[Rspec-devel] define_instance_method, stub_with, and mock_with

[David Astels]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On 4-Sep-06, at 9:37 PM, Jay Levitt wrote:

> David Astels wrote:
>> On 4-Sep-06, at 8:49 PM, David Astels wrote:
>>
>>> I will be looking hard at mocha/stubba tonight, but they look like a
>>> good way to go.  I'm quite happy to have us depricate rSpec's mock
>>> framework in favour of mocha/stubba.
>>
>> Well.. on first glance Mocha seems significantly inferior to rSpec
>> mocks.  No support for consecutive calls, primitive argument
>> matching, no ordering, and a rather clunky syntax.
>
> Is that inherent to Mocha's structure, though, or just a function  
> of the
> fact that Mocha's all of 45 days old?  I wonder if contributing these
> features to Mocha isn't better than creating a
> pluggable-framework-framework.  In fact, I think I can argue that in
> either case:
>
Probably.  Both mocha and rspec-mock are built partly on the same  
base (Schmock)... it should be easy to add these capabilities to Mocha.

> If these features are inherent to the mock object, then there's no  
> way a
> pluggable framework could support such verifications on Mocha objects.
> If, however, they're inherent to the verification layer, then  
> there's no
> reason they can't be contributed to Mocha, thus getting rspec out  
> of the
> mock business.

Verification is internal to the mock and it's expectations.  The call  
to verify the mock is exposed to the outer framework (test/unit or  
rspec).

Either way is fine with me.  I do wonder what's better for the  
community... a single, common mock framework... or some selection/ 
variation?

I generally tend to favour the latter.

Dave

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (Darwin)

iD8DBQFE/M4qauez/L4x7g4RAmX1AKCQU+Gy7gMYJLvG3iyRx+TGcePO8wCgzFoD
HoRDo1L2k6LZhbipQ5kl9Pc=
=DNnF
-----END PGP SIGNATURE-----