Setting up a development environment

Question

I'm looking for feedback on improving the structure of this code. Also, I could not find the command line executable for Chrome.

General feedback is also requested.

#
#
#
# Divider - configures bash, git, grunt, sublime and chrome
#
#
#
user1=foo
config_bash() {
 rm ~/.bash_profile
 ln -s ~/root/config/bash/login.sh ~/.bash_profile
 source ~/.bash_profile
 echo "Bash configured."
}
config_git() {
 local a="1ドル"
 if [ $# -eq 0 ]
 then
 a="default"
 fi
 if [ $# -eq 1 ]
 then
 git config --global user.name "$a"
 git config --global user.email "$a@$a.com" 
 fi
 git config --global push.default matching
 git config --global core.editor "subl"
 git remote add godaddy [email protected]:~/root.git
 git remote add heroku https://git.heroku.com/frozen-dusk-2587.git
 echo "Git configured." 
}
config_grunt() {
 sudo npm install -g grunt-cli
 mkdir ~/root_install/grunt
 ln -s ~/root/config/grunt/package.json ~/root_install/grunt/package.json
 ln -s ~/root/config/grunt/Gruntfile.js ~/root_install/grunt/Gruntfile.js
 cd ~/root_install/grunt
 npm install
 cd ~/root
 echo "Grunt configured."
} 
config_sublime_2() {
 ln -s /Applications/Sublime\ Text\ 2.app/Contents/SharedSupport/bin/subl /usr/local/bin/subl
 rm -rf ~/Library/Application\ Support/Sublime\ Text\ 2/Packages/User
 ln -s ~/root/config/sublime ~/Library/Application\ Support/Sublime\ Text\ 2/Packages/User
 echo "Sublime configured."
}
reset_sublime_2() {
 rm -rf ~/Library/Application\ Support/Sublime\ Text\ 2
}
config_chrome(){
 chmod 0444 ~/Library/Application\ Support/Google/Chrome/Default/History
 echo "Chrome configured."
}
config_all() {
 config_bash
 config_git
 config_grunt
 config_sublime_2
 config_chrome
 list
}
list(){
 local a=$(which bash) b=$(which git) c=$(which grunt) d=$(which subl)
 local g=$(which node) i=$(which heroku)
 echo "****" 
 echo "Your bash executble is here: $a." 
 echo "Your git executble is here: $b." 
 echo "Your grunt executble is here: $c."
 echo "Your sublime executble is here: $d."
 echo "Your node executble is here: $g."
 echo "Your heroku executble is here: $i."
 echo "****"
}

• janos: I rolled back to the version before your refactoring. It is our policy to not edit the code after answers were received, as they may be invalidated by the changes. You may post your updated code as a new question to get further reviews.
• janos: See also: What you may and may not do after receiving answers
• Quill: @janos' comment still applies, please don't update your question with updated code.
• cade galt 0: Her is the update per code 1.2.1 section 8 subsection alpha - codereview.stackexchange.com/questions/122605/…

Answer 1

Simplify

Since $# cannot be both 0 and 1 at the same time, it would be better to join the two if conditions here:

local a="1ドル"
if [ $# -eq 0 ]
then
 a="default"
fi
if [ $# -eq 1 ]
then
 git config --global user.name "$a"
 git config --global user.email "$a@$a.com" 
fi

But, actually, what if there are 2 or more parameters instead of 0 and 1? It can happen by a programming mistake. It would be more robust to rewrite this by checking the values of the parameters, rather than their total count:

if [ "1ドル" ]
then
 git config --global user.name "1ドル"
 git config --global user.email "1ドル@1ドル.com" 
fi

Notice that I eliminated the a local variable, as it was pointless (a="default" set but never actually used).

Avoid changing directories in scripts

This is a bit fragile:

cd ~/root_install/grunt
npm install
cd ~/root

A cd command may fail, for example if the directory was not yet created. In general it's not a good idea to change directories inside a script, as the script might perform actions in unexpected directories when things go wrong.

A good solution in such situations is to wrap commands within (...), like this:

(cd ~/root_install/grunt; npm install)

Directory changes executed inside (...) will not affect the environment outside.

You can learn more about this syntax in the Compound Commands section in man bash:

 (list) list is executed in a subshell environment (see COMMAND EXECU-
 TION ENVIRONMENT below). Variable assignments and builtin com-
 mands that affect the shell's environment do not remain in
 effect after the command completes. The return status is the
 exit status of list.

Robustness

Note that if you re-run the script a 2nd time, some commands will fail, for example these:

mkdir ~/root_install/grunt
ln -s ~/root/config/grunt/package.json ~/root_install/grunt/package.json
ln -s ~/root/config/grunt/Gruntfile.js ~/root_install/grunt/Gruntfile.js

If the directory and the symlinks already exist, the script will print error messages for these lines. You can avoid that by adding some flags:

mkdir -p ~/root_install/grunt
ln -sf ~/root/config/grunt/package.json ~/root_install/grunt/package.json
ln -sf ~/root/config/grunt/Gruntfile.js ~/root_install/grunt/Gruntfile.js

Naming

These variable names are not very helpful:

local a=$(which bash) b=$(which git) c=$(which grunt) d=$(which subl)
local g=$(which node) i=$(which heroku)

Better name these after the commands they represent:

local bash=$(which bash) git=$(which git) # and so on ...

• janos: As I mentioned above, after you assign a="default", the variable is never used again. It's a pointless assignment
• janos: First a is assigned to 1ドル. If there are no arguments, then a is assigned to default, but then a is not used again. If there is one argument, then a is used by git config commands. Like I said, the value default never gets used.
• cade galt 0: that was a logic error I see thanks for pointing it out, but for correct working code I need 3 cases ( 0, 1, and more than 1 ) and a variable.
• Phrancis: @cadegalt0 nobody on this site gets paid to participate.
• janos: Votes are notoriously difficult to predict. Just posting better code doesn't guarantee more upvotes. If you want to get more upvotes, read the how-to-ask page in the help center very carefully, and also look at other successful posts in the relevant tags. In general, what seems to gather more upvotes is serious hard work on the quality of a post.