I'd read that when renaming files in Git, you should commit any changes, perform your rename and then stage your renamed file. Git will recognise the file from the contents, rather than seeing it as a new untracked file, and keep the change history.
However, doing just this tonight I ended up reverting to git mv.
> $ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
#
I renamed my stylesheet in Finder from iphone.css to mobile.css:
> $ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
#
# Changed but not updated:
# (use "git add/rm <file>..." to update what will be committed)
# (use "git checkout -- <file>..." to discard changes in working directory)
#
# deleted: css/iphone.css
#
# Untracked files:
# (use "git add <file>..." to include in what will be committed)
#
# css/mobile.css
So Git now thinks I've deleted one CSS file, and added a new one. It is not what I want. Let’s undo the rename and let Git do the work.
> $ git reset HEAD .
Unstaged changes after reset:
M css/iphone.css
M index.html
I am back to where I began:
> $ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
#
Let's use git mv instead:
> $ git mv css/iphone.css css/mobile.css
> $ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# renamed: css/iphone.css -> css/mobile.css
#
# Changed but not updated:
# (use "git add <file>..." to update what will be committed)
# (use "git checkout -- <file>..." to discard changes in working directory)
#
# modified: index.html
#
It looks like we're good. So why didn't Git recognise the rename the first time around when I used Finder?
For git mv the manual page says
The index is updated after successful completion, […]
So, at first, you have to update the index on your own
(by using git add mobile.css). However git status
will still show two different files:
$ git status
# On branch master
warning: LF will be replaced by CRLF in index.html
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
# new file: mobile.css
#
# Changed but not updated:
# (use "git add/rm <file>..." to update what will be committed)
# (use "git checkout -- <file>..." to discard changes in working directory)
#
# deleted: iphone.css
#
You can get a different output by running git commit --dry-run -a, which results in what you
expect:
Tanascius@H181 /d/temp/blo (master)
$ git commit --dry-run -a
# On branch master
warning: LF will be replaced by CRLF in index.html
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
# renamed: iphone.css -> mobile.css
#
I can't tell you exactly why we see these differences
between git status and git commit --dry-run -a, but
here is a hint from Linus:
git really doesn't even care about the whole "rename detection" internally, and any commits you have done with renames are totally independent of the heuristics we then use to show the renames.
A dry-run uses the real renaming mechanisms, while a
git status probably doesn't.