Bash / Git-如何使用当前的git分支名称制作脚本

Question

New to bash scripting. Know enough to run something simple from intro videos via LinuxAcademy.

I'd like to pull my current Git branch name into the script as a variable and use that variable to make a directory.

From there, I think I can automate the rest of the process.

No script yet, but I do know what I'd like it to do; here it is in plain language.

1. cd somedir/dir
2. git branch (pull name as a variable)
3. mkdir (as branch name variable)
4. cd (newdir)
5. touch xyz.html
6. mkdir css img
7. cd css touch xyz.css reset.css
8. cd ../..

Answer 1

There are two different good ways to do this, and a number of bad ways.

The most common bad way that you'll find is a variation on this theme:

 branch=$(git branch | grep '\\*' | cut -d' ' -f2) # don't do this! 

Replace this with:

branch=$(git rev-parse --abbrev-ref HEAD)

Note that if HEAD is detached (a normal enough state that means not on any branch at all , typically found, eg, when you're in the middle of a rebase that you have not yet finished), this just prints HEAD . Using this as the branch name often works and does the right thing, so that's OK as long as, well, that's OK. (The cut version of this prints (HEAD , which doesn't work.)

The other way to do it is:

branch=$(git symbolic-ref --short HEAD) || exit

(which you can make fancier if you like). With this variant, if you're not on a branch—if you're in that detached HEAD mode—the git symbolic-ref command will produce a complaint:

fatal: ref HEAD is not a symbolic ref

and the || exit || exit clause will make your script terminate immediately, rather than proceeding to use $branch as if it were a valid branch name when it's not. Use this code fragment if using the literal string HEAD instead of an actual branch name is the wrong thing to do.

Answer 2

In bash context, you would use a simple git command

git rev-parse --abbrev-ref HEAD

to output current branch in its short form. ( doc )