I'm using git diff on tags in a Bash window and its doing what I want for the most part.
$ git diff -- numstat TAG2 TAG1
Giving this output:
4 6 dir1/dir2/dir3/dir4/thisIsMyFile.cs
How can I get this output:
4 6 thisIsMyFile.cs
I don't need/want the path (dir1/dir2/dir3/dir4/), just the file name (thisIsMyFile.cs).
As far as I know, there's no way to do this with Git alone (i.e., just giving Git diff the right arguments). You have to use a Bash pipeline to post-process the Git output. I suggest:
git diff --numstat TAG2 TAG1 | while IFS= read -r line; do
printf '%s\t' "$(cut -f-2 <<<"$line")"
basename "$(cut -f2- <<<"$line")"
done
This is a little more complicated because it is designed to preserve the original formatting and to parse correctly even if there are files with tabs or spaces in them. It could be simplified if you can guarantee no file has whitespace in its file name, or if you are interested in using a more powerful scripting language than Bash to parse the lines.
You can use awk:
git diff --numstat | awk -F'[/\t]' '{print $1,$2,$NF}'
What is the awk command doing?
-F'[/\t]' sets the field delimiter to either tab or /. (Output from git is tab delimited). $NF is the last field in the line - the filename.
Note: The above command will fail when the filename contains a tab. If you can live with that restriction, have fun!
use basename -a $(git diff --numstat TAG1 TAG2)
NOTE I'm using bash syntax to execute a sub-command.
