Hi Don,
It must be something over here. Because a projection change does not
remove the layers. I can even force garbage collection on the profiler
and the layer objects still show as live. I'll keep looking. Thanks for
your help.
On 09/19/2011 03:56 PM, Don Dietrick wrote:
> Hi Marc,
>
> I think what's happening is the MapBean is holding on to removed
> layers until the projection changes. Once a layer has been removed
> and the map projection is changed, the layer should be garbage
> collected.
>
> The purpose of this is to keep the layer in the dark about when it's
> actually removed, in case the user clicks a layer off/on/etc without
> changing the projection. This prevents having layers redo work that
> is valid for the current projection and keeps them ready to
> immediately repaint their contents if re-added to the map.
>
> I just ran a test using a profiler, and my layers were gone after the
> projection change.
>
> Hope this helps,
>
> Don
>
> On Mon, Sep 19, 2011 at 5:15 PM, Marc Benstein<MARC.D.BENSTEIN_at_saic.com> wrote:
>> Just to confirm that this is potentially an openmap issue. I added
>> com.bbn.openmap.layer.test.TestLayer and removed the layer using MapHandler
>> as well as LayerHandler. There is definitely a leak. However, it could be in
>> the JDK.
>>
>> On 09/19/2011 02:55 PM, Marc Benstein wrote:
>>>
>>> Don,
>>> Thanks for the reply. I was wrong before _all_ layers that I add to
>>> MapHandler or the LayerHandler fail to release the added layer object on
>>> remove.
>>>
>>> Something is keeping these objects around in memory. The setList(null)
>>> call only frees the data which is the majority of memory. However, the
>>> layer object is still sitting around somewhere and the setList(null) is
>>> not needed if the layer itself gets freed. I will dig further. Unless
>>> you know of the cause.
>>>
>>> On 09/16/2011 04:51 PM, Don Dietrick wrote:
>>>>
>>>> Hi Marc,
>>>>
>>>> Removing them from the MapHandler should work, although I generally
>>>> deal with the LayerHandler directly. Adding setList(null) is a good
>>>> idea for dispose(), though, I'll add that.
>>>>
>>>> - Don
>>>>
>>>> On Fri, Sep 16, 2011 at 6:36 PM, Marc
>>>> Benstein<MARC.D.BENSTEIN_at_saic.com> wrote:
>>>>>
>>>>> Hi openmap-users,
>>>>>
>>>>> I am having trouble removing layers from the MapHandler. I have a few
>>>>> layers
>>>>> that extend OMGraphicHandlerLayer and some that extend Layer. I add
>>>>> them all
>>>>> to MapHandler using the add(Object) method. However, even after I remove
>>>>> them using MapHandler.remove(Object) they are still sitting around
>>>>> leaking
>>>>> memory. I found that if I override dispose() and call super.dispose()
>>>>> along
>>>>> with setList(null) the layers that extend OMGraphicHandlerLayer are
>>>>> garbage
>>>>> collected. Unfortunately, the classes that extend Layer do not get gc'd.
>>>>> They are very simple layers that only override
>>>>> projectionChanged(ProjectionEvent) and paintComponent(Graphics).
>>>>> Consequently, dispose() along with remove(Object) calls do not remove
>>>>> them
>>>>> completely.
>>>>>
>>>>> I have a suggestion to the maintainers to override dispose() in
>>>>> OMGraphicHandlerLayer and calling setList(null) along with
>>>>> super.dispose().
>>>>>
>>>>> How do I _really_ remove the classes that extend Layer from the
>>>>> MapHandler?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> --
>>>>> Marc Benstein
>>>>> 505-830-6763
>>>>>
>>>>> --
>>>>> [To unsubscribe to this list send an email to "majdart_at_bbn.com"
>>>>> with the following text in the BODY of the message "unsubscribe
>>>>> openmap-users"]
>>>>>
>>>
>>> --
>>> [To unsubscribe to this list send an email to "majdart_at_bbn.com"
>>> with the following text in the BODY of the message "unsubscribe
>>> openmap-users"]
>>
>> --
>> [To unsubscribe to this list send an email to "majdart_at_bbn.com"
>> with the following text in the BODY of the message "unsubscribe
>> openmap-users"]
>>
--
[To unsubscribe to this list send an email to "majdart_at_bbn.com"
with the following text in the BODY of the message "unsubscribe openmap-users"]
Received on Mon Sep 19 2011 - 18:01:22 EDT