Figure out changed lines in a commit in a git repo

git show can be used to figure out the changed lines in a commit. I recently worked on a task at work to remove all unused texts. I used a script and it was not feasible for QA to test and make sure that I had not removed some text that is in use by going through the entire product.

QA reached out to me asking the list of strings removed, and I used git show to get them.

The command was:

git show <commit-hash> --unified=0 -- <file-path>

Lets break it down.

1. git show -> Is used to get the changes in the commit with the
2. --unified=0 -> Is only gives the list of changed lines without any surrounding context lines.
3. -- -> Tells that the next values are not a git option.
4. -> Is the absolute or relative path of the file in git repo.

I was only interested in getting the removed lines so the output of this command was filtered through grep to only get removed lines(-)

So now the command was:

git show --unified=0 <commit-hash> -- <file-path> | grep '^-'

• grep '^-' -> Matches all lines starting(^) with a -

To make the output cleaner, I further filtered it with cut -c2-(remove leading -). I also removed empty lines with grep -v '^[[:space:]]*$'. I then redirected the output to file, so it can be shared with QA.

So the final command is:

git show --unified=0 <commit-hash> -- <file-path> | grep '^-' | cut -c2- | grep -v '^[[:space:]]*$' > <output-file-name>

• cut -c2- -> Extract characters starting from 2nd position(- is in 1st position, it is 1 based indexing)
• grep -v '^[[:space:]]*$' -> Remove all lines with only whitespaces, -v does inverted matching. ^[[:space:]]*$ matches lines with 0 or more whitespaces.

References: man git-show, man grep, man cut