I wanted to revisit this again... Planning to kill this setting (with
API deprecation) to reduce branching throughout the code. Please voice
your objections if you have any.
Andrus
On Jan 25, 2006, at 1:04 AM, Andrus Adamchik wrote:
Hmm... now that I am going through the places where it is used, it
doesn't look that bad anymore (partially due to the
GenericSelectQuery -> QueryMetadata refactoring I've done in the
last few days). I guess I'll keep it for now, as it may become more
important once we implement other types of inheritance in addition
to the single-table variety.
Sorry for the noise :-)
Andrus
On Jan 24, 2006, at 5:33 PM, Andrus Adamchik wrote:
Great. There is no objections. I'll go ahead with it.
Andrus
On Jan 23, 2006, at 3:39 PM, Cris Daniluk wrote:
I've never used it, but I *have* been annoyed by reading through the
spaghetti code it spawns :)
On 1/23/06, Gentry, Michael (Contractor) <[EMAIL PROTECTED]
> wrote:
I've never used it ...
-----Original Message-----
From: Andrus Adamchik [mailto:[EMAIL PROTECTED]
Sent: Saturday, January 21, 2006 4:40 PM
To: Cayenne Devel
Subject: Deprecating "isResolvingInherited"
Anybody ever set "isResolvingInherited" to FALSE?
Supposedly this can buy some performance improvements, but the
resulting object graph ends up being inconsistent, besides we
need to
pass this flag around in many places, resulting in spaghetti code.
I am thinking of deprecating this flag. Any objections?
Andrus
