on_error.md

on_error

Counterpart to the on_error: attribute.

Takes a function definition (or a variable pointing to one of them) and makes sure the function is called when an error occurs at the current level or below.

sequence
  set f.l []
  on_error (def err \ push f.l err.error.msg)
  push f.l 0
  push f.l x # <-- will fail because `x` is unknown
  push f.l 1

Where the field l ends up containing [ 0, "cannot find \"x\"" ].

set f.l []

define error_handler err
  push f.l err.error.msg

sequence
  on_error error_handler
  # ...

on and on_error

"on_error" is made to allow for on error, so that:

sequence
  on error
    push f.l err.msg # a block with an `err` variable
  # ...

gets turned into:

sequence
  on_error
    def err # a anonymous function definition with an `err` argument
      push f.l err.msg
  # ...

on_error kriteria

Similarly to Ruby, one may catch certain types of errors.

In the follow example, Charly is tasked with "it failed" and the err if there is a Ruby error of class RuntimeError happens in the sequence:

sequence
  on_error class: 'RuntimeError' (def err \ charly "it failed", err)
  alice 'do this'
  bob 'do that'

One can shorten it to:

sequence
  on_error 'RuntimeError' (def err \ charly "it failed", err)
  alice 'do this'
  bob 'do that'

But this short version will check the Ruby class name and then the error message.

In the next example, Charly is tasked with "it timed out" if the error class or the error message match the regex /timeout/:

sequence
  on_error (/timeout/) (def err \ charly "it timed out")
  on_error (def err \ charly "it failed", err)
  alice 'do this'
  bob 'do that'

Order matters.

Please note that you can't set a criteria when you're using the on_error: attribute, as in:

sequence on_error: (def err \ charly "it failed", err)
  alice 'do this'
  bob 'do that'

see also

On, on_cancel.

• source
• on_error spec