Remove the second instance of a character from a string

Question

I am working on making a git command I use more useful. A common task I do is to grep my git rev-list --all. I wrote (aka cut and pasted another StackExchange answer) git command to do this for me.

~/bin/git-search:

!/bin/bash
function _search() {
    git rev-list --all | (
        while read revision; do
            git grep -F $1 $revision
        done
    )
}
_search $1

The output from this looks like:

f26ce56cf6b17401292c494f906b2b6a9071ca75:filename.py:grepped string

I usually take these results and run git show along with the commit and file path to see that particular version of the file. git show takes the input of {COMMIT HASH}:path/to/file.

What I'd ideally like is to have my git function stick a whitespace where the second : is, which would allow me more easily copy and paste the output of git-search into git show, ie:

f26ce56cf6b17401292c494f906b2b6a9071ca75:filename.py grepped string

I'd like to use BASH for this as I am already using BASH. My initial solution was to use Python but that seems needless to me. I just am not sure how best to achieve this in BASH.

Answer 1

sed 's/:/ /2'

This would change the second : character to a space.

You could stick that in as an extra stage of your function's pipeline:

#!/bin/sh

git rev-list --all |
while read revision; do
    git grep -F "$1" "$revision"
done |
sed 's/:/ /2'

(I actually deleted the function as it didn't seem to be needed; note also the quoting of the variable expansions; oh, and it's a /bin/sh script since it's not using any bash-specific features (neither did yours, except for the unneeded function keyword))