I was playing with the following scenario:
IFS set to backslash,
using read (with and without -r),
triggering field splitting by passing multiple names to read,
triggering field splitting by expanding an unquoted parameter.
And, I found the following results:
(note: the ones with a # , that mean
I've discovered a bash bug that causes the "jobs" command to produce
incorrect output after an ERR trap has been defined and triggered.
I've seen it with bash 4.3.11(1)-release on Linux Mint 17 and bash
4.3.30(1)-release on Debian 6.
I have *not* seen it with 3.2.25(1)-release on CentOS 5.10.
The
