Parsers should parse scenario outline like standard scenarios

[dcendents]

Hi,

• I use cfparse when I want to parse a list.
• I use scenario outlines when I want to test my scenario with different values

I cannot mix both conditions together.

e.g.:

Scenario Outline: mix case
  Given the emails <emails>
  ...

Examples:
  | emails |
  | someone@somewhere.com |
  | someone@somewhere.com, someone.else@somewhere.com |


Scenario: other scenario
  Given the emails email1@server.net, email2@server.net
  ...

It would be very useful if I could then have the following step

@given(parsers.cfparse("Given the emails {emails:str+}". extra_types=dict(str=str)), target_fixture="emails")
def emails(emails)
    return emails

Ok I could split the list myself, but it's ugly

@given("Given the emails <emails>")
@given(parsers.cfparse("Given the emails {emails:str+}". extra_types=dict(str=str)), target_fixture="emails")
def emails(emails)
    return emails if isinstance(emails, list) else emails.split(",")

The same kind of support has been asked a few times in #293 but since it has nothing to do with the original problem I fear it has been lost through all the discussions.

Thanks

[TBBle]

We've had to use exactly this workaround of the double-@given, once with <> examples and once with parsers.parse, so implementing this would be great.

I recently worked in Godog (the official Cucumber implementation for Go) and it works this way; the steps never see the <example> placeholder, they parse the resulting interpolated statement like if it was written literally, and for pytest-bdd that would let us apply all the various benefits of parsers.X (built-in or custom) as desired.

I haven't checked #433 to see if that's how it worked, but it was reverted in #440 so I guess that's academic.