Respond to Prompt in Bash Script
I work on a project that requires I frequently build and destroy a virtual machine. I don't enjoy having to do so but virtual machines can be notoriously difficult to prop up, especially when you have a complex app living within it.
Manually typing the same commands over and over can be mind-numbing so I've create a bash script to handle all of the work for me. One hiccup to the process is needing to confirm removal of an existing virtual machine; using a pipe and echo allows me to answer the prompt:
# ... some directives here
# Remove the machine, confirming "y" when asked by docker-machine
echo 'y' | docker-machine rm default
# ... more directives here
Using echo I pipe a y response to docker-machine's confirmation prompt, thus allowing the script to move forward with other tasks without the need for manual intervention.
I don't, however, know how to handle multiple prompts -- can you tell me and everyone else?
![Serving Fonts from CDN]()
For maximum performance, we all know we must put our assets on CDN (another domain). Along with those assets are custom web fonts. Unfortunately custom web fonts via CDN (or any cross-domain font request) don't work in Firefox or Internet Explorer (correctly so, by spec) though...
![Create Namespaced Classes with MooTools]()
MooTools has always gotten a bit of grief for not inherently using and standardizing namespaced-based JavaScript classes like the Dojo Toolkit does. Many developers create their classes as globals which is generally frowned up. I mostly disagree with that stance, but each to their own. In any event...
![CSS pointer-events]()
The responsibilities taken on by CSS seems to be increasingly blurring with JavaScript. Consider the -webkit-touch-callout CSS property, which prevents iOS's link dialog menu when you tap and hold a clickable element. The pointer-events property is even more JavaScript-like, preventing:
click actions from doing...
![dat.gui: Exceptional JavaScript Interface Controller]()
We all love trusted JavaScript frameworks like MooTools, jQuery, and Dojo, but there's a big push toward using focused micro-frameworks for smaller purposes. Of course, there are positives and negatives to using them. Positives include smaller JS footprint (especially good for mobile) and less cruft, negatives...
As far as I know this is usually done with the yes command:
This should also respond to multiple prompts.
Ah yes, I did experiment with that. But what if I know the multiple prompt responses should be “y”, “n”, “3”? That would be useful.
To expand on what Andreas said, you can build a canned file of responses, one per line (extra blanks for the equivalent of hitting enter) and sending that to your program:
Can’t you stream a file in? Like in
Try a here string.
while IFS= read -r; do echo "E- $REPLY" done <<<"y n 3"Will print the three lines of responses. Assuming the docker prompt is similar to read it should work.
The ‘expect’ utility can be installed on most Linux boxes.
https://en.wikipedia.org/wiki/Expect
I agree about ‘expect’ as Rohit mentions, it’s usually what you resort to when piping doesn’t work for different reasons (such as passwords being read from physical terminal only, input being prematurely consumed by a loop, etc).
The tool is {advanced|complicated} enough to maybe warrant a future post ;-)
In my team we use Vagrant and Ansible to do these kinds of things. They are very helpful tool for popping up a dev env with customized apps / configs.
I have not tried Ansible for Docker but there is a page for it https://www.ansible.com/docker