mmmmmm DBXTalk and image save sounded a bell... When we developed OpenDBXDriver, we idea was to make it as simple as possible for the final user. One thing we did was to automatically disconnect all opened connections at system shutdown...so that not to let connections open in the DB.�

Now I realized that we have this:

DBXConnection class >> shutDown

self

allInstancesDo: [:each | each shutDown]

but it should actually be�

DBXConnection class >> shutDown: quitting

� � quitting ifTrue: [ self allInstancesDo: [:each | each shutDown]

I fixed this some time ago (OpenDBXDriver-MarianoMartinezPeck.38)

Maybe this could be related?�

Paul are you also using DBXTalk just by chance in that image?

Cheers,

On Wed, Sep 4, 2013 at 1:00 PM, Esteban Lorenzano <estebanlm@gmail.com> wrote:

phil was working with 3.0 (afaik)
mariano was having problems with flushing monticello caches.

I think it was not the same.

Esteban

On Sep 4, 2013, at 5:32 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:

> On 09/04/2013 12:54 AM, Esteban Lorenzano wrote:
>> I'm not sure what we are seeing in pharo3 is the same problem reported in pharo2.
>> Of course if it is, we will backport.
>> But "inflation" problem in ph3 is very consistent and repeatable, the problem reported in ph2... well, I never seen it before, and we have just one report (and I assume that we have more than one hundred production projects with that version).
>> So is likely something different (also some of the detected leaks are due to changes in ph3, not present in the older version).
>>
>> Esteban
>
>
> heres two more reports about image size in another thread (Phil & Mariano):
>
> http://forum.world.st/Large-images-reasons-WAS-Re-Pharo-2-0-with-Seaside-DBXTalk-GlorpDBX-Magritte-3-TWBS-is-getting-slower-tp4701743.html
>
>
>

--
Mariano
http://marianopeck.wordpress.com