Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Ben Caradoc-Davies
Connections must be idle for at least 5 minutes (and up to 10) with these settings: minEvictableIdleTimeMillis="30" timeBetweenEvictionRunsMillis="30" If requests are very infrequent, then I agree that they will be much slower. For more frequent requests there should be no loss of perfo

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Andrea Aime
On Wed, Jun 4, 2014 at 8:37 AM, Ben Caradoc-Davies < ben.caradoc-dav...@csiro.au> wrote: > What about minIdle=0? That seemed to fix my problems with firewalls > between GeoServer and the database. > Yes, that combination would work, but the cure will imho damage the patient in the process, as it

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Ben Caradoc-Davies
What about minIdle=0? That seemed to fix my problems with firewalls between GeoServer and the database. (Still supporting upgrade to DBCP 2.0.) On 04/06/14 14:35, Andrea Aime wrote: > Those are supported, but they are not useful, as the pool won't evict > connections > below the minIdle threshol

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Andrea Aime
On Wed, Jun 4, 2014 at 3:19 AM, Ben Caradoc-Davies < ben.caradoc-dav...@csiro.au> wrote: > So DBCP 1.4 does not support > minEvictableIdleTimeMillis/timeBetweenEvictionRunsMillis > or equivalent? > http://docs.geoserver.org/latest/en/user/data/app- > schema/data-stores.html#jndi Those are suppo

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Ben Caradoc-Davies
So DBCP 1.4 does not support minEvictableIdleTimeMillis/timeBetweenEvictionRunsMillis or equivalent? http://docs.geoserver.org/latest/en/user/data/app-schema/data-stores.html#jndi +1 for upgrade to DBCP 2.0 based on our positive experience with 1.4. Kind regards, Ben. On 03/06/14 16:40, Andrea

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Jody Garnett
One more question - were you going to back port this work for the stable series? Usually we have at least a show of making a module available for public review prior to going to extension status. I would be happy to make a beta release (focused on Java 7 and a new wfs implementation) in order to m

Re: [Geotools-devel] Extending process annotations to allow generic metadata specification

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 3:57 PM, Andrea Aime wrote: > And here we go with a pull request: > https://github.com/geotools/geotools/pull/469 > Bah, I'm losing it... I did implement the version using @KVP. Pull request now updated to use simple strings as in the second example Cheers Andrea -- ==

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Niels Charlier
No, these appear not to be present in the wfs-ng module. Kind Regards Niels On 03/06/14 15:12, Andrea Aime wrote: On Tue, Jun 3, 2014 at 3:09 PM, Niels Charlier > wrote: Well I ported the tinyows and mapserver tests from wfs to wfs-ng... There is a num

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Justin Deoliveira
On Tue, Jun 3, 2014 at 6:23 AM, Andrea Aime wrote: > On Tue, Jun 3, 2014 at 2:19 PM, Niels Charlier wrote: > >> As far as I know atm, everything is done as requested. I suggest we >> move the discussion to the github pull request, and start from there. The >> goal is still to get the module sup

Re: [Geotools-devel] Extending process annotations to allow generic metadata specification

2014-06-03 Thread Andrea Aime
And here we go with a pull request: https://github.com/geotools/geotools/pull/469 Cheers Andrea On Sat, May 31, 2014 at 3:45 PM, Andrea Aime wrote: > On Sat, May 31, 2014 at 3:41 PM, Jody Garnett > wrote: > >> >> So is meta taking a list of strings, each string parsed as a key value pai

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 3:12 PM, Andrea Aime wrote: > The parameters I'm concerned, because they were recently (roughly one year > ago) added, are the following: > > final String namespaceOverride = (String) NAMESPACE.lookUp(params); > final Boolean useDefaultSRS = (Boolean) > USED

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 3:09 PM, Niels Charlier wrote: > Well I ported the tinyows and mapserver tests from wfs to wfs-ng... >> > There is a number of issues with cascading those two, I'm not 100% sure the tests cover them all. The parameters I'm concerned, because they were recently (roughly one

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Niels Charlier
On 03/06/14 14:36, Andrea Aime wrote: > > This might become a problem for GeoServer users that have WFS data > stores already configured in > GeoServer. Would it be difficult to create, say, a factory that acts > as a bridge for those? > > Also the old WFS store had a number of configurations abo

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Jody Garnett
It seems pretty good, if anything can survive uDig testing :) Note we can also use the class path trick you did for the shapefile & shapefile-ng handover. Allow new factory to work against the old parameters, but perform a check to see if the old implementation is on the class path first. If it is

[Geotools-devel] [jira] (GEOT-4811) Allow process annotations to specify free metadata entries for the associated Parameter objects

2014-06-03 Thread Andrea Aime (JIRA)
Title: Message Title Andrea Aime created an issue

[Geotools-devel] Jenkins build is back to normal : GeoTools-11.x #58

2014-06-03 Thread winbuild
See -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 2:37 PM, Jody Garnett wrote: > I would recommend killing the old wfs module (it is unsupported for a > reason - namely it has not attracted funding). With that "out of the way" > wfs-ng is free to take its place. > Yep, that's the simplest and cleaner approach, that's why

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 2:32 PM, Niels Charlier wrote: > (like, if I just replace the dependencies we have towards the old wfs > module, what will happen?) > > > In an existing configuration, you'd have to make new data stores, because > wfs-ng is a different type of data store. > This might beco

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Jody Garnett
I would recommend killing the old wfs module (it is unsupported for a reason - namely it has not attracted funding). With that "out of the way" wfs-ng is free to take its place. I kind of wish we could perform this change while wfs and wfs-ng are both unsupported - but you indicated your contract

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Niels Charlier
On 03/06/14 14:23, Andrea Aime wrote: On Tue, Jun 3, 2014 at 2:19 PM, Niels Charlier > wrote: As far as I know atm, everything is done as requested. I suggest we move the discussion to the github pull request, and start from there. The goal is still to get the

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Andrea Aime
On Tue, Jun 3, 2014 at 2:19 PM, Niels Charlier wrote: > As far as I know atm, everything is done as requested. I suggest we move > the discussion to the github pull request, and start from there. The goal > is still to get the module supported asap. > Hi, about the module getting supported, tha

[Geotools-devel] Build failed in Jenkins: GeoTools-11.x #57

2014-06-03 Thread winbuild
See Changes: [lago.88] Modified ImageIO-Ext version from 1.1.9 to 1.1.10 -- [...truncated 833 lines...] [INFO] Feature-Pregeneralized SKIPPED [INFO] grass

Re: [Geotools-devel] wfs-ng improvements

2014-06-03 Thread Niels Charlier
Hi Jody, Justin, I updated the branch / pull request to fix the getInfo stuff. I tested from udig (works) as well as added unit tests for this functionality. As far as I know atm, everything is done as requested. I suggest we move the discussion to the github pull request, and start from ther

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Christian Mueller
+1 from here On Tue, Jun 3, 2014 at 1:19 PM, Ian Turton wrote: > +1 from me > > Ian > > > On 3 June 2014 09:40, Andrea Aime wrote: > >> On Mon, Apr 28, 2014 at 4:31 PM, Justin Deoliveira < >> jdeol...@boundlessgeo.com> wrote: >> >>> I'll take the conservative approach of "if it ain't broke don

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Ian Turton
+1 from me Ian On 3 June 2014 09:40, Andrea Aime wrote: > On Mon, Apr 28, 2014 at 4:31 PM, Justin Deoliveira < > jdeol...@boundlessgeo.com> wrote: > >> I'll take the conservative approach of "if it ain't broke don't fix it". >> With the java7 upgrade still ongoing we already have enough moving

Re: [Geotools-devel] Upgrading our connection pool library

2014-06-03 Thread Andrea Aime
On Mon, Apr 28, 2014 at 4:31 PM, Justin Deoliveira < jdeol...@boundlessgeo.com> wrote: > I'll take the conservative approach of "if it ain't broke don't fix it". > With the java7 upgrade still ongoing we already have enough moving parts so > perhaps we should put this one off for now since it does