persistence@glassfish.java.net

RE: Issue 401: Moving size() outside a for loop?

From: Gordon Yorke <gordon.yorke_at_oracle.com>
Date: Fri, 17 Mar 2006 17:00:35 -0500

Hello Marina,
        Feel free to formulate the changes and we can try to integrate them and test them but given the timeline we may not get them in before the Monday cutoff. (CTS takes 5 hours here and I believe 12 hours for Shelly)
--Gordon

-----Original Message-----
From: Marina Vatkina [mailto:Marina.Vatkina_at_Sun.COM]
Sent: Friday, March 17, 2006 4:49 PM
To: persistence_at_glassfish.dev.java.net
Subject: Re: Issue 401: Moving size() outside a for loop?


Hello Gordon,

Yes, I know. This is why I asked. One the other hand, is it worth trying?
After I fix hints/property names to be toplink..., I can make the changes
and we can run all the tests (CTS, perf, unit, etc.). If it causes at
least 1 regression, I'll attach the changes to the bug report, and we'll
decide what to do with this later.

Would you agree?

thanks,
-marina


Gordon Yorke wrote:
> Hello Marina,
> I think changing 63 files at this point (so close to our CTS goal) could be problematic?
> --Gordon
>
> -----Original Message-----
> From: Marina Vatkina [mailto:Marina.Vatkina_at_Sun.COM]
> Sent: Friday, March 17, 2006 4:39 PM
> To: persistence
> Subject: Issue 401: Moving size() outside a for loop?
>
>
> Peter, Tom,
>
> There are 63 files where I found size() in the for loop (using grep,
> so obviously I need to go over each case to double check).
>
> Do you see any problem with moving all of them out of the for loop?
>
> I also saw (at least once) where the code "manually" creates an
> Object[] from a collection. Do you see a problem if I replace it
> with toArray() call?
>
> While it's a lot of changes, I can make them before Monday so that
> we can try to run them with performance benchmarks and see if it
> makes a difference.
>
> What do you think?
>
> thanks,
> -marina