Why git stores copy of whole file for single line change?

Question

Say we go for a sample run:

git init
echo -e 'hede\nhodo' > new.txt
git add new.txt
git commit -m 'commit 1'
 1 file changed, 2 insertions(+)
echo -e 'hede\n123' > new.txt
git add new.txt
git commit -m 'commit 2'
 1 file changed, 1 insertion(+), 1 deletion(-)

So I know SHA of two versions. When I look into them, I see:

git cat-file -p ad0daa422be9531700a069e6800b18c065f755bb
hede
hodo

git cat-file -p 6fbf5b80b76bbe270bb41789c34851d622ae1f63
hede
123

It seems git copied blob, but 'hede' remained redundant. I know it is not a big impact for small files but just asking: Does git optimize this into differences later on? Or is this the way what makes git efficient?

Thanks

Check this answer http://stackoverflow.com/questions/8198105/how-does-git-store-files and note the comment by Jan Hudec — Michael Ver, Mar 27 '14 at 07:01

score 4 · Accepted Answer · edited May 23 '17 at 12:15

4

As I mention in "How does git store files?", Git:

manages content (which is why you see 2 complete blobs)
stores internally delta (and achieve an high-compression level that way, though pack files)

So it is efficient when it matters (clone/push/pull), while providing easy and quick access locally.

edited May 23 '17 at 12:15

Community

1
1

answered Mar 27 '14 at 07:09

VonC

1,262,500
529
4,410
5,250

Thank you that was the information I was looking for. Jan Hudec's comment is clear as well. – hurturk Mar 27 '14 at 16:02

Why git stores copy of whole file for single line change?

1 Answers1