cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kelven Yang <kelven.y...@citrix.com>
Subject Re: Review Request 12134: removed txn attribute from DB annotation
Date Wed, 04 Sep 2013 18:09:19 GMT


On 9/4/13 10:18 AM, "Darren Shepherd" <darren.s.shepherd@gmail.com> wrote:

>On 09/04/2013 10:04 AM, Alex Huang wrote:
>> Kelven and Darren should review this.
>>
>> That attribute used to be always read to determine if a method should
>>be intercepted at load time.  My understanding is the switch to using
>>spring as the injection mechanism caused this to be turned off and now
>>every single method is intercepted, which is terrible.  I don¹t know if
>>Kelven and/or Darren has any plans to fix this.  If they don¹t plan on
>>fixing it, then we should just commit it.
>>
>
>Alex,
>
>My plan is to essentially remove the existence of @DB.  I'm working on
>that now. 

+1 for removing @DB. The approach to mimic all the semantics in Spring for
@DB that I did last time was due to the time constraint I had. I tried to
minimized the code changes, therefore mimic @DB semantics is the cheapest
strategy.

Giving for what @DB is doing, it merely just helps developer to do a per
call-chain context open/close automatically, And open transaction context
is a very cheap operation which can be done by just doing it at
entry-point of the very top of the call chain. I should have done that
last time.

-Kelven

> In the meantime I think its good to just remove the txn
>attribute as the patch does.  I clicked "ship it", but deferred to you
>to commit.
>
>Darren


Mime
View raw message