[Dbix-class] Re: 0.08 breaks DBIx::Class::EncodedColumn

[Guillermo Roditi]

On Tue, Feb 3, 2009 at 9:13 PM, Guillermo Roditi <groditi at gmail.com> wrote:

>
> Ok, so it looks like somehow something changed and now set_column is
>>> getting called with the encoded value and the value is being re-encoded
>>>
>>>
>> AHA! So i think I got it. create does new() and then insert()
>>
>> Looks like old insert did not use set_column, and the latest version doe=
s.
>>
>>
>> What to do, fix it in EncodedColumn and break back compat or is this a
>> dbic issue?
>
>
> OK sorry for the volume, but now I am stuck because I can't even fix it by
> removing my override of new since EncodedColumn hashes all the values on
> set, and so they are plaintext between new and insert, which would break
> back compat. this is a hairy one. something will have to break, so  if you
> use EncodedColumn please put in your $0.02 now so we can fix it in the way
> that will not screw over too many people...
>

Failing test added to the multi_stuff branch, as instructed by mst.

-- =

Guillermo Roditi (groditi)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.scsys.co.uk/pipermail/dbix-class/attachments/20090204/1a7=
12eca/attachment.htm