我怎样才能在 git 中获取文件的每次提交的代码行

Question

i tried using the below statement in git:

$ wc -l file name 

Above statement gives me only for the latest version. But i need for each and every commit, so how can i get it?

Answer 1

The command you posted has nothing in common with git . wc is a general purpose Unix tool that counts the bytes, characters, words or lines in a file (its name stands for word count ). It cannot be used to count the number of lines of code, is it is language agnostic and cannot tell a line of code from a comment.

However, if you need the total number of lines in the file you can use the command:

git log --numstat --format=oneline -- file.txt | grep file.txt$

to get a history of changes on the specified file.

The output contains three columns separated by a variable number of spaces. The first column contains the number of lines added on each commit, the second column contains the number of lines removed on that commit, the third column is always the path of the file inside the project.

The commits are displayed in reverse order, as usual; the first line is the most recent commit that modified the file, the last line is the commit when the file was created (and should contain 0 on the second column).

It's easy to write a simple script (in your preferred scripting language) to parse the output and generate the number of lines of the file on each commit.

Update

The above command line doesn't preserve the commit hashes. Running only the git log part of it generates two lines for each commit. The first line contains the commit hash and the first line of the commit message, separated by a space. The second line is the one explained above. Still an easy job for a small parsing script.

In bash , the parsing script could look like this:

#!/bin/bash

if [ ! -f $1 ]; then echo Cannot find the file \"$1\".; exit 1; fi

LINES=$(wc -l $1 | sed 's/^ *\([0-9]*\) .*$/\1/')
git log --numstat --format=oneline -- $1 | while read -r HASH COMMENT; do
  read -r INSERTED DELETED FILENAME
  echo $HASH $LINES
  LINES=$(($LINES-$INSERTED+$DELETED))
done

It accepts the file path as its first argument and produces a list of lines; each line contains the commit hash and the number of lines the input file had after that commit.

Answer 2

You can use git log to get all of the commit hashes for a given $filename and check these out using a for -loop in your shell. For example (in Zsh):

git stash
for sha in $(git log --format=format:%H -- $filename); do
    git checkout $sha -- $filename
    echo "$(wc -l $filename) @ $sha" 
done
git stash pop