7 ms·

Beyond grep: ack, a code search tool for programmers

70 points by andrelaszlo 13y ago

michaelmior 13y ago
A similar post popped up a couple weeks ago, and one comment suggested https://github.com/ggreer/the_silver_searcher https://github.com/ggreer/the_silver_searcher. I was using ack at the time, but I'm really liking ag as I've found to be quite a bit faster. The one thing I wish it had was being able to restrict searches by file type.
- rwl4 13y ago
  I second this. My friend kept suggesting I switch from Greg to ack. I searched for it and found ag and fell in love with it's speed. That same friend has now switched too.
- bencoder 13y ago
  Yep. I switched from grep to ack then to ag. ag is incredibly fast. You can limit the files it searches with -G <pattern>
  - meowface 13y ago
    Thanks, I was looking for this as well. ag --help could probably be reworked a little bit; --include makes more sense than -G / --file-search-regex in my opinion.
  - michaelmior 13y ago
    Thanks. Didn't read the manpage too carefully. Still not quite as easy as with ack, but probably good enough for me.
  - mh- 13y ago
    for those who are frequently searching paths with a large number of files stored on an SSD or ramdisk/tmpfs, the bottleneck is very much the CPU time. in these cases ag is noticeably faster (orders of magnitude, in some cases), especially if you're just searching a literal and not with a regex pattern. the author has done some really cool performance hacks, and written some great blog posts[1] along the way. [1]: http://geoff.greer.fm/2012/09/03/profiling-ag-writing-my-own-scandir/ http://geoff.greer.fm/2012/09/03/profiling-ag-writing-my-own...
    - e12e 13y ago
      I was a bit surprised by a simple test I just ran on a Linux 3.10 tree: time rgrep aes . > /dev/null real 0m0.900s user 0m0.548s sys 0m0.340s # Somewhat similar output to ag: time rgrep -n --color aes . > /dev/null real 0m1.177s user 0m0.876s sys 0m0.288s time ag aes > /dev/null real 0m1.147s user 0m1.040s sys 0m0.548s # Using fixed strings in grep, limiting us to searching c-files time rgrep -n --color -F --include='*.c' aes . > /dev/null real 0m0.936s user 0m0.720s sys 0m0.208s time ag -G \.c aes > /dev/null real 0m1.130s user 0m1.140s sys 0m0.428s This is on an encrypted volume sitting on top of a low end SSD, all runs with hot chache. The times here are from ag in Debian -- I tried a build from upstream git -- but with essentially the same time. I guess ag might make sense under OS X -- but there doesn't appear to be any (speed) advantages under GNU/Linux.
      - mh- 13y ago
        nice, thanks for sharing. I was actually contrasting it to ack in my speed claims. I don't expect that it'd be that much faster than GNU grep in most situations.
- soveran 13y ago
  Not quite what you want, but you can restrict the search to filenames matching certain pattern with the -G option. For example, ag <pattern> -G \.c will only search on C files.
- stottc 13y ago
  You can do something like : ag -G ".txt$" hello
- mtarnovan 13y ago
  Indeed, I've also been using ack for a while, and switched to ag. Much faster.
- meowface 13y ago
  ag is indeed incredibly fast and extremely useful. It's replaced grep for me, any time I need to do a recursive search.
- ams6110 13y ago
  I'm pretty accustomed to using find | xargs grep. This lets me completely tailor the file I search using find's filters (e.g., only search files with a .py extension, and skip .svn/ directory). E.g. assuming I'm in my project's root directory: find . -name .svn -prune -o -name '*.py' -print | xargs grep -Hi 'string or RE I am looking for'
  - pyre 13y ago
    Or you can do: ack --python PATTERN [DIRECTORY]
  - michaelmior 13y ago
    I used to do the same, but this accomplishes the same thing and is much more concise ag -i -G '\.py$' 'string or RE I am looking for'
  - troels 13y ago
    I'm pretty happy with `find|xargs grep` myself. Of course, I have an alias that hides the arcane incantation away under a single `frep`, but underneath it uses those tried and tested unix tools. An alias is usually simpler to carry with me to new systems as well.
    - ams6110 13y ago
      My alias is my .bash_history :). I grant that these other tools are great, but I like using the standard utilities because, among other things, once you learn 'find' and 'xargs' you can piece together all kinds of other commands and you don't have to remember a tool-specific syntax (at least for the 'find' and 'xargs' parts).
- davidgerard 13y ago
  Yep. Came here to recommend ag. It's just fantastic.
allannienhuis 13y ago
ack is the bomb.
hippich 13y ago
If you use GIT, checkout also 'git grep' command.
- davvid 13y ago
  Yes, more people need to know about "git grep". Did you know you can do things like `git grep -e foo --and -e bar --and --not -e baz`? "git grep" is great.
- pcl 13y ago
  It'd be cool to see these sorts of new search tools integrate into the git ecosystem. I wonder how much faster 'git ack' would be than just straight-up ack -- in other words, could ack benefit from any of git's working tree and delta-based optimizations / awareness?
  - keeperofdakeys 13y ago
    Besides only searching the files in the working tree, and more easily searching previous revisions, there won't be any speed improvements. Even if a file is stored using deltas, you still need to fully reconstruct it (more processing).
    - pcl 13y ago
      For full-file queries, sure. But I could imagine asking temporal questions like "show me changes made in the last month that introduces API x", for example.
eliben 13y ago
If you don't want/have Perl installed or are just a Python aficionado - check out https://github.com/eliben/pss/ https://github.com/eliben/pss/ - the Python-based ack clone. It's also actively developed.
zhufenggood 13y ago
if you use vim and vundle Bundle 'mileszs/ack.vim'
clumsysmurf 13y ago
Anyone know of a tool like ack/ag which also lets me search for a term in proximity to another?" For example, find <foo> within 5 lines of <bar>"
- andrelaszlo 13y ago
  I guess you could setup Sphinx to index your code :) "If it's worth doing, it's worth overdoing!" http://sphinxsearch.com/ http://sphinxsearch.com/
- ekns 13y ago
  You could do this with grep or ack or ag: grep bar -A 5 -B 5 files... | grep foo -A n prints n lines following the match (bar) and -B n prints n lines before the match.
  - bencoder 13y ago
    -C x does the same as both -A x and -B x together (A - after, B - before, C - context)

70 points by andrelaszlo 13y ago

A similar post popped up a couple weeks ago, and one comment suggested https://github.com/ggreer/the_silver_searcher https://github.com/ggreer/the_silver_searcher. I was using ack at the time, but I'm really liking ag as I've found to be quite a bit faster. The one thing I wish it had was being able to restrict searches by file type.

rwl4 13y ago
I second this. My friend kept suggesting I switch from Greg to ack. I searched for it and found ag and fell in love with it's speed. That same friend has now switched too.
bencoder 13y ago
Yep. I switched from grep to ack then to ag. ag is incredibly fast. You can limit the files it searches with -G <pattern>
- meowface 13y ago
  Thanks, I was looking for this as well. ag --help could probably be reworked a little bit; --include makes more sense than -G / --file-search-regex in my opinion.
- michaelmior 13y ago
  Thanks. Didn't read the manpage too carefully. Still not quite as easy as with ack, but probably good enough for me.
- mh- 13y ago
  for those who are frequently searching paths with a large number of files stored on an SSD or ramdisk/tmpfs, the bottleneck is very much the CPU time. in these cases ag is noticeably faster (orders of magnitude, in some cases), especially if you're just searching a literal and not with a regex pattern. the author has done some really cool performance hacks, and written some great blog posts[1] along the way. [1]: http://geoff.greer.fm/2012/09/03/profiling-ag-writing-my-own-scandir/ http://geoff.greer.fm/2012/09/03/profiling-ag-writing-my-own...
  - e12e 13y ago
    I was a bit surprised by a simple test I just ran on a Linux 3.10 tree: time rgrep aes . > /dev/null real 0m0.900s user 0m0.548s sys 0m0.340s # Somewhat similar output to ag: time rgrep -n --color aes . > /dev/null real 0m1.177s user 0m0.876s sys 0m0.288s time ag aes > /dev/null real 0m1.147s user 0m1.040s sys 0m0.548s # Using fixed strings in grep, limiting us to searching c-files time rgrep -n --color -F --include='*.c' aes . > /dev/null real 0m0.936s user 0m0.720s sys 0m0.208s time ag -G \.c aes > /dev/null real 0m1.130s user 0m1.140s sys 0m0.428s This is on an encrypted volume sitting on top of a low end SSD, all runs with hot chache. The times here are from ag in Debian -- I tried a build from upstream git -- but with essentially the same time. I guess ag might make sense under OS X -- but there doesn't appear to be any (speed) advantages under GNU/Linux.
    - mh- 13y ago
      nice, thanks for sharing. I was actually contrasting it to ack in my speed claims. I don't expect that it'd be that much faster than GNU grep in most situations.
soveran 13y ago
Not quite what you want, but you can restrict the search to filenames matching certain pattern with the -G option. For example, ag <pattern> -G \.c will only search on C files.
stottc 13y ago
You can do something like : ag -G ".txt$" hello
mtarnovan 13y ago
Indeed, I've also been using ack for a while, and switched to ag. Much faster.
meowface 13y ago
ag is indeed incredibly fast and extremely useful. It's replaced grep for me, any time I need to do a recursive search.
ams6110 13y ago
I'm pretty accustomed to using find | xargs grep. This lets me completely tailor the file I search using find's filters (e.g., only search files with a .py extension, and skip .svn/ directory). E.g. assuming I'm in my project's root directory: find . -name .svn -prune -o -name '*.py' -print | xargs grep -Hi 'string or RE I am looking for'
- pyre 13y ago
  Or you can do: ack --python PATTERN [DIRECTORY]
- michaelmior 13y ago
  I used to do the same, but this accomplishes the same thing and is much more concise ag -i -G '\.py$' 'string or RE I am looking for'
- troels 13y ago
  I'm pretty happy with `find|xargs grep` myself. Of course, I have an alias that hides the arcane incantation away under a single `frep`, but underneath it uses those tried and tested unix tools. An alias is usually simpler to carry with me to new systems as well.
  - ams6110 13y ago
    My alias is my .bash_history :). I grant that these other tools are great, but I like using the standard utilities because, among other things, once you learn 'find' and 'xargs' you can piece together all kinds of other commands and you don't have to remember a tool-specific syntax (at least for the 'find' and 'xargs' parts).
davidgerard 13y ago
Yep. Came here to recommend ag. It's just fantastic.

allannienhuis 13y ago

ack is the bomb.

hippich 13y ago

If you use GIT, checkout also 'git grep' command.

davvid 13y ago
Yes, more people need to know about "git grep". Did you know you can do things like `git grep -e foo --and -e bar --and --not -e baz`? "git grep" is great.
pcl 13y ago
It'd be cool to see these sorts of new search tools integrate into the git ecosystem. I wonder how much faster 'git ack' would be than just straight-up ack -- in other words, could ack benefit from any of git's working tree and delta-based optimizations / awareness?
- keeperofdakeys 13y ago
  Besides only searching the files in the working tree, and more easily searching previous revisions, there won't be any speed improvements. Even if a file is stored using deltas, you still need to fully reconstruct it (more processing).
  - pcl 13y ago
    For full-file queries, sure. But I could imagine asking temporal questions like "show me changes made in the last month that introduces API x", for example.

eliben 13y ago

If you don't want/have Perl installed or are just a Python aficionado - check out https://github.com/eliben/pss/ https://github.com/eliben/pss/ - the Python-based ack clone. It's also actively developed.

zhufenggood 13y ago

if you use vim and vundle Bundle 'mileszs/ack.vim'

clumsysmurf 13y ago

Anyone know of a tool like ack/ag which also lets me search for a term in proximity to another?" For example, find <foo> within 5 lines of <bar>"

andrelaszlo 13y ago
I guess you could setup Sphinx to index your code :) "If it's worth doing, it's worth overdoing!" http://sphinxsearch.com/ http://sphinxsearch.com/
ekns 13y ago
You could do this with grep or ack or ag: grep bar -A 5 -B 5 files... | grep foo -A n prints n lines following the match (bar) and -B n prints n lines before the match.
- bencoder 13y ago
  -C x does the same as both -A x and -B x together (A - after, B - before, C - context)