Reputation: 27119
I'd like to change the pwd of the current shell from within a ruby script. So:
> pwd
/tmp
> ruby cdscript.rb
> pwd
/usr/bin
This is the code I have right now:
exec('cd /usr/bin')
Unfortunately, cd is a builtin command. So:
`exec': No such file or directory - cd (Errno:ENOENT)
Any workaround for this?
No way to get it working in ruby itself, so I switched gears. I modified the script to output the target directory path and then defined a function in .bashrc that would pass the arguments through the script and then cd to the right directory. Not self-contained as I would have hoped, but it did the job.
Thanks for the replies, folks.
Upvotes: 5
Views: 10768
Reputation: 3567
As other answers have pointed out, running either system("cd foo")
or `cd foo`
within a Ruby script will only change the directory of the Ruby subprocess. As soon as the Ruby script finishes, you'll get popped back to the directory you were in before.
A way around this is to have the Ruby script only be responsible for printing out a directory name and then calling that script with a wrapping shell function that cd
s into the directory the Ruby script prints out.
If you'd like your Ruby script to interact with the user use STDERR.puts
instead of puts
.
Here's an example Ruby script cd.rb
that will try to find a matching directory from user input:
#!/usr/bin/env ruby
pattern = ARGV.join(" ")
dirs = Dir.glob("*").filter{|path| File.directory?(path)}.filter{|path| path.include?(pattern)}
if dirs.length == 0
STDERR.puts "Error: no matching directories found."
print "."
exit 1
elsif dirs.length == 1
print dirs.first
else
while true do
STDERR.puts "Which one?"
dirs.each_with_index{|dir, i|
STDERR.puts "#{i+1}. #{dir}"
}
STDERR.print "=> "
i = STDIN.gets.chomp.to_i - 1
if dirs[i]
print dirs[i]
exit
else
STDERR.puts "Error: uhh what? Try that again...\n\n"
end
end
end
And here's the wrapping sh
or zsh
script to stick in your .bashrc
or .zshrc
that will pass along arguments to cd.rb
and then cd
into the directory that Ruby script prints out:
function ,cd () {
cd "`cd.rb $@`"
}
Upvotes: 0
Reputation: 389
You can try to use system
instead of exec
.
It works for me.
Like system("cd #{new_path} && <statement> && cd #{original_path}")
Upvotes: 0
Reputation: 1278
Stumbled across this while searching to do the same thing.
I was able to solve this by running multiple statements within the backticks.
'cd #{path} && <statement> && cd ..'
Upvotes: 6
Reputation: 10155
As other answers already pointed out, you can change the pwd inside your ruby script, but it only affects the child process (your ruby script). A parent process' pwd cannot be changed from a child process.
One workaround could be, to have the script output the shell command(s) to execute and call it in backticks (i.e. the shell executes the script and takes its output as commands to execute)
myscript.rb:
# … fancy code to build somepath …
puts "cd #{somepath}"
to call it:
`ruby myscript.rb`
make it a simple command by using an alias:
alias myscript='`ruby /path/to/myscript.rb`'
Unfortunately, this way you can't have your script output text to the user since any text the script outputs is executed by the shell (though there are more workarounds to this as well).
Upvotes: 4
Reputation: 370425
You can't change the working directory (or the environment variables for that matter) of the shell that invoked you in a ruby script (or any other kind of application). The only commands that can change the pwd of the current shell are shell built-ins.
Upvotes: 0
Reputation: 22220
You won't be able to change the working directory of your shell.
When it executes ruby, it forks so you get a new environment, with a new working directory.
Upvotes: 8
Reputation: 340476
Dir.chdir("/usr/bin")
will change the pwd within the ruby script, but that won't change the shell's PWD as the script is executed in a child process.
Upvotes: 14