Handling file renames in git
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
#
Rename 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. Not what I want, lets undo the rename and let git do the work.
> $ git reset HEAD .
Unstaged changes after reset:
M css/iphone.css
M index.html
Back to where I began.
> $ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# modified: index.html
#
Lets 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
#
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.
You have to add the two modified files to the index before git will recognize it as a move.
The only difference between mv old new and git mv old new is that the git mv also adds the files to the index.
mv old new then git add -A would have worked, too.
Note that you can't just use git add . because that doesn't add removals to the index.
See Difference between "git add -A" and "git add ."
Best thing is to try it for yourself.
mkdir test
cd test
git init
touch aaa.txt
git add .
git commit -a -m "New file"
mv aaa.txt bbb.txt
git add .
git status
git commit --dry-run -a
Now git status and git commit --dry-run -a shows two different results where git status shows bbb.txt as a new file/ aaa.txt is deleted, and the --dry-run commands shows the actual rename.
~/test$ git status
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# new file: bbb.txt
#
# Changes not staged for commit:
# (use "git add/rm <file>..." to update what will be committed)
# (use "git checkout -- <file>..." to discard changes in working directory)
#
# deleted: aaa.txt
#
/test$ git commit --dry-run -a
# On branch master
# Changes to be committed:
# (use "git reset HEAD <file>..." to unstage)
#
# renamed: aaa.txt -> bbb.txt
#
Now go ahead and do the check-in.
git commit -a -m "Rename"
Now you can see that the file is in fact renamed, and what's shown in git status is wrong.
Moral of the story: If you're not sure whether your file got renamed, issue a "git commit --dry-run -a". If its showing that the file is renamed, you're good to go.
链接地址: http://www.djcxy.com/p/7896.html上一篇: 在新存储库上推送原点主控错误
下一篇: 在git中处理文件重命名