-B option of git log
Hello,
As far as I understand the documentation, -B of git-log should help correct rename detection. But it does not seem to work for me.
Let me get a setup:
$> git init .
Initialized empty Git repository in /tmp/t2/.git/
$> echo 'Lorem ipsum doler sed. Lorem ipsum doler sed. Lorem ipsum doler sed. Lorem ipsum doler sed.' > a
$> git add a
$> git commit -m 'Init.'
[master (root-commit) b78205c] Init.
1 file changed, 1 insertion(+) create mode 100644 a
$> mv a b
$> echo 'new' > a
$> git add -A .
$> git commit -m '2nd.'
[master a30ca49] 2nd.
2 files changed, 2 insertions(+), 1 deletion(-) create mode 100644 b
Now, I get the following:
$> git log --oneline -B20%/80% -M20% --name-status a30ca49 2nd.
M a
A b
b78205c Init.
A a
But I would expect that git-log shows me a rename from a to b and a new a.
What is my misunderstanding? (And I tried it also with files with a couple of lines.)
SEcki
Quoted 21 lines · Eckhard Maass
Let me get a setup:
$> git init .
Initialized empty Git repository in/tmp/t2/.git/
$> echo 'Lorem ipsum doler sed. Lorem ipsum doler sed. Lorem ipsum doler sed. Lorem ipsum doler sed.' > a
$> git add a
$> git commit -m 'Init.'
[master (root-commit) b78205c] Init.1 file changed, 1 insertion(+) create mode 100644 a$> mv a b
$> echo 'new' > a
$> git add -A .
$> git commit -m '2nd.'
[master a30ca49] 2nd.2 files changed, 2 insertions(+), 1 deletion(-) create mode 100644 bNow, I get the following:
Quoted 8 lines
$> git log --oneline -B20%/80% -M20% --name-status a30ca49 2nd.
M a
A b
b78205c Init.
A aBut I would expect that git-log shows me a rename from a to b and a new a.
I think the problem is that your test file is so tiny that it falls afoul of git's MINIMUM_BREAK_SIZE heuristic of 400 bytes (which prevents false positives on tiny files). Try replacing your "Lorem ipsum" echo with something like "seq 1 150", and I think you will find the result you are expecting.
-Peff
Quoted 5 lines · Jeff King
I think the problem is that your test file is so tiny that it falls afoul of git's
MINIMUM_BREAK_SIZEheuristic of 400 bytes (which prevents false positives on tiny files). Try replacing your "Lorem ipsum" echo with something like "seq 1 150", and I think you will find the result you are expecting.
Thanks. Two points, though:
With bigger files, I do get something like:
| M100 d | R100 d e
The rename is fine. But I am a bit puzzled mit the M - I, somehow, would have expected an A for add and not a M.
Secondly, would it be a good idea to enhance the documentation on that point to clarify this minimum size?
SEcki