Difference

Difference#

The git diff command allows you to chech difference between condition of repository.

Untracked files#

Untracked files wont be displayed in git diff.

The following example just init git repository and add a file there and the git diff command doesn’t print anything in this case.

%%bash
mkdir diff_example
cd diff_example

git init &> /dev/null
echo "some text" > test_file
git diff

cd ..
rm -r diff_example

Changed files#

By default git diff shows unstaged changes. In fact, any change to the file can be described as an insert/delete. So the following subsections show what it’s like to have insert/delete lines in the git diff results.

Insert line#

Main point of the next cell is to add line "second line" to the file commited before. In git diff results it’ll look like +second line.

%%bash
mkdir diff_example
cd diff_example

git init &> /dev/null
echo "first line" > test_file
git add test_file &> /dev/null
git commit -m "initial commit" &> /dev/null


echo "second line" >> test_file
git diff

cd ..
rm -r diff_example

diff --git a/test_file b/test_file
index 08fe272..06fcdd7 100644
--- a/test_file
+++ b/test_file
@@ -1 +1,2 @@
 first line
+second line

Delete line#

In this case I created files with two lines "first line" and "second line". After committing, I save the new version of the file without the second line. Gid diff shows it as -second line.

%%bash
mkdir diff_example
cd diff_example
git init &> /dev/null

cat > test_file << EOF
first line
second line
EOF
git add test_file
git commit -m "initial commit" &> /dev/null

cat > test_file << EOF
first line
EOF
git diff

cd ..
rm -r diff_example

diff --git a/test_file b/test_file
index 06fcdd7..08fe272 100644
--- a/test_file
+++ b/test_file
@@ -1,2 +1 @@
 first line
-second line

New file#

In this section, I’ll show you what a newly added file looks like in git diff.

Empty file#

In the following example, I’ve added an empty second_file and compared it with the state of the repository in the previous commit, if there is no second_file' line, the new file mode` will signal that this file has been created.

%%bash
mkdir diff_example
cd diff_example
git init &> /dev/null

touch first_file
git add first_file
git commit -m "first commit" &> /dev/null

touch second_file
git add second_file &> /dev/null

git diff --staged
cat second_file

cd ..
rm -r diff_example

diff --git a/second_file b/second_file
new file mode 100644
index 0000000..e69de29

File with content#

In this example I have created a file second_file and added some content to it. It’ll share all the details of adding empty file option - new file mode 100644 signal that new file has been created. But there is also a section with changes to the file below. Note that the source file from the commit to which we are comparing the state is marked as --- /dev/null (in Linux this directory is used for information that should simply be deleted).

%%bash
mkdir diff_example
cd diff_example
git init &> /dev/null

touch first_file
git add first_file
git commit -m "first commit" &> /dev/null

echo "some content" > second_file
git add second_file &> /dev/null

git diff --staged
cat second_file

cd ..
rm -r diff_example

diff --git a/second_file b/second_file
new file mode 100644
index 0000000..2ef267e
--- /dev/null
+++ b/second_file
@@ -0,0 +1 @@
+some content
some content

–staged option#

By default git diff shows only changed but not staged files. To get information about staged files use --staged option.

New file#

By using the --staged option, you can even see the difference between a newly created (but added) file.

So in the following example I have created and added to the repo file and compared the results with and without the --staged option.

%%bash
mkdir diff_example
cd diff_example

git init &> /dev/null
echo "some text" > staged_file
git add staged_file

echo "=====just diff===="
git diff

echo
echo "=====diff staged====="
git diff --staged

cd ..
rm -r diff_example

=====just diff====

=====diff staged=====
diff --git a/staged_file b/staged_file
new file mode 100644
index 0000000..7b57bd2
--- /dev/null
+++ b/staged_file
@@ -0,0 +1 @@
+some text

Unstaged changes#

If you use the --staged option, you won’t see any difference when unstaging files. In the example:

In the first commit I just added first_file;
I added second_file and staged it;
I changed first_file but did not stage it;
Using git diff' without the –staged’ option I only got changes in first_file because it wasn’t staged;
Using git diff --staged I only got information about second_file but not about first_file.

%%bash
mkdir diff_example
cd diff_example
git init &> /dev/null

echo "first text" > first_file
git add first_file
git commit -m "first file add" &> /dev/null

echo "sefond text" > second_file
git add second_file
echo "changed first text" > first_file

echo "=====just diff===="
git diff

echo
echo "=====diff staged====="
git diff --staged

cd ..
rm -r diff_example

=====just diff====
diff --git a/first_file b/first_file
index 38181e5..fc9cf92 100644
--- a/first_file
+++ b/first_file
@@ -1 +1 @@
-first text
+changed first text

=====diff staged=====
diff --git a/second_file b/second_file
new file mode 100644
index 0000000..422845a
--- /dev/null
+++ b/second_file
@@ -0,0 +1 @@
+sefond text

Compare two commits#

You can compare two arbitrary commits using the syntax git diff <basic commit> <comparison commit>.

Note basic commit not necessary to be earlier that comparison commit.

So in the following example:

In cycle created few commits each add some line to file:
- Note line val "hash$i=$(git rev-parse HEAD)" just saves hash of last commit to variable hash<i>, so, for example, you can get hash of first commit from variable $hash1;
git log just help to understand current state of the repository;
Then I copare commits Line 1 added and Line 2 added in both options:
- Line 1 added as basic commit and Line 2 added as comparison commit - +Line 2show that it was added in comparison commit relatively basic commit;
- Line 2 added as basic commit and Line 1 added as comparison commit - -Line 2show that it was deleted in comparison commit relatively basic commit.

%%bash
mkdir diff_example
cd diff_example
git init &> /dev/null

for i in {1..3}
do
  echo "Line $i" >> file
  git add file
  git commit -m "Line $i added" &> /dev/null
  eval "hash$i=$(git rev-parse HEAD)"
done

echo "=====git log====="
git log --oneline --decorate

echo
echo "=====git diff \$hash1 \$hash2======="
git diff $hash1 $hash2
echo
echo "=====git diff \$hash2 \$hash1======="
git diff $hash2 $hash1

cd ..
rm -r diff_example

=====git log=====
c9578cb (HEAD -> master) Line 3 added
ff19e36 Line 2 added
bd6fd11 Line 1 added

=====git diff $hash1 $hash2=======
diff --git a/file b/file
index 3be9c81..c82de6a 100644
--- a/file
+++ b/file
@@ -1 +1,2 @@
 Line 1
+Line 2

=====git diff $hash2 $hash1=======
diff --git a/file b/file
index c82de6a..3be9c81 100644
--- a/file
+++ b/file
@@ -1,2 +1 @@
 Line 1
-Line 2

Difference

Contents

Difference#

Untracked files#

Changed files#

Insert line#

Delete line#

New file#

Empty file#

File with content#

–staged option#

New file#

Unstaged changes#

Compare two commits#