Beta 5 or RC 1?

[Michael Smith-5]

Steve,

When we (DM and I) were testing this during the code sprint, using valgrind,
we could not reproduce it. I don't know if it was something specific with
that build that we compiled on the WMS shootout box or something else. It
might have also been that we misdiagnosed the problem as a memory leak when
we were just hammering the box to hard.

I'll try to run some valgrind tests on the shootout box and see if I get any
additional info.




On 10/29/09 5:59 AM, "Steve Lime" <[hidden email]> wrote:

> I added you as a cc on the ticket (#3187). I would think any single map draw
> with Oracle should
> leak memory so a test case should be easy to drum up. I'd look to the
> msLayerWhichItems()
> area first.
>
> Steve
>
>>>> On 10/28/2009 at 11:46 AM, in message
> <[hidden email]>, Alan Boudreault
> <[hidden email]> wrote:
>> Maybe I could invest in that bug if nobody is already on it. Does anyone
>> could
>> attach a test case ?
>>
>> Alan
>>
>> On October 28, 2009 12:03:25 pm Paul Ramsey wrote:
>>> The Oracle FCGI memory leaks seems like a blocker to me. It is new in
>>> 5.6, didn't exist in 5.4, and makes Oracle FCGI unusable.
>>>
>>> P.
>>>
>>> On Wed, Oct 28, 2009 at 8:10 AM, Steve Lime <[hidden email]>
>> wrote:
>>>> Hi all: I know folks are recovering from the conference or on walk about,
>>>> etc... We should think about the next release though. We can either cut
>>>> another beta or the first release candidate. I don't think there are any
>>>> *big* bugs although there still seem to be some unresolved rendering
>>>> issues between 5.4 and 5.6 pointed out by Havard.We can still fix issues
>>>> within an RC but they should be pretty minor. Any preferences?
>>>>
>>>> Steve
>>>>
>>>> _______________________________________________
>>>> mapserver-dev mailing list
>>>> [hidden email]
>>>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>>>
>>> _______________________________________________
>>> mapserver-dev mailing list
>>> [hidden email]
>>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>>>
>
> _______________________________________________
> mapserver-dev mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev

_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

[Steve Lime]

Hi all: I know folks are recovering from the conference or on walk about, etc... We should think about
the next release though. We can either cut another beta or the first release candidate. I don't think
there are any *big* bugs although there still seem to be some unresolved rendering issues between
5.4 and 5.6 pointed out by Havard.We can still fix issues within an RC but they should be pretty
minor. Any preferences?

Steve

_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

[Paul Ramsey]

The Oracle FCGI memory leaks seems like a blocker to me. It is new in
5.6, didn't exist in 5.4, and makes Oracle FCGI unusable.

P.

On Wed, Oct 28, 2009 at 8:10 AM, Steve Lime <[hidden email]> wrote:

> Hi all: I know folks are recovering from the conference or on walk about, etc... We should think about
> the next release though. We can either cut another beta or the first release candidate. I don't think
> there are any *big* bugs although there still seem to be some unresolved rendering issues between
> 5.4 and 5.6 pointed out by Havard.We can still fix issues within an RC but they should be pretty
> minor. Any preferences?
>
> Steve
>
> _______________________________________________
> mapserver-dev mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>

_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

[Alan Boudreault]

Maybe I could invest in that bug if nobody is already on it. Does anyone could
attach a test case ?

Alan

On October 28, 2009 12:03:25 pm Paul Ramsey wrote:
> The Oracle FCGI memory leaks seems like a blocker to me. It is new in
> 5.6, didn't exist in 5.4, and makes Oracle FCGI unusable.
>
> P.
>
> On Wed, Oct 28, 2009 at 8:10 AM, Steve Lime <[hidden email]>
wrote:

> > Hi all: I know folks are recovering from the conference or on walk about,
> > etc... We should think about the next release though. We can either cut
> > another beta or the first release candidate. I don't think there are any
> > *big* bugs although there still seem to be some unresolved rendering
> > issues between 5.4 and 5.6 pointed out by Havard.We can still fix issues
> > within an RC but they should be pretty minor. Any preferences?
> >
> > Steve
> >
> > _______________________________________________
> > mapserver-dev mailing list
> > [hidden email]
> > http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>
> _______________________________________________
> mapserver-dev mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>

--
Alan Boudreault
Mapgears
http://www.mapgears.com
_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

[Steve Lime]

I added you as a cc on the ticket (#3187). I would think any single map draw with Oracle should
leak memory so a test case should be easy to drum up. I'd look to the msLayerWhichItems()
area first.

Steve

>>> On 10/28/2009 at 11:46 AM, in message
<[hidden email]>, Alan Boudreault
<[hidden email]> wrote:

> Maybe I could invest in that bug if nobody is already on it. Does anyone
> could
> attach a test case ?
>
> Alan
>
> On October 28, 2009 12:03:25 pm Paul Ramsey wrote:
>> The Oracle FCGI memory leaks seems like a blocker to me. It is new in
>> 5.6, didn't exist in 5.4, and makes Oracle FCGI unusable.
>>
>> P.
>>
>> On Wed, Oct 28, 2009 at 8:10 AM, Steve Lime <[hidden email]>
> wrote:
>> > Hi all: I know folks are recovering from the conference or on walk about,
>> > etc... We should think about the next release though. We can either cut
>> > another beta or the first release candidate. I don't think there are any
>> > *big* bugs although there still seem to be some unresolved rendering
>> > issues between 5.4 and 5.6 pointed out by Havard.We can still fix issues
>> > within an RC but they should be pretty minor. Any preferences?
>> >
>> > Steve
>> >
>> > _______________________________________________
>> > mapserver-dev mailing list
>> > [hidden email]
>> > http://lists.osgeo.org/mailman/listinfo/mapserver-dev 
>>
>> _______________________________________________
>> mapserver-dev mailing list
>> [hidden email]
>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev 
>>

_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev

[Daniel Morissette]

Steve Lime wrote:
> Hi all: I know folks are recovering from the conference or on walk about, etc... We should think about
> the next release though. We can either cut another beta or the first release candidate. I don't think
> there are any *big* bugs although there still seem to be some unresolved rendering issues between
> 5.4 and 5.6 pointed out by Havard.We can still fix issues within an RC but they should be pretty
> minor. Any preferences?
>

I just got back this morning and haven't had a chance to go through the
accumulated mail yet.

Do you have pointers to the rendering issues that you are referring to?

With respect to the Oracle leak, it may make Oracle hardly usable under
FCGI, but it is very hard to reproduce and if we treat it as a blocker
we could be stuck for a few weeks. Alan is currently working on it, he
has been able to reproduce it earlier but cannot any more (need to
figure out what happened). I suggest we don't treat it as a blocker and
release a first RC with this known issue. We'd try to fix it ASAP,
before the final release if possible, or in a point release if we cannot
solve it in time.

Daniel
--
Daniel Morissette
http://www.mapgears.com/
_______________________________________________
mapserver-dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapserver-dev