the end of the line -- so we're actually searching for things after the EOL, which doesn't make sense. One solution would be to escape (to Ack, not the shell) those characters:

    $ ack '\$\.browser'

But that's quite a bit of thought, and we may have some other control characters that we don't notice ([, -, {). So another way to do it would be:

    $ ack -Q '$.browser'

This switch makes Ack escape all necessary characters so that the string we're passing is considered a literal and not a regular expression. However, we still must quote the $.browser so the shell doesn't interpolate when it sees the $.

Ack by default returns lines it matches on. If you want to see context for the code in the results, you can use the -C flag, and provide the number of lines of context to include from above and below the match:

    $ ack -C 1 return

    ttdev.pm-37-    my $nslashes = ($str =~ tr[/][/]);
    ttdev.pm:38:    return sprintf( "%02d%s", $nslashes, $str );
    ttdev.pm-39-}

Here, ack shows the line before and after the line matching return. If you want rows before the match, use -B. For only rows after the match, use -A.

To be continued...

Something went wrong with that request. Please try again.