CODEWITHSUNDEEP
gitshellautomation
Saurabh Kumar
Saurabh Kumar

Reputation: 5945

A better way to do git clone

This is lazy programmer request. I would like to create a shell script automate the following process: 

git clone <remote-repo-url>
cd <cloned-folder>
open <cloned-folder>

So the idea here is to clone a URL and then immediately cd into the cloned-folder. The trick here is to identify the cloned-folder from the url pattern.

For now we can assume that url structure is in this pattern .../<cloned-folder>.git i.e. the url.

I am wondering if we can actually do with using awk or some tools like that. The part i am stuck is finding a appropriate regex, I guess.

USE CASE: Here the use case is if you clone a url, you want to be in the repofolder as soon as possible. The is the pre-requirement if you want to run any git command like git log or mate . which we do 99% of the time.

Thanks in advance.

Upvotes: 8

Views: 13756

Answers (5)

Hongbo Liu
Hongbo Liu

Reputation: 3064

Instead of getting directory name from url, we could just cd into the latest created directory. Here is the corresponding shell script:

gclcd() {
    git clone --recursive $* && cd "$(ls -t | head -1)"
}

Upvotes: 1

Louis Maddox
Louis Maddox

Reputation: 5566

The bash parameter version is quite nice, I went with basename cuz it seemed cleaner

function gclocd {
  # git clone and enter the cloned file
  if [[ -z "$2" ]]; then
    reponame=$(basename "$1" ".git");
  else
    reponame=$2;
  fi
    git clone "$1" "$reponame";
  cd "$reponame";
}

I use short memorable tab-completable function names for a bunch of things like this and put them in my .bashrc as aliases/functions, e.g. gcmmt for git commit -m. lazy may well already be a prefix for things in your shell, which means it's that little bit more to type each time...

Upvotes: 0

罗泽轩
罗泽轩

Reputation: 1673

An improvement from @dbr's answer:

function lazyclone {
    reponame=${1##*/}
    reponame=${reponame%.git}
    git clone "$1" "$reponame";
    cd "$reponame";
}

With the help of Bash's parameter expansion, we can get rid of awk and sed.

Upvotes: 1

dbr
dbr

Reputation: 169573

bash function to do this (works in zsh also):

function lazyclone {
    url=$1;
    reponame=$(echo $url | awk -F/ '{print $NF}' | sed -e 's/.git$//');
    git clone $url $reponame;
    cd $reponame;
}

The awk command prints the part after the last / (e.g from http://example.com/myrepo.git to myrepo.git). The sed command removes the trailing .git

Usage:

$ pwd
~/
$ lazyclone https://github.com/dbr/tvdb_api.git
tvdb_api
Cloning into 'tvdb_api'...
remote: Counting objects: 1477, done.
remote: Compressing objects: 100% (534/534), done.
remote: Total 1477 (delta 952), reused 1462 (delta 940)
Receiving objects: 100% (1477/1477), 268.48 KiB | 202 KiB/s, done.
Resolving deltas: 100% (952/952), done.
$ pwd
~/tvdb_api

Upvotes: 15

jordanm
jordanm

Reputation: 34934

With git clone, you can specify the folder to clone to, instead of allowing it to be named automatically.

dir=myclone
git clone git://somerepo "$dir"
cd "$dir"
open "$dir"

Upvotes: 2

Related Questions