I agree with kilon. Sorry if it is an stupid question but why not identify
them by grouping into "examples" protocols?
The rule could be matching "examples*" named protocols, then you can group
examples that match different features ("examples-featureA",
"examples-featureB" ...).
On 27 October 2014 13:56, Peter Uhnák <[email protected]> wrote:
> > So you can both name your example whatever you want and keep your code
> separate , clean and nice.
> I believe the reason for convention was clickable examples in system
> browser. So as far as tests go even if they are in separate classes you can
> only run methods beginning with 'test' and not other support (setUp/..)
> methods. The same would go for examples. You may need other non-example
> methods to run the example so you need to discriminate them.
>
> Also shouldn't this convention for pragmas vs method names be more global?
> For example
> ComposableModel class>>defaultSpec vs <spec> and so on.
>
> Peter
>
> On Mon, Oct 27, 2014 at 1:33 PM, kilon alios <[email protected]>
> wrote:
>
>> I am not a fan of pragmas myself, but then I am also not an enemy.
>>
>> Personal I would prefer if we would follow a similar approach to unit
>> testing, meaning separate classes for examples. I dont like the idea of
>> mixing regular code with examples even just for a single example. Separate
>> class will also mean there would be no need for the name of the method to
>> contain "example". So you can both name your example whatever you want and
>> keep your code separate , clean and nice.
>>
>> On Mon, Oct 27, 2014 at 2:15 PM, stepharo <[email protected]> wrote:
>>
>>> Hi doru and others
>>>
>>> so finally what is the conventions we should use for example?
>>>
>>> exampleFooBar
>>> <example>
>>>
>>> Stef
>>>
>>>
>>>
>>
>
