rspec users - Feb 2010 - Testing code that integrates with Rspec

Hey all,

I''ve been using Rspec for a while now, but this is the first time
I''ve run
up against the following problem: how do you use Rspec to test code that
uses Rspec?

Basically my code takes as input a representation of example groups, and it
should dynamically create those example groups and then run them. I borrowed
this code from a open-source project that doesn''t appear to have tests
for
it; the code works, but I''d like to get it under test so I can improve
it.

Going through the Rspec code (v 1.3.0), it appears that this should be
possible.  There''s Spec::Runner::Options, which looks to be a decent
entry
point for kicking off an Rspec run (with #run_examples). It has something
attached to it called a Reporter, which looks like an observer that, if I
could mock it, I could sense exactly what my run of Rspec did.

Unfortunately, when I wrote a spec to invoke this code, both the code under
test and the spec itself seemed to talk to the same Reporter (the one that
got created first, for the the spec). So my spec and my code are entangled.

Any suggestions?

Here''s a gist with a stripped down version of my code and a description
of
what output it produced. I can work on this gist to make it actually
runnable if anybody is interested.

Thanks a bunch,
Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/rspec-users/attachments/20100223/51e83e61/attachment.html>

Oops, forgot the gist:

http://gist.github.com/312383

Thanks,
Mike

On Feb 23, 9:19?pm, Mike Grafton <graf... at gmail.com>
wrote:
> Hey all,
>
> I''ve been using Rspec for a while now, but this is the first time
I''ve run
> up against the following problem: how do you use Rspec to test code that
> uses Rspec?
>
> Basically my code takes as input a representation of example groups, and it
> should dynamically create those example groups and then run them. I
borrowed
> this code from a open-source project that doesn''t appear to have
tests for
> it; the code works, but I''d like to get it under test so I can
improve it.
>
> Going through the Rspec code (v 1.3.0), it appears that this should be
> possible. ?There''s Spec::Runner::Options, which looks to be a
decent entry
> point for kicking off an Rspec run (with #run_examples). It has something
> attached to it called a Reporter, which looks like an observer that, if I
> could mock it, I could sense exactly what my run of Rspec did.
>
> Unfortunately, when I wrote a spec to invoke this code, both the code under
> test and the spec itself seemed to talk to the same Reporter (the one that
> got created first, for the the spec). So my spec and my code are entangled.
>
> Any suggestions?
>
> Here''s a gist with a stripped down version of my code and a
description of
> what output it produced. I can work on this gist to make it actually
> runnable if anybody is interested.
>
> Thanks a bunch,
> Mike
>
> _______________________________________________
> rspec-users mailing list
> rspec-us... at
rubyforge.orghttp://rubyforge.org/mailman/listinfo/rspec-users