[Catalyst] Duplicate entries with C::P::Session::Store::DBIC and MySQL - new findings

[Tobias Kremer]

Quoting Daniel Westermann-Clark <dwc at pobox.com>:

> On 2008-08-26 14:18:18 +0200, Tobias Kremer wrote:
> > Just out of pure curiosity: Why is it that there are dedicated
> > "flash:<hash>" entries in the storage for the flash? Wouldn't the
> > session be enough?
>
> The "flash:" rows were used for compatibility with Store::DBI.  We can
> break compatibility if people find the it not very useful.

We definitely should! IMHO five queries per request to the database just for the
session and flash stuff is inacceptable. The alternative approach could very
well get rid of the race-condition, too. I think it's really worth it.

I have to admit that I don't understand what compatibility with Store::DBI
you're talking about. AFAICT the Store::DBI source doesn't separate between
"session" and "flash" (does it support the flash at all?). Maybe you (or Andy)
can shed some light on this!

--Tobias