[ebeip90]

If you're using ZSH, this will give you a Python-style backtrace showing the specific line of each failure, even with nested function calls.

https://gist.github.com/zachriggle/8574964d2e3078cdfae84b574...

e.g.

    An error occurred on ./test:18
    Frame 1 (./test:21)
    18           false
    19       }
    20
    21   >>> a

    Frame 2 (./test:6)
    3        source TRAPERR.zsh
    4
    5        a() {
    6    >>>     b
    7        }
    8
    9        b() {

    Frame 3 (./test:10)
    7        }
    8
    9        b() {
    10   >>>     c
    11       }
    12
    13       c() {

    Frame 4 (./test:14)
    11       }
    12
    13       c() {
    14   >>>     d
    15       }
    16
    17       d() {

    Frame 5 (./test:18)
    15       }
    16
    17       d() {
    18   >>>     false
    19       }
    20
    21       a

[drran]

I can do this in bash too (partially), with `set -eE` and bash-modules:

  #!/bin/bash
  . import.sh strict log
  a() {
    b
  }
  b() {
    c
  }
  c() {
    d
  }
  d() {
    false
  }
  
  a

Result:

  $ ./test.sh
  [test.sh] PANIC: Uncaught error.
  at d(./test.sh:13)
  at c(./test.sh:10)
  at b(./test.sh:7)
  at a(./test.sh:4)
  at main(./test.sh:16)

However, it stops to work at some point, e.g. in `for` loop, or in a subshell. zsh is better in this regard.