[Geoserver-devel] POST once more

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.

It seems that the POST functionality has been broken again.

There is a dispatcher servlet in front that decides which service to
use, and that decides whether it should use the POST or GET method.
The problem is that when you POST the dispatcher do not check the
content-type http header to check what it is actually getting.
It do now, as it used to do, consider all POST's as data streams.
Meaning that it thinks the request body only contain xml input.
That is an incorrect assumption.

Consider the html form:

<form action="http://localhost/geoserver/wms&quot; method="get">
<input name="bbox" value="593036.1,6604888.0,753648.25,6787297.5"/>
<input name="layers" value="topp:hordaland"/>
<input name="Format" value="image/png"/>
<input name="request" value="GetMap"/>
<input name="width" value="420"/>
<input name="height" value="477"/>
<input name="SRS" value="EPSG:32632"/>
<input name="SLD_BODY" value="[LOADS OF XML]"/>
</form>

The above form submit works fine if the sld xml is not too large.
If it gets too large you need to POST it in the request body instead of
as url parameters. In the browser this is simple:

<form action="http://localhost/geoserver/wms&quot; method="post">
...
same parameters as above
...
</form>

This is posted and the sld xml can be as large as it needs to. However
this POST do not work, even when it got the exact same parameters as the
GET form.

The problem is that the dispatcher only checks if the method is post or
get, but do not consider the content type. The browser submits a form
with the content type: application/x-www-form-urlencoded
If this content type is received it is a parametrized post, meaning that
request.getParameter("Format") will work on both submits, get and post.
However, the request.getQueryString() do not work when it is posted, as
there is no query.

When the above code is posted the dispatcher creates a temporary file on
the disk, which the system thinks is pure xml: wfsdispatch4947686tmp

This file contains:
bbox=593036.1%2C6604888.0%2C753648.25%2C6787297.5&layers=topp%3Ahordaland&Format=image%2Fpng&request=GetMap&width=420&height=477&srs=EPSG%3A32632&SLD_BODY=[LOADS
OF XML]

Which in fact is what has been posted, and not the xml itself. This
because the dispatcher takes the request stream and just pipe it through
to the temporary file.

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?

Best
Magne

Magne Skjeret ha scritto:

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:

Magne Skjeret ha scritto:

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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:1004,4567246f239661804284693!

--
Justin Deoliveira
jdeolive@anonymised.com
The Open Planning Project
http://topp.openplans.org

I have submitted a new patch that will handle this case. Perhaps you
could review it Magne?

The original patch lives in the GetMap servlet, in the isUrlEncoded method.

-Justin

Justin Deoliveira wrote:

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:

Magne Skjeret ha scritto:

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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

--
Justin Deoliveira
jdeolive@anonymised.com
The Open Planning Project
http://topp.openplans.org

Magne Skjeret wrote:

G'Day - thanks for finding this (sure wish we had your review earlier).

I was just going to comment on one thing ....

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
  

This was not a political decision - the svn server went belly-up and we had to canvas the developer
list and ask people to set up themselves an account on codehaus Xircles and so on.

So no policitcs; you probably just missed the email. If you can add yourself to the project (their
are directions on the wiki) I can sign in and hit the okay button.

All the best,
Jody

Best
Magne

-------------------------------------------------------------------------
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
  

Justin Deoliveira wrote:

I have submitted a new patch that will handle this case. Perhaps you
could review it Magne?
  

Hi

Yes, from what I can read this should work.
Great work Justin.

One major advantage with your fix next to mine is that this will work
for any request, but mine only for GetMap requests. The advantage of
centralized code. Does this actually mean that the check in the GetMap
servlet can be removed, as the dispatcher will call the correct method
directly? No point in wasting valuable cycles if we can avoid it.

Magne

The original patch lives in the GetMap servlet, in the isUrlEncoded method.

-Justin

Justin Deoliveira wrote:
  

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:
    

Magne Skjeret ha scritto:
      

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.
        

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
        

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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

Jody Garnett wrote:

Magne Skjeret wrote:

G'Day - thanks for finding this (sure wish we had your review earlier).
  

Hi
Sorry about that. :frowning:
I have been so busy lately that I hardly have time to pay attention to
what is going on in the mailing-list.

I was just going to comment on one thing ....
  

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
  

This was not a political decision - the svn server went belly-up and we
had to canvas the developer
list and ask people to set up themselves an account on codehaus Xircles
and so on.

So no policitcs; you probably just missed the email. If you can add
yourself to the project (their
are directions on the wiki) I can sign in and hit the okay button.
  

Oh, no. I meant if you had done a political desission on how to handle
the input on GET vs. POST.
I think I remember the mail that we were to register again, and that is
why I said that I did not think I had commiter access anymore.

Sorry about the confusion there.

Magne

Cool, thanks for the review. And good point, that code GetMap can be
safely removed. I have applied the patch to the dispatcher. Andrea is
doing to some testing and it things are ok it will get into the next
release I believe.

-Justin

Magne Skjeret wrote:

Justin Deoliveira wrote:

I have submitted a new patch that will handle this case. Perhaps you
could review it Magne?
  

Hi

Yes, from what I can read this should work.
Great work Justin.

One major advantage with your fix next to mine is that this will work
for any request, but mine only for GetMap requests. The advantage of
centralized code. Does this actually mean that the check in the GetMap
servlet can be removed, as the dispatcher will call the correct method
directly? No point in wasting valuable cycles if we can avoid it.

Magne

The original patch lives in the GetMap servlet, in the isUrlEncoded method.

-Justin

Justin Deoliveira wrote:
  

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:
    

Magne Skjeret ha scritto:
      

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.
        

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
        

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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

!DSPAM:1004,45673c34252181510810322!

--
Justin Deoliveira
jdeolive@anonymised.com
The Open Planning Project
http://topp.openplans.org

Justin Deoliveira wrote:

Hi

If someone is interested in seeing how we use the geoserver, here is a link
for the Norwegian institute of public health site that me and my peers
have created.

http://norgeshelsa.no/norgeshelsaen/

I think most of the indicators have a map attached. Use the map icon on
the toolbar, or follow the link
*http://tinyurl.com/ucfsj to a prepared map.
When viewing the map try to turn on labels and the values and see how
labels actually are removed. The algorithm that decide there is space or
not is a bit overeager. We have even added the vendor specific items in
the xml to allow label overlap, but it still do not work properly.

Well, I will report that issue to jira to have a proper discussion on it.

Magne
*

Cool, thanks for the review. And good point, that code GetMap can be
safely removed. I have applied the patch to the dispatcher. Andrea is
doing to some testing and it things are ok it will get into the next
release I believe.

-Justin

Magne Skjeret wrote:
  

Justin Deoliveira wrote:
    

I have submitted a new patch that will handle this case. Perhaps you
could review it Magne?
  

Hi

Yes, from what I can read this should work.
Great work Justin.

One major advantage with your fix next to mine is that this will work
for any request, but mine only for GetMap requests. The advantage of
centralized code. Does this actually mean that the check in the GetMap
servlet can be removed, as the dispatcher will call the correct method
directly? No point in wasting valuable cycles if we can avoid it.

Magne

The original patch lives in the GetMap servlet, in the isUrlEncoded method.

-Justin

Justin Deoliveira wrote:
  

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:
    

Magne Skjeret ha scritto:
      

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.
        

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
        

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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

!DSPAM:1004,45673c34252181510810322!

Looks great Magne!! Always nice to hear about a GeoServer success
story!! Hopefully we can sort out the labelling issues soon.

-Justin

Magne Skjeret wrote:

Justin Deoliveira wrote:

Hi

If someone is interested in seeing how we use the geoserver, here is a link
for the Norwegian institute of public health site that me and my peers
have created.

http://norgeshelsa.no/norgeshelsaen/

I think most of the indicators have a map attached. Use the map icon on
the toolbar, or follow the link
*http://tinyurl.com/ucfsj to a prepared map.
When viewing the map try to turn on labels and the values and see how
labels actually are removed. The algorithm that decide there is space or
not is a bit overeager. We have even added the vendor specific items in
the xml to allow label overlap, but it still do not work properly.

Well, I will report that issue to jira to have a proper discussion on it.

Magne
*

Cool, thanks for the review. And good point, that code GetMap can be
safely removed. I have applied the patch to the dispatcher. Andrea is
doing to some testing and it things are ok it will get into the next
release I believe.

-Justin

Magne Skjeret wrote:
  

Justin Deoliveira wrote:
    

I have submitted a new patch that will handle this case. Perhaps you
could review it Magne?
  

Hi

Yes, from what I can read this should work.
Great work Justin.

One major advantage with your fix next to mine is that this will work
for any request, but mine only for GetMap requests. The advantage of
centralized code. Does this actually mean that the check in the GetMap
servlet can be removed, as the dispatcher will call the correct method
directly? No point in wasting valuable cycles if we can avoid it.

Magne

The original patch lives in the GetMap servlet, in the isUrlEncoded method.

-Justin

Justin Deoliveira wrote:
  

The bug in question is http://jira.codehaus.org/browse/GEOS-441.

I remember this patch, and yes the dispatcher does not take this into
account unfortunately, this is one is my bad :(.

The patch is relatively straight forward. All is needed is that in a
post request, if the content type is "application/x-www-form-urlencoded"
then it should just be treated as a get request.

-Justin

aaime wrote:
    

Magne Skjeret ha scritto:
      

Hi

Long time since I have been here now, and I am sorry that I have come
back to complain a bit.

We just upgraded our server with geoserver 1.4rc3.
        

...

I know that geoserver supports to stream xml directly in the body of the
request. The content type is then not
"application/x-www-form-urlencoded", but rather "text/xml". When
text/xml content type is received it should be fine to stream the
content directly in. However if the content type is
application/x-www-form-urlencoded then it is safe to pass it on to the
doGet method as long as you do not use request.getQueryString();

This is what I did my fix for the geoserver 1.3 release. To differ
between form submits and "data" submits.

I can fix this myself, but I am unsure if I got commiter access anymore.
Or maybe this has been a political decision that I am unaware of?
        

I think you may still have, but nevertheless 1.4.x is in deep freeze, and
we are releasing 1.4.0rc4 today, so do not commit.
Provide us with a patch instead, so that we can review it, and hopefully
integrate it in 1.4.0rc4 (there won't be a rc5, I cannot hold commits
on gt2 2.2.x anymore...)

I hope Chris shows up and speaks up too, but in the meantime, are there
any references about your old patch so that I can put it in context?
(jira issues, mailing list posts...).

Cheers
Andrea Aime

-------------------------------------------------------------------------
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

-------------------------------------------------------------------------
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:1004,4567506e263261116498154!

--
Justin Deoliveira
jdeolive@anonymised.com
The Open Planning Project
http://topp.openplans.org

Magne Skjeret ha scritto:

Justin Deoliveira wrote:

Hi

If someone is interested in seeing how we use the geoserver, here is a link
for the Norwegian institute of public health site that me and my peers
have created.

http://norgeshelsa.no/norgeshelsaen/

I think most of the indicators have a map attached. Use the map icon on
the toolbar, or follow the link
*http://tinyurl.com/ucfsj to a prepared map.
When viewing the map try to turn on labels and the values and see how
labels actually are removed. The algorithm that decide there is space or
not is a bit overeager.

Indeed, not nice. Well, we'll have to work on the labelling algorithm
anyways in order to properly support tiling, this will be a good occasion to try and fix this issue too.

And yes, a jira entry for this is much appreciated :slight_smile:

Cheers
Andrea

P: Btw, very nice web application you have over there :slight_smile: