[Dbix-class] register_source() regression in 0.08102

[Peter Rabbitson]

Ian Wells wrote:
> 2009/5/7 Ronald J Kimball <rkimball+dbixclass at pangeamedia.com>:
>> What is the correct way to implement this with 0.08102?  I'm skeptical about
>> register_extra_source(), because I'm not registering an additional source
>> for UserSession; I'm registering a new source for UserSession200905.
> 
> The way the document reads to me:
> 
>> register_extra_source
>> Arguments: $moniker, $result_source
>> As "register_source" but should be used if the result class already has a source and you want to register an extra one.
> 
> ...it's not that you're registering a second ResultSource called
> UserSession (which you clearly aren't), it's that you're registering a
> second source *of* UserSession Result classes.  If that's the case the
> error message could do with the word 'Result' in it (which it would
> have if you used load_namespaces, but I guess you're using
> load_classes; perhaps the message could be changed for clarity?).  If
> that's not the case the docs could do with a touch-up.
> 
> In passing, I note that ResultSource's documentation refer to the
> '[rR]esult class' a couple of times, but don't actually define what
> the term means for a ResultSource.  Indeed, there's no implication
> that a ResultSource class has a corresponding Result class or how you
> might find out what it is, but there is an implication that the
> 'Result class' is used during deployment.  Perhaps someone who knows
> how the two relate could add it to the documentation in there
> somewhere?
> 

It is even better if the people who see a problem with the documentation
submit patches for the affected areas. Even if your text is technically
incorrect it will be written to clarify an area that is no-brainer to the
seasoned user. The no-brainer part is extremely important, as it means
that the document will never be fixed satisfactory unless you provide the
text that works for YOU on first reading.

So - patches welcome :)