Code review comment for lp:~lifeless/storm/with

On Wed, Mar 23, 2011 at 11:24 AM, Gustavo Niemeyer <email address hidden> wrote:
>
>> I don't think we'd want the example you have of a full Person table defined on
>> top of a WITH clause: unless it loaded into the storm cache in the same place
>> our regular Person table does (which IIRC would raise errors?).
>
> My point is that whether you find this useful right now or not, it will just work,
> because it's just another way to define a table.

Ok

>> @all
>> I'm still at a loss on how to code this up - it sounds nice but I need
>> pointers: I got as far as I did with this on blind luck following the path of
>> least resistance. I'm not asking for the code to be written for me, but I need
>> to know where existing similar tests are that I can copy; how the compiler
>> works so I can tell how to change it - or at least some hints pointing me in
>> the appropriate directions...
>
> It's hard to help you with plain "I don't get it. What do I do?" questions
> after we provide you some guidance.  We need more specific questions from
> you taking into account what we already said, otherwise we'll end up just
> saying the same thing again.

I don't mean to whinge, but AFAICT I haven't had any guidance : no
code pointers, design or structure pointers, other than 'something
like ClassAlias'. As a specific example; you don't like 'with_', but I
can't see any way to say to the compiler that a custom type should
compile to the front of the query - before the
SELECT/UPDATE/INSERT/DELETE.

> We can also have a voice call at some point to have a more concrete
> implementation conversation, if you're finding hard to understand the
> guidance we're providing and would find better to chat over it.

That could be very helpful - thanks. I'm timeshifted towards UK at the
moment while we work through a backlog of RT tickets impacting the LP
deployment environment, but I should have some overlap with you in
your afternoons, or in a week or so I'll be back to more NZ normal
time and have better overlap.

I've pushed up a sketch that may be what you meant, and toggled this
back to needs review.

Cheers,
Rob