Feature Suggestion: Limit length of output line

[jeffg-archive]

I search repos that contain large, machine-generated JSON files (single line 500k file, for example). If these files contain my search pattern, the output spams my terminal with the full file contents.

I want to do something equivalent to this, but natively in ack (preserving ack's formatting):

$ ack <pattern> | cut -b 1-200
path/to/file1:124: blah blah pattern
path/to/file2:7: pattern blah blah blah blah blah ... <truncated>

For example, this imaginary option:

$ ack --max-line-length 200 <pattern>
path/to/file1
124: blah blah pattern
path/to/file2
7: pattern blah blah blah blah blah ... <truncated>

For reference, see this stackexchange question and answers: https://unix.stackexchange.com/questions/163726/limit-grep-context-to-n-characters-on-line

[petdance]

I think we might already have a ticket on this suggestion. It's a common one.

[hftf]

For reference, see also #234 and beyondgrep/ack2#596 and beyondgrep/ack2#345. More on StackOverflow/StackExchange:

1. How to truncate long matching lines returned by grep or ack (2010)
2. grep/ack -o with characters of context (not lines) (2011)

[philipp-classen]

If compatibility with ag is a concern, the feature is called --width there. ag --width NUM will cut after NUM characters and replace the rest with [...].

[petdance]

We would also have to allow a --nowidth or --width=0 so that you can override from the command line what I expect many people will do which is put a --width=100 or whatever in their .ackrc.

[n1vux]

Unless matches in the auto-generated nonwrapped JSON are interesting, the usual recommendation is to use affirmation or negative filetypes --nojson or --cpp .

Where --width=80 would really help is when there are a mix of uninteresting long lines, interesting long lines, and interesting short lines in the same project.

[stephenostermiller]

I too would like some sort of feature to deal with suppressing or truncating matches from long lines. My workaround is to filter the output of ack using grep to remove results that are longer than 300 characters.

ack my-seach-string | grep -vE '.{300,}'

but because using ack with a pipe turns off color by default, I usually turn that back on with a flag:

ack --color my-seach-string | grep -vE '.{300,}'

It would be nice to be able to put something in my .ackrc to ignore or truncate long lines by default that I could then override on the command line if I needed to.

[n1vux]

For explicit reading here in situ -
the 2. Stack-overflow reference in hftf's comment above has a usable work-around for characters of context for a long line:

ack -o '.{0,20}string.{0,20}'
will output up to 20 chars before and after the match.

Folks probably want highlighting of just the match of string within that context, so there's still work to do here.

a concrete proposal / straw dog would be ack --CC=20 string is effectively syntactic sugar for -o '.{0,20}string.{0,20}' but with highlighting of just string .

[epa]

Personally I see this as similar to the treatment of binary files by grep(1). Rather than print unreadable stuff to the terminal it says 'binary file matches'. I would like to see ack say 'foo:55:very long line matches' or even 'foo: single-line file matches'. I'm not so worried about the exact details of how to print the match with context, or highlighting, or what the maximum length should be.

(Most of the time these single-line files are generated code anyway and could probably be skipped, but you can't do that on file type, as JSON or XML files can equally well be human-written code or generated cruft. They are still just about readable in an editor, but not really printable to a terminal. Having ack print only a summary of whether a match is found would be good enough.)

[n1vux]

An additional workaround when dealing with JS/JSON, XML/HTML, wrapped TXT, and other super-long "lines" --

Ack option PAGER=less allows specifying options on the pager program.
E.g., -R is recommended so that your Ack highlighting works with the pager instead of displaying as ^[3m;.

I'm now using PAGER='less -iRSF' when such strangeness is expected.

-i or --ignore-case
-R or --RAW-CONTROL-CHARS
-S or --chop-long-lines
-F or --quit-if-one-screen

-S will show the first screen-width of the lines (lines matched or if -ABC, context lines too) , with a > at margin showing there's scroll-right text. ⬅️ ➡️ scroll your window around the buffer.

(If you're using a different pager than less, maybe yours has a similar feature; but less should be available for all modern desktop OS today, if not already in the default set of commands. For Win 10/11, this may require WSL or Cygwin or other GNU overlay.)

[stephenostermiller]

Instead of setting this as a env variable for all programs, you can configure ack to use less with these options with its config file. Add two lines to ~/.ackrc:

--pager
less -iRSF