Hi all,
the question about missing JAI prompted me to check
1.5.x functionality without JAI... well, no surprise,
it does not start.
This is not good, Geoserver should start just fine instead,
and have only WMS and WFS working, whilst WCS should be
completely disabled in this configuration (eventually
with an error message stating that it cannot run without
JAI).
Or I am missing something and PSC decided that it's ok
to require JAI around? Don't remember, sorry.
Would it be difficult to do so? Moreover, are then any
parts of WMS that do require JAI even when rendering
just vector data?
No I think its generally decided that forcing users to have JAI just to run GeoServer is not acceptable. Brent can you confirm that we fixed this on trunk. I found some "stub" jars (that were being used for 1.3.x ) and I posted them in the maven repository.
We might be able to pull the same trick off for 1.5.x, however I think now with the GC stuff on there are some additional jai requirements. Hopefully they can clarify.
-Justin
Andrea Aime wrote:
Hi all,
the question about missing JAI prompted me to check
1.5.x functionality without JAI... well, no surprise,
it does not start.
This is not good, Geoserver should start just fine instead,
and have only WMS and WFS working, whilst WCS should be
completely disabled in this configuration (eventually
with an error message stating that it cannot run without
JAI).
Or I am missing something and PSC decided that it's ok
to require JAI around? Don't remember, sorry.
Would it be difficult to do so? Moreover, are then any
parts of WMS that do require JAI even when rendering
just vector data?
Out of the box it is supposed to run without JAI, and 1.3/1.4 do that. There was an issue where people couldn't compil geoserver 1.4 because it required JAI, but we included the jars so now it is possible.
I don't think the PSC voted on anything about JAI and I think we are sticking with "it should work without it". Mostly because JAI is a pain to install, and previous experience has shown that people do not like to have to install it.
I think it would be semi difficult to modify the GUI to recognize that JAI is absent and adjust accordingly, but not impossible. I'm assuming disabling the service would be quite easy.
Brent Owens
(The Open Planning Project)
Andrea Aime wrote:
Hi all,
the question about missing JAI prompted me to check
1.5.x functionality without JAI... well, no surprise,
it does not start.
This is not good, Geoserver should start just fine instead,
and have only WMS and WFS working, whilst WCS should be
completely disabled in this configuration (eventually
with an error message stating that it cannot run without
JAI).
Or I am missing something and PSC decided that it's ok
to require JAI around? Don't remember, sorry.
Would it be difficult to do so? Moreover, are then any
parts of WMS that do require JAI even when rendering
just vector data?
Hi all,
the question about missing JAI prompted me to check
1.5.x functionality without JAI... well, no surprise,
it does not start.
This is not good, Geoserver should start just fine instead,
and have only WMS and WFS working, whilst WCS should be
completely disabled in this configuration (eventually
with an error message stating that it cannot run without
JAI).
Or I am missing something and PSC decided that it's ok
to require JAI around? Don't remember, sorry.
No definitely not. It's a pretty firm requirement for me that most of GeoServer work fine without having to install JAI. Though on 1.4.x we get away by providing some jars (since gt2 has weird random requirements for JAI that don't actually depend on what it does well - though rumor is Martin got rid of them for 2.3).
I've been aware that WCS branch had that dependency. Fixing this is probably the thing for me that would move it from 'beta' to 'release candidate' status. Speaking of which, we still need that beta1 release.
Chris
Would it be difficult to do so? Moreover, are then any
parts of WMS that do require JAI even when rendering
just vector data?
I've been aware that WCS branch had that dependency. Fixing this is probably the thing for me that would move it from 'beta' to 'release candidate' status. Speaking of which, we still need that beta1 release.
Yeah, I know. I think Simone and Alessio are waiting for tomorrow's gt2 2.3 release, thought the release policy was changed so that we can release non final stuff against gt2 snapshots no?
Yes trunk does require JAI to run and to compile. 1.4 still works without it however.
Brent Owens
(The Open Planning Project)
Justin Deoliveira wrote:
No I think its generally decided that forcing users to have JAI just to run GeoServer is not acceptable. Brent can you confirm that we fixed this on trunk. I found some "stub" jars (that were being used for 1.3.x ) and I posted them in the maven repository.
We might be able to pull the same trick off for 1.5.x, however I think now with the GC stuff on there are some additional jai requirements. Hopefully they can clarify.
-Justin
Andrea Aime wrote:
Hi all,
the question about missing JAI prompted me to check
1.5.x functionality without JAI... well, no surprise,
it does not start.
This is not good, Geoserver should start just fine instead,
and have only WMS and WFS working, whilst WCS should be
completely disabled in this configuration (eventually
with an error message stating that it cannot run without
JAI).
Or I am missing something and PSC decided that it's ok
to require JAI around? Don't remember, sorry.
Would it be difficult to do so? Moreover, are then any
parts of WMS that do require JAI even when rendering
just vector data?
Leaving aside WCS, It is going to be quite difficult to have WMS
working with rasters without having JAI and ImageIO installed.
Simone.
On 11/30/06, Brent Owens <brentowens@anonymised.com> wrote:
Yes trunk does require JAI to run and to compile. 1.4 still works
without it however.
Brent Owens
(The Open Planning Project)
Justin Deoliveira wrote:
> No I think its generally decided that forcing users to have JAI just to
> run GeoServer is not acceptable. Brent can you confirm that we fixed
> this on trunk. I found some "stub" jars (that were being used for 1.3.x
> ) and I posted them in the maven repository.
>
> We might be able to pull the same trick off for 1.5.x, however I think
> now with the GC stuff on there are some additional jai requirements.
> Hopefully they can clarify.
>
> -Justin
>
> Andrea Aime wrote:
>
>> Hi all,
>> the question about missing JAI prompted me to check
>> 1.5.x functionality without JAI... well, no surprise,
>> it does not start.
>>
>> This is not good, Geoserver should start just fine instead,
>> and have only WMS and WFS working, whilst WCS should be
>> completely disabled in this configuration (eventually
>> with an error message stating that it cannot run without
>> JAI).
>>
>> Or I am missing something and PSC decided that it's ok
>> to require JAI around? Don't remember, sorry.
>>
>> Would it be difficult to do so? Moreover, are then any
>> parts of WMS that do require JAI even when rendering
>> just vector data?
>>
>> Cheers
>> Andrea
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Geoserver-devel mailing list
>> Geoserver-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>>
>> !DSPAM:1004,456f0903250172051017194!
>>
>
Could be possibly dynamically figure out if jai is installed at runtime, and if not, disable the rendering of rasters, and WCS?
-Justin
Simone Giannecchini wrote:
Leaving aside WCS, It is going to be quite difficult to have WMS
working with rasters without having JAI and ImageIO installed.
Simone.
On 11/30/06, Brent Owens <brentowens@anonymised.com> wrote:
Yes trunk does require JAI to run and to compile. 1.4 still works
without it however.
Brent Owens
(The Open Planning Project)
Justin Deoliveira wrote:
> No I think its generally decided that forcing users to have JAI just to
> run GeoServer is not acceptable. Brent can you confirm that we fixed
> this on trunk. I found some "stub" jars (that were being used for 1.3.x
> ) and I posted them in the maven repository.
>
> We might be able to pull the same trick off for 1.5.x, however I think
> now with the GC stuff on there are some additional jai requirements.
> Hopefully they can clarify.
>
> -Justin
>
> Andrea Aime wrote:
>
>> Hi all,
>> the question about missing JAI prompted me to check
>> 1.5.x functionality without JAI... well, no surprise,
>> it does not start.
>>
>> This is not good, Geoserver should start just fine instead,
>> and have only WMS and WFS working, whilst WCS should be
>> completely disabled in this configuration (eventually
>> with an error message stating that it cannot run without
>> JAI).
>>
>> Or I am missing something and PSC decided that it's ok
>> to require JAI around? Don't remember, sorry.
>>
>> Would it be difficult to do so? Moreover, are then any
>> parts of WMS that do require JAI even when rendering
>> just vector data?
>>
>> Cheers
>> Andrea
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Geoserver-devel mailing list
>> Geoserver-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>>
>> >>
>>
>
On 11/30/06, Justin Deoliveira <jdeolive@anonymised.com> wrote:
Could be possibly dynamically figure out if jai is installed at runtime,
Sure, just do a class.forname and catch an exception.
and if not, disable the rendering of rasters, and WCS?
I will answer with a question:
1>Is it possible to disable a runtime the rendering of features and the WFS?
Ciao,
Simone.
-Justin
Simone Giannecchini wrote:
> Leaving aside WCS, It is going to be quite difficult to have WMS
> working with rasters without having JAI and ImageIO installed.
>
> Simone.
>
> On 11/30/06, Brent Owens <brentowens@anonymised.com> wrote:
>> Yes trunk does require JAI to run and to compile. 1.4 still works
>> without it however.
>>
>> Brent Owens
>> (The Open Planning Project)
>>
>> Justin Deoliveira wrote:
>> > No I think its generally decided that forcing users to have JAI just to
>> > run GeoServer is not acceptable. Brent can you confirm that we fixed
>> > this on trunk. I found some "stub" jars (that were being used for 1.3.x
>> > ) and I posted them in the maven repository.
>> >
>> > We might be able to pull the same trick off for 1.5.x, however I think
>> > now with the GC stuff on there are some additional jai requirements.
>> > Hopefully they can clarify.
>> >
>> > -Justin
>> >
>> > Andrea Aime wrote:
>> >
>> >> Hi all,
>> >> the question about missing JAI prompted me to check
>> >> 1.5.x functionality without JAI... well, no surprise,
>> >> it does not start.
>> >>
>> >> This is not good, Geoserver should start just fine instead,
>> >> and have only WMS and WFS working, whilst WCS should be
>> >> completely disabled in this configuration (eventually
>> >> with an error message stating that it cannot run without
>> >> JAI).
>> >>
>> >> Or I am missing something and PSC decided that it's ok
>> >> to require JAI around? Don't remember, sorry.
>> >>
>> >> Would it be difficult to do so? Moreover, are then any
>> >> parts of WMS that do require JAI even when rendering
>> >> just vector data?
>> >>
>> >> Cheers
>> >> Andrea
>> >>
>> -------------------------------------------------------------------------
>> >> Take Surveys. Earn Cash. Influence the Future of IT
>> >> Join SourceForge.net's Techsay panel and you'll get the chance to
>> share your
>> >> opinions on IT & business topics through brief surveys - and earn cash
>> >>
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> >> _______________________________________________
>> >> Geoserver-devel mailing list
>> >> Geoserver-devel@lists.sourceforge.net
>> >> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>> >>
>> >> >>
>> >>
>> >
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to
>> share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Geoserver-devel mailing list
>> Geoserver-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>>
>
On 11/30/06, Justin Deoliveira <jdeolive@anonymised.com> wrote:
Could be possibly dynamically figure out if jai is installed at runtime,
Sure, just do a class.forname and catch an exception.
and if not, disable the rendering of rasters, and WCS?
I will answer with a question:
1>Is it possible to disable a runtime the rendering of features and the WFS?
I guess the real problem here is that removing a module is not
really possible because configuration and user interface are not
modularized, so we are stuck with something which is not really
modular, once a component has a configuration part, you cannot
get away without it...
On 11/30/06, Justin Deoliveira <jdeolive@anonymised.com> wrote:
Could be possibly dynamically figure out if jai is installed at runtime,
Sure, just do a class.forname and catch an exception.
and if not, disable the rendering of rasters, and WCS?
I will answer with a question:
1>Is it possible to disable a runtime the rendering of features and the WFS?
I guess the real problem here is that removing a module is not
really possible because configuration and user interface are not
modularized, so we are stuck with something which is not really
modular, once a component has a configuration part, you cannot
get away without it...
Yeah, our system isn't really conclusive to disabling certain modules, its kind of a all or nothing thing at the moment. On the spring side of things we could pull it off though. All it would take is a bean fired up at container startup that did the Class.forName check , and if everything a go, load the appropriate bean definitions.
We could hack something into the GUI to disable the options. Won't be pretty and will take some work.
Brent Owens
(The Open Planning Project)
Andrea Aime wrote:
Simone Giannecchini ha scritto:
On 11/30/06, Justin Deoliveira <jdeolive@anonymised.com> wrote:
Could be possibly dynamically figure out if jai is installed at runtime,
Sure, just do a class.forname and catch an exception.
and if not, disable the rendering of rasters, and WCS?
I will answer with a question:
1>Is it possible to disable a runtime the rendering of features and the WFS?
I guess the real problem here is that removing a module is not
really possible because configuration and user interface are not
modularized, so we are stuck with something which is not really
modular, once a component has a configuration part, you cannot
get away without it...
The source code for the jai-imageio-core and jai-imageio-demos projects is licensed under the open source Berkeley Software Distribution (BSD) License.
I weirdly couldn't find links to downloads from there directly, so maybe we have to build it ourselves to have it be bsd? But regardless the source is bsd, so I think we're fine.
Chris
By not including ImageIO into the distribution we give up a few
new WMS formats but it should be working... should, nobody really tried that...