[CDBI] Re: Transaction over parent and children has_a objects

Christopher H. Laco claco at chrislaco.com
Tue Aug 9 15:56:52 BST 2005


Edward J. Sabol wrote:
>>Does killing AutoCommit locally automatically kill autoupdate(1)?
> 
> 
> Not exactly, but AutoCommit trumps autoupdate.
> 
> 
Does that mean that with AutoCommit turned off locally, field updates 
won't be sent to the server even if autoupdate(1), or that they will 
still be sent to the server but not committed?

I guess I'm really trying to solve two problems. First, that I make as 
few roud trip to the DB as possible when updating fields under on 
circumstance, and also that all of the parent/child changes get one 
transaction.

Maybe I'm making this to hard on myself. Maybe I just change the 
parent/child classes __PACKAGE__->autoupdate(0), but flip that bit on 
when someone creates on using new(). I think this means all user created 
instances are autoupdate(1), by all instances inflated from 
has_a/has_many are still in __PACKAGE__->autoupdate(0) land.

-=Chris
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3178 bytes
Desc: S/MIME Cryptographic Signature
Url : http://lists.digitalcraftsmen.net/mailman/private/classdbi/attachments/20050809/014f2e1d/smime.bin


More information about the ClassDBI mailing list