As evidenced by its output, etc-update does not check for valid input before trying to process it. Reproducible: Always Steps to Reproduce: 1. Run etc-update. 2. At a prompt, use "1y" as input. Actual Results: /usr/sbin/etc-update: line 215: read: 1y: value too great for base (error token is "1y") Expected Results: The software should recover from this error and say that "1y" was not an option.
It's better than it used to be (bug 143881). I suppose we can stop trying to read it into an integer type in order to avoid that error all together.
I'd have to be more familiar with the language etc-update is written in, but the data type isn't the problem. It does need to be an integer, unless non-integer input is expected at some point. In my brief testing, it seemed like the select syntax of bash was more robust. Am I correct there? Would select be better?
Isn't it fixed already? :)
In svn r12053 I've fixed it to show a more human readable error message, like this: Value '1y' is not valid. Please enter an integer value: (In reply to comment #3) > Isn't it fixed already? :) Well, it was fixed in the sense that the program would continue to run. However, the error message was not very good.
This is fixed in 2.1.6_rc1 and 2.2_rc16.