In talking to Frank today I learned that gdal is trying to push 1.7
out the door; as I understand it (from a conversation at foss4g) we
are hanging back a version due to memory leaks.
Is there any chance that we can be more specific about memory leaks?
Or can we test against 1.7 in time to provide feedback prior to 1.7
release.
Jody
Jody Garnett ha scritto:
In talking to Frank today I learned that gdal is trying to push 1.7
out the door; as I understand it (from a conversation at foss4g) we
are hanging back a version due to memory leaks.
Due to unforeseeable instability actually. GDAL is well tested in its
own common case, which is short usage: generally speaking desktop apps
don't run more than a few hours, and MapServer is more often than not
running without fastcgi enabled. Even when it is a segfault brings
down one worker, which is promptly replaced by the FastCGI runtime.
When it comes to GeoServer we have a single process running for weeks
or months. If it segfaults the service goes down, period.
It's much a more demanding working environment.
That means we need to test it a lot more and it makes sense to use
releases that are known to be very stable to start with, also given
time is usually short and thus there is none to be spent debugging
GDAL proper (the bindings themselves have to be debugged/improved to
start with).
Only Simone and his crew have the skill to port the bindings to GDAL 1.7
and at the moment, afaik, are working on a port towards 1.6.
Not sure if he has any time to check out 1.7, I cannot answer for him.
Just trying to explain that this is not as easy as it may seem.
Cheers
Andrea
--
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.
Ciao Jody,
please read below...
-------------------------------------------------------
Ing. Simone Giannecchini
GeoSolutions S.A.S.
Founder - Software Engineer
Via Carignoni 51
55041 Camaiore (LU)
Italy
phone: +39 0584983027
fax: +39 0584983027
mob: +39 333 8128928
http://www.geo-solutions.it
http://geo-solutions.blogspot.com/
http://simboss.blogspot.com/
http://www.linkedin.com/in/simonegiannecchini
-------------------------------------------------------
On Tue, Nov 24, 2009 at 6:00 AM, Jody Garnett <jody.garnett@anonymised.com> wrote:
In talking to Frank today I learned that gdal is trying to push 1.7
out the door; as I understand it (from a conversation at foss4g) we
are hanging back a version due to memory leaks.
This is my own policy. We decide to switch between version of gdal
only it the z in X.Y.Z is > 1. Infact right now on imageio-ext trunk
we are playing with gdal 1.6.x. This is done to try and track version
where memory leaks and or problems should have been already adressed.
Of course time and resources play a role, switching a version means
quite some work, therefore it is not that easy to do that
automatically when a new version comes out.
Is there any chance that we can be more specific about memory leaks?
Or can we test against 1.7 in time to provide feedback prior to 1.7
release.
I can talk to daniele tomorrow. But testing 1.7.x is not going to be
feasible in the short term, which means not before than next year. I
don't think this would work with the GDAL timeline. Anyway daniele has
been interacting with Even around our changes to the Java bindings and
it seems that most of them where already introduced in trunk,
therefore moving forward with GDAL versions should not be too painful
Ciao,
Simone.
Jody
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
Thanks Andrea; I did not imagine it was easy (only that there is a
window of opportunity made known to me by frank).
Jody
On Tue, Nov 24, 2009 at 7:11 PM, Andrea Aime <aaime@anonymised.com> wrote:
Jody Garnett ha scritto:
In talking to Frank today I learned that gdal is trying to push 1.7
out the door; as I understand it (from a conversation at foss4g) we
are hanging back a version due to memory leaks.Due to unforeseeable instability actually. GDAL is well tested in its
own common case, which is short usage: generally speaking desktop apps
don't run more than a few hours, and MapServer is more often than not
running without fastcgi enabled. Even when it is a segfault brings
down one worker, which is promptly replaced by the FastCGI runtime.When it comes to GeoServer we have a single process running for weeks
or months. If it segfaults the service goes down, period.
It's much a more demanding working environment.That means we need to test it a lot more and it makes sense to use
releases that are known to be very stable to start with, also given
time is usually short and thus there is none to be spent debugging
GDAL proper (the bindings themselves have to be debugged/improved to
start with).Only Simone and his crew have the skill to port the bindings to GDAL 1.7
and at the moment, afaik, are working on a port towards 1.6.Not sure if he has any time to check out 1.7, I cannot answer for him.
Just trying to explain that this is not as easy as it may seem.Cheers
Andrea--
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.