[DBIx-Class-Devel] [dbix-class] Doc patch for explaining custom ResultSet classes (#49)

[Samuel Kaufman]

Ah OK, thanks for the feedback. I'll check it out and see what makes sense,
then I'll ask for comments in #dbix-class.
On Jul 1, 2014 5:52 AM, "Peter Rabbitson" <notifications at github.com> wrote:

> Ok. While on the surface this looked like a trivial thing it actually
> turned out to be much more involved. These pieces were committed to
> current/for_cpan_index so far:
>
>    - Some refactor of the docs 293a924
>    <https://github.com/dbsrgits/dbix-class/commit/293a9242a> and
>    downright purge of useless deps 77a6448
>    <https://github.com/dbsrgits/dbix-class/commit/77a6448dc>
>    - An easier-to-follow version of your datetime example a0a0da0
>    <https://github.com/dbsrgits/dbix-class/commit/a0a0da0ae>
>
> The remaining piece is pushed to wip/moo_resultsets_docs, and has the
> following outstanding problem that needs attention: The Moose example has
> only BUILDARGS yet the Moo example has both BUILDARGS and FOREIGNBUILDARGS.
> Please investigate, see which one is right and resubmit changes to 15c43cf
> <https://github.com/dbsrgits/dbix-class/commit/15c43cfda> (the example
> code has to be flawless, as the cargocult of this will be high)
>
> As an aside - there was an issue brought up yesterday on IRC concerning
> Moose-based Result classes (not ResultSet classes). Perhaps a separate
> patch to document that inconsistency would also be of use:
>
> <corgifex> ribasushi: 1) is using Moose with DBIx::Class supported at all? 2) is MooseX::NonMoose + extends 'DBIx::Class::Core' the right way?  3) is 'lazy' a sensible workaround for DBIx::Class not calling the constructor? 4) is there a better way to do this?
> <ribasushi> corgifex: in this case it depends what you mean by "supported"
> <ribasushi> corgifex:  the core DBIC result class has 2 separate constructors, which was a good idea at the time it was written, so if you expect constructor-based thingies to work, you will have to do some duct taping
> <ribasushi> corgifex: and as you already properly noted default non-lazy attributes do not work without constructor-based instantiation
> <ribasushi> corgifex: so it all depends on "how do you want things to interface" - it will be a tradeoff either way, and you need to decide where to make it
>
> Cheers and thanks again for looking into this
>
> —
> Reply to this email directly or view it on GitHub
> <https://github.com/dbsrgits/dbix-class/pull/49#issuecomment-47637403>.
>

---
Reply to this email directly or view it on GitHub:
https://github.com/dbsrgits/dbix-class/pull/49#issuecomment-47650130
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.scsys.co.uk/pipermail/dbix-class-devel/attachments/20140701/5394ffcf/attachment.htm>