MapServer 5.0.0 released!

[Daniel Morissette]

The MapServer team is proud to announce the release of MapServer 5.0!

Perhaps the most important new feature in this release is support for
map rendering using the Anti-grain Geometry (AGG) graphics library
(www.antigrain.com). AGG brings high quality 2-D rendering capabilities
to MapServer that will allow users to create maps with vastly improved
cartographic quality. Equally important is that using AGG results in a
minimal reduction in performance. All symbology supported by the default
GD (www.libgd.org) graphics library are also supported using AGG. We are
excited about the future possibilities of bringing high-end cartography
to on-demand web mapping.

This new version also includes dozens of small bug fixes, enhancements
and performance improvements. Among the new features are: style and
label attribute attribute binding; lookup table-based raster color
correction; dynamic charting (pie and bar); explicit label prioritizing;
enhanced debugging and logging; dynamic allocation for layers, classes,
styles and symbols; improved memory management and garbage collection
for MapScript; numerous improvements to OGC specification support, and
lots more. A list of RFCs related to this release is included at the end
of this message.

The source code release is available at the usual location on the
MapServer site:

   http://mapserver.gis.umn.edu/download/current

This page also contains links to binary distributions that will
hopefully be updated with 5.0 binaries shortly.

Users of 4.10 and older should review the 5.0 migration guide which
documents the changes that must be made to your application when
upgrading from MapServer 4.10 to 5.0:

   http://mapserver.gis.umn.edu/download/current/MIGRATION_GUIDE.TXT/

Finally, here is a summary of MapServer 5.0 feature/enhancement with the
related RFCs. For more details on a specific item, please refer to the
corresponding RFC document at http://mapserver.gis.umn.edu/development/rfc

New features:
------------

- MS RFC 19: Added Style and Label attribute binding

- MS RFC 21: Raster Color Correction via color lookup table

- MS RFC 27: Added label priority

- MS RFC 29: Added dynamic charting (pie and bar charts)

- MS RFC 31: New mechanism to load/set objects via URL using mapfile syntax

- MS RFC 32: Added support for map rendering using the AGG library for
   better output quality


Long time issues resolved:
-------------------------

- MS RFC 17: Use dynamic allocation for symbols, layers, classes and styles
   (got rid of the static limit on the number of instances of each in a map)

- MS RFC 24: Improved memory management and garbage collection for MapScript

- MS RFC 26: Terminology cleanup (layer transparency renamed to opacity,
   scale becomes scaledenom, symbol style becomes symbol pattern)

- MS RFC 28: Enhanced the debug/logging mechanism to facilitate
   troubleshooting and tuning applications. Added support for multiple
   debug levels and more control on output location.

[Gregor at HostGIS]

So, where's the release party? :)

Thanks a million, Steve, Daniel, and the zillions of others. Mapserver
makes low-cost cartography possible.

--
Gregor Mosheh / Greg Allensworth
System Administrator, HostGIS cartographic development & hosting services
http://www.HostGIS.com/

"Remember that no one cares if you can back up,
  only if you can restore." - AMANDA

[Daniel Morissette]

Gregor Mosheh wrote:
> So, where's the release party? :)
>

Victoria, BC, next week!  :)

Daniel
--
Daniel Morissette
http://www.mapgears.com/

[Nolte, Tim]

Some javascript/style in this post has been disabled (why?)

Firstly, thanks for all of the excellent work on MapServer! The availability of a quality package to the community is something not to be taken lightly. One question I have about this final release is in direct response to your statement, "Equally important is that using AGG results in a minimal reduction in performance.". I guess I'm really wondering how you are quantifying that? My hope is that this final release of 5.0.0 has additional performance updates for using AGG. I've struggled with some performance issues on our current MapServer and finally had to move back to using GD because of the performance difference. I had a layer that rendered in about 14 secs via AGG but then only 3 secs via GD. This really ended up being the deal breaker for using AGG. I'll be looking over the 5.0 migration guide to see if there are some tips there that will help with performance. I'd love to be able to go back to using AGG as I prefer the quality of the maps.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web:    www.ipcswirelessinc.com

[Howard Butler]

I would also like to add a link to the HISTORY document, which shows  
in more detail how impressive and how much work this release was.  It  
is also the place to find out quickly if your pet issue or bug was  
fixed this release.

<http://trac.osgeo.org/mapserver/browser/tags/rel-5-0-0/mapserver/ 
HISTORY.TXT>

Great work guys!  See you next week to drink some beers.

Howard

On Sep 17, 2007, at 3:23 PM, Daniel Morissette wrote:

> The MapServer team is proud to announce the release of MapServer 5.0!
>
> Perhaps the most important new feature in this release is support  
> for map rendering using the Anti-grain Geometry (AGG) graphics  
> library (www.antigrain.com). AGG brings high quality 2-D rendering  
> capabilities to MapServer that will allow users to create maps with  
> vastly improved cartographic quality. Equally important is that  
> using AGG results in a minimal reduction in performance. All  
> symbology supported by the default GD (www.libgd.org) graphics  
> library are also supported using AGG. We are excited about the  
> future possibilities of bringing high-end cartography to on-demand  
> web mapping.
>
> This new version also includes dozens of small bug fixes,  
> enhancements and performance improvements. Among the new features  
> are: style and label attribute attribute binding; lookup table-
> based raster color correction; dynamic charting (pie and bar);  
> explicit label prioritizing; enhanced debugging and logging;  
> dynamic allocation for layers, classes, styles and symbols;  
> improved memory management and garbage collection for MapScript;  
> numerous improvements to OGC specification support, and lots more.  
> A list of RFCs related to this release is included at the end of  
> this message.
>
> The source code release is available at the usual location on the  
> MapServer site:
>
>   http://mapserver.gis.umn.edu/download/current
>
> This page also contains links to binary distributions that will  
> hopefully be updated with 5.0 binaries shortly.
>
> Users of 4.10 and older should review the 5.0 migration guide which  
> documents the changes that must be made to your application when  
> upgrading from MapServer 4.10 to 5.0:
>
>   http://mapserver.gis.umn.edu/download/current/MIGRATION_GUIDE.TXT/
>
> Finally, here is a summary of MapServer 5.0 feature/enhancement  
> with the related RFCs. For more details on a specific item, please  
> refer to the corresponding RFC document at http://
> mapserver.gis.umn.edu/development/rfc
>
> New features:
> ------------
>
> - MS RFC 19: Added Style and Label attribute binding
>
> - MS RFC 21: Raster Color Correction via color lookup table
>
> - MS RFC 27: Added label priority
>
> - MS RFC 29: Added dynamic charting (pie and bar charts)
>
> - MS RFC 31: New mechanism to load/set objects via URL using  
> mapfile syntax
>
> - MS RFC 32: Added support for map rendering using the AGG library for
>   better output quality
>
>
> Long time issues resolved:
> -------------------------
>
> - MS RFC 17: Use dynamic allocation for symbols, layers, classes  
> and styles
>   (got rid of the static limit on the number of instances of each  
> in a map)
>
> - MS RFC 24: Improved memory management and garbage collection for  
> MapScript
>
> - MS RFC 26: Terminology cleanup (layer transparency renamed to  
> opacity,
>   scale becomes scaledenom, symbol style becomes symbol pattern)
>
> - MS RFC 28: Enhanced the debug/logging mechanism to facilitate
>   troubleshooting and tuning applications. Added support for multiple
>   debug levels and more control on output location.

[P Kishor]

My compliments to the chefs... drink 5.0.0 meters of beer. I will miss
you all, but will make sure to tipple a few in your honor in Paris, a
pale imitation of Victoria BC.

On 9/17/07, Howard Butler <[hidden email]> wrote:

> I would also like to add a link to the HISTORY document, which shows
> in more detail how impressive and how much work this release was.  It
> is also the place to find out quickly if your pet issue or bug was
> fixed this release.
>
> <http://trac.osgeo.org/mapserver/browser/tags/rel-5-0-0/mapserver/
> HISTORY.TXT>
>
> Great work guys!  See you next week to drink some beers.
>
>
> Howard
>
> On Sep 17, 2007, at 3:23 PM, Daniel Morissette wrote:
>
> > The MapServer team is proud to announce the release of MapServer 5.0!
> >
> > Perhaps the most important new feature in this release is support
> > for map rendering using the Anti-grain Geometry (AGG) graphics
> > library (www.antigrain.com). AGG brings high quality 2-D rendering
> > capabilities to MapServer that will allow users to create maps with
> > vastly improved cartographic quality. Equally important is that
> > using AGG results in a minimal reduction in performance. All
> > symbology supported by the default GD (www.libgd.org) graphics
> > library are also supported using AGG. We are excited about the
> > future possibilities of bringing high-end cartography to on-demand
> > web mapping.
> >
> > This new version also includes dozens of small bug fixes,
> > enhancements and performance improvements. Among the new features
> > are: style and label attribute attribute binding; lookup table-
> > based raster color correction; dynamic charting (pie and bar);
> > explicit label prioritizing; enhanced debugging and logging;
> > dynamic allocation for layers, classes, styles and symbols;
> > improved memory management and garbage collection for MapScript;
> > numerous improvements to OGC specification support, and lots more.
> > A list of RFCs related to this release is included at the end of
> > this message.
> >
> > The source code release is available at the usual location on the
> > MapServer site:
> >
> >   http://mapserver.gis.umn.edu/download/current
> >
> > This page also contains links to binary distributions that will
> > hopefully be updated with 5.0 binaries shortly.
> >
> > Users of 4.10 and older should review the 5.0 migration guide which
> > documents the changes that must be made to your application when
> > upgrading from MapServer 4.10 to 5.0:
> >
> >   http://mapserver.gis.umn.edu/download/current/MIGRATION_GUIDE.TXT/
> >
> > Finally, here is a summary of MapServer 5.0 feature/enhancement
> > with the related RFCs. For more details on a specific item, please
> > refer to the corresponding RFC document at http://
> > mapserver.gis.umn.edu/development/rfc
> >
> > New features:
> > ------------
> >
> > - MS RFC 19: Added Style and Label attribute binding
> >
> > - MS RFC 21: Raster Color Correction via color lookup table
> >
> > - MS RFC 27: Added label priority
> >
> > - MS RFC 29: Added dynamic charting (pie and bar charts)
> >
> > - MS RFC 31: New mechanism to load/set objects via URL using
> > mapfile syntax
> >
> > - MS RFC 32: Added support for map rendering using the AGG library for
> >   better output quality
> >
> >
> > Long time issues resolved:
> > -------------------------
> >
> > - MS RFC 17: Use dynamic allocation for symbols, layers, classes
> > and styles
> >   (got rid of the static limit on the number of instances of each
> > in a map)
> >
> > - MS RFC 24: Improved memory management and garbage collection for
> > MapScript
> >
> > - MS RFC 26: Terminology cleanup (layer transparency renamed to
> > opacity,
> >   scale becomes scaledenom, symbol style becomes symbol pattern)
> >
> > - MS RFC 28: Enhanced the debug/logging mechanism to facilitate
> >   troubleshooting and tuning applications. Added support for multiple
> >   debug levels and more control on output location.
>

--
Puneet Kishor
http://punkish.eidesis.org/
Nelson Institute for Environmental Studies
http://www.nelson.wisc.edu/
Open Source Geospatial Foundation (OSGeo)
http://www.osgeo.org/
Summer 2007 S&T Policy Fellow, The National Academies
http://www.nas.edu/

[Kralidis,Tom [Ontario]]

 
Great work all.  I won't be at foss4g this year, but will live vicariously through everyone this time around.
 
..Tom
 

On 9/17/07, Howard Butler <[hidden email]> wrote:

> I would also like to add a link to the HISTORY document, which shows
> in more detail how impressive and how much work this release was.  It
> is also the place to find out quickly if your pet issue or bug was
> fixed this release.
>
> <http://trac.osgeo.org/mapserver/browser/tags/rel-5-0-0/mapserver/
> HISTORY.TXT>
>
> Great work guys!  See you next week to drink some beers.
>
>
> Howard
>
> On Sep 17, 2007, at 3:23 PM, Daniel Morissette wrote:
>
> > The MapServer team is proud to announce the release of MapServer 5.0!
> >
> > Perhaps the most important new feature in this release is support
> > for map rendering using the Anti-grain Geometry (AGG) graphics
> > library (www.antigrain.com). AGG brings high quality 2-D rendering
> > capabilities to MapServer that will allow users to create maps with
> > vastly improved cartographic quality. Equally important is that
> > using AGG results in a minimal reduction in performance. All
> > symbology supported by the default GD (www.libgd.org) graphics
> > library are also supported using AGG. We are excited about the
> > future possibilities of bringing high-end cartography to on-demand
> > web mapping.
> >
> > This new version also includes dozens of small bug fixes,
> > enhancements and performance improvements. Among the new features
> > are: style and label attribute attribute binding; lookup table-
> > based raster color correction; dynamic charting (pie and bar);
> > explicit label prioritizing; enhanced debugging and logging;
> > dynamic allocation for layers, classes, styles and symbols;
> > improved memory management and garbage collection for MapScript;
> > numerous improvements to OGC specification support, and lots more.
> > A list of RFCs related to this release is included at the end of
> > this message.
> >
> > The source code release is available at the usual location on the
> > MapServer site:
> >
> >   http://mapserver.gis.umn.edu/download/current
> >
> > This page also contains links to binary distributions that will
> > hopefully be updated with 5.0 binaries shortly.
> >
> > Users of 4.10 and older should review the 5.0 migration guide which
> > documents the changes that must be made to your application when
> > upgrading from MapServer 4.10 to 5.0:
> >
> >   http://mapserver.gis.umn.edu/download/current/MIGRATION_GUIDE.TXT/
> >
> > Finally, here is a summary of MapServer 5.0 feature/enhancement
> > with the related RFCs. For more details on a specific item, please
> > refer to the corresponding RFC document at http:// <http:///>
> > mapserver.gis.umn.edu/development/rfc
> >
> > New features:
> > ------------
> >
> > - MS RFC 19: Added Style and Label attribute binding
> >
> > - MS RFC 21: Raster Color Correction via color lookup table
> >
> > - MS RFC 27: Added label priority
> >
> > - MS RFC 29: Added dynamic charting (pie and bar charts)
> >
> > - MS RFC 31: New mechanism to load/set objects via URL using
> > mapfile syntax
> >
> > - MS RFC 32: Added support for map rendering using the AGG library for
> >   better output quality
> >
> >
> > Long time issues resolved:
> > -------------------------
> >
> > - MS RFC 17: Use dynamic allocation for symbols, layers, classes
> > and styles
> >   (got rid of the static limit on the number of instances of each
> > in a map)
> >
> > - MS RFC 24: Improved memory management and garbage collection for
> > MapScript
> >
> > - MS RFC 26: Terminology cleanup (layer transparency renamed to
> > opacity,
> >   scale becomes scaledenom, symbol style becomes symbol pattern)
> >
> > - MS RFC 28: Enhanced the debug/logging mechanism to facilitate
> >   troubleshooting and tuning applications. Added support for multiple
> >   debug levels and more control on output location.
>

[Daniel Morissette]

Nolte, Tim wrote:
> [...]  One question I have about this final release is in
> direct response to your statement, "Equally important is that using AGG
> results in a minimal reduction in performance.". I guess I'm really
> wondering how you are quantifying that?


That's based on feedback we got from most people who tested the new AGG
rendering. In most cases the impact was minimal and I think we even
heard of some cases where performance was better.

However I think there may have been a few exceptions, and your case is
clearly one. I didn't work directly on the AGG rendering myself so I
cannot comment on what could result in this performance hit. I'll let
those who know that part better comment on that.

BTW, if you could share a small test case (data + mapfile) that
demonstrates the difference then that would increase your chances of
getting help.

Daniel
--
Daniel Morissette
http://www.mapgears.com/

[Nolte, Tim]

I hope that you didn't take my comment as though I was questioning the
validity of the comment. I was really just curious because of my
performance issues. I'll throw together a test case, who should I send
it to, just the list?

I posted (actually double posted because I'm blind..grr..not really) a
second note about being able to turn of antialiasing for a layer. It
really the 5 layers I'm seeing such a drastic performance difference
with. These layers are not very nice polygons, they are pretty detailed
with a lot of curves. I'm no expert by any means so there is probably
more I can do to speed up these layers. I have already broke them up
into separate smaller shapefiles as well as index them. Thanks.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com


-----Original Message-----
From: UMN MapServer Users List [mailto:[hidden email]] On
Behalf Of Daniel Morissette
Sent: Tuesday, September 18, 2007 10:21 AM
To: [hidden email]
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

Nolte, Tim wrote:
> [...]  One question I have about this final release is in
> direct response to your statement, "Equally important is that using
AGG
> results in a minimal reduction in performance.". I guess I'm really
> wondering how you are quantifying that?


That's based on feedback we got from most people who tested the new AGG
rendering. In most cases the impact was minimal and I think we even
heard of some cases where performance was better.

However I think there may have been a few exceptions, and your case is
clearly one. I didn't work directly on the AGG rendering myself so I
cannot comment on what could result in this performance hit. I'll let
those who know that part better comment on that.

BTW, if you could share a small test case (data + mapfile) that
demonstrates the difference then that would increase your chances of
getting help.

Daniel
--
Daniel Morissette
http://www.mapgears.com/

[thomas bonfort]

On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> I hope that you didn't take my comment as though I was questioning the
> validity of the comment. I was really just curious because of my
> performance issues. I'll throw together a test case, who should I send
> it to, just the list?

you open a bug report and attach the archive to it...
or you can send the archive directly to me

cheers, tb

[thomas bonfort]

Tim,

Attached is a patch you can apply to mapagg.cpp - if you can't apply
it I'll mail you the whole file.

it basically switches to aliased rendering if you use
SYMBOL
TYPE SIMPLE
ANTIALIASED FALSE
NAME 'aliased'
END

it'll work *only* on line and polygon layers, if you specify a
STYLE
NAME 'aliased'
COLOR  r g b
END

nothing other is supported (line widths, or polygon outlines)

I'd be interested to hear about timings if you can try it

tb

[fastlinesandpolygons.patch]

---

Index: mapagg.cpp
===================================================================
--- mapagg.cpp  (revision 6894)
+++ mapagg.cpp  (working copy)
@@ -101,7 +101,6 @@
 #endif
 
 
-
 typedef agg::rgba8 color_type;
 typedef agg::font_engine_freetype_int16 font_engine_type;
 typedef agg::font_cache_manager<font_engine_type> font_manager_type;
@@ -113,6 +112,8 @@
 typedef agg::renderer_scanline_aa_solid<renderer_base> renderer_aa;
 typedef agg::renderer_outline_aa<renderer_base> renderer_oaa;
 typedef agg::renderer_primitives<renderer_base> renderer_prim;
+typedef agg::renderer_scanline_bin_solid<renderer_base> renderer_bin;
+
 typedef agg::rasterizer_outline_aa<renderer_oaa> rasterizer_outline_aa;
 typedef agg::rasterizer_outline <renderer_prim> rasterizer_outline;
 typedef agg::rasterizer_scanline_aa<> rasterizer_scanline;
@@ -228,6 +229,9 @@
         ren_base(thePixelFormat),
         ren_base_pre(thePixelFormat_pre),
         ren_aa(ren_base),
+        ren_prim(ren_base),
+        ren_bin(ren_base),
+        ras_oa(ren_prim),
         m_fman(m_feng)
         {
             
@@ -279,7 +283,7 @@
         ras_aa.reset();
         ras_aa.filling_rule(agg::fill_non_zero);
         ren_aa.color(msToAGGColor(c));
-
+        
         if (dashstylelength <= 0) {
             agg::conv_stroke<agg::path_storage> stroke(p);  
             stroke.width(width);
@@ -301,6 +305,14 @@
         agg::render_scanlines(ras_aa, sl, ren_aa);    
     }
     
+    ///render an aliased 1 pixel width polyline
+    ///@param p the path_storage containing the vertexes of the polyline
+    ///@param c the color of the line
+    void renderPolylineFast(agg::path_storage &p,colorObj *c) {
+        ren_prim.line_color(msToAGGColor(c));
+        ras_oa.add_path(p);
+    }
+    
     ///brush a polyline with a vector symbol. draws the vector symbol in a temporary
     ///image that will be used as a brush for rendering the polyline. this function
     ///doesn't do any actual rendering, it only creates the brush and forwards to the
@@ -393,6 +405,18 @@
             agg::render_scanlines ( ras_aa, sl, ren_aa );
         }
     }
+    
+    ///render a non antialiased shape represented by an agg::path_storage
+    ///@param path the path containing the geometry to render
+    ///@param color fill color or null for no fill
+    void renderPathSolidFast(agg::path_storage &path, colorObj *color) {
+        ras_aa.reset();
+        ren_bin.color(msToAGGColor(color));
+        ras_aa.add_path(path);
+        agg::render_scanlines(ras_aa, m_sl_bin, ren_bin);
+    }
+    
+    
 
     ///render a fill pattern clipped by a shape
     ///this has only been tested when pattern is a stroke, and clipper a path.
@@ -720,8 +744,12 @@
     renderer_base ren_base;
     renderer_base_pre ren_base_pre;
     renderer_aa ren_aa;
+    renderer_prim ren_prim;
+    renderer_bin ren_bin;
     scanline sl;
+    agg::scanline_bin m_sl_bin;
     rasterizer_scanline ras_aa;
+    rasterizer_outline ras_oa;
     font_engine_type m_feng;
     font_manager_type m_fman;
     agg::rgba msToAGGColor(colorObj *c) {
@@ -762,6 +790,8 @@
     return pNewImage;
 }
 
+
+
 ///internally used function to get the cached AGG renderer.
 AGGMapserverRenderer* getAGGRenderer(imageObj *image) {
     return (AGGMapserverRenderer*)image->imageextra;
@@ -767,6 +797,7 @@
     return (AGGMapserverRenderer*)image->imageextra;
 }
 
+
 // ----------------------------------------------------------------------
 // Utility function to initialize the color of an image.  The background
 // color is passed, but the outputFormatObj is consulted to see if the
@@ -1483,7 +1514,10 @@
         else
             nwidth=width;
         //render the polyline with optional dashing
-        ren->renderPolyline(line,color,nwidth,symbol->patternlength,symbol->pattern);
+        if(symbol->type==MS_SYMBOL_SIMPLE && symbol->antialias==MS_FALSE && symbol->patternlength<=0)
+            ren->renderPolylineFast(line,color);
+        else
+            ren->renderPolyline(line,color,nwidth,symbol->patternlength,symbol->pattern);
     }
     else if(symbol->type==MS_SYMBOL_TRUETYPE) {
         //specific function that treats truetype symbols
@@ -1602,6 +1636,10 @@
 
     AGGMapserverRenderer* ren = getAGGRenderer(image);
     agg::path_storage polygon = shapePolygonToPath(p,0,0);
+    if(symbol->type==MS_SYMBOL_SIMPLE &&symbol->antialias==MS_FALSE) { // solid fill
+                ren->renderPathSolidFast(polygon,&(style->color));
+                return; // done simple case
+    }
     if(style->symbol == 0 || symbol->type==MS_SYMBOL_SIMPLE) {
         // simply draw a solid fill and outline of the specified colors
         if(MS_VALID_COLOR(style->outlinecolor))

---

[thomas bonfort]

oops typo...
and of course it'll work much better if your symbol uses ANTIALIAS
FALSE instead of ANTIALIASED FALSE ;)

On 9/18/07, thomas bonfort <[hidden email]> wrote:

> Tim,
>
> Attached is a patch you can apply to mapagg.cpp - if you can't apply
> it I'll mail you the whole file.
>
> it basically switches to aliased rendering if you use
> SYMBOL
> TYPE SIMPLE
> ANTIALIASED FALSE
> NAME 'aliased'
> END
>
> it'll work *only* on line and polygon layers, if you specify a
> STYLE
> NAME 'aliased'
> COLOR  r g b
> END
>
> nothing other is supported (line widths, or polygon outlines)
>
> I'd be interested to hear about timings if you can try it
>
> tb
>
>

[Nolte, Tim]

Thomas,

I'll give this a try, and see what I get. Thanks. I'll let you how it
works. Just so I get this straight I add "ANTIALIAS FALSE" to my layer
after the patch and that layer won't be antialiased, correct?


----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com


-----Original Message-----
From: thomas bonfort [mailto:[hidden email]]
Sent: Tuesday, September 18, 2007 12:31 PM
To: Nolte, Tim
Cc: [hidden email]
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

Tim,

Attached is a patch you can apply to mapagg.cpp - if you can't apply
it I'll mail you the whole file.

it basically switches to aliased rendering if you use
SYMBOL
TYPE SIMPLE
ANTIALIASED FALSE
NAME 'aliased'
END

it'll work *only* on line and polygon layers, if you specify a
STYLE
NAME 'aliased'
COLOR  r g b
END

nothing other is supported (line widths, or polygon outlines)

I'd be interested to hear about timings if you can try it

tb

[thomas bonfort]

On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> Thomas,
>
> I'll give this a try, and see what I get. Thanks. I'll let you how it
  > works. Just so I get this straight I add "ANTIALIAS FALSE" to my layer
> after the patch and that layer won't be antialiased, correct?

nope, that's not it..
you create the symbol definition in your symbol file or inlined in
your mapfile. and then you set the layers you want rendered fast to
LAYER
NAME blablba
DATA ...
 CLASS
  STYLE
   COLOR r g b
   SYMBOL 'aliased'
  END
 END
END

it's slower than gd but did speed up dense line drawings by a factor
of 3 when i tried it, less with polygons.

tb

>
>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
>
>
> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Tuesday, September 18, 2007 12:31 PM
> To: Nolte, Tim
> Cc: [hidden email]
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> Tim,
>
> Attached is a patch you can apply to mapagg.cpp - if you can't apply
> it I'll mail you the whole file.
>
> it basically switches to aliased rendering if you use
> SYMBOL
> TYPE SIMPLE
> ANTIALIASED FALSE
> NAME 'aliased'
> END
>
> it'll work *only* on line and polygon layers, if you specify a
> STYLE
> NAME 'aliased'
> COLOR  r g b
> END
>
> nothing other is supported (line widths, or polygon outlines)
>
> I'd be interested to hear about timings if you can try it
>
> tb
>

[Nolte, Tim]

Thomas,

You've totally made my day. I was finally able to get your patch applied
today. I tested it out and it works wonderfully. My mapfile went from a
whopping 17sec down to a little over 5sec. I don't know if this feature
would ever actually make it into the full release but it's a keeper as
far as I'm concerned.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com


-----Original Message-----
From: thomas bonfort [mailto:[hidden email]]
Sent: Tuesday, September 18, 2007 1:46 PM
To: Nolte, Tim
Cc: [hidden email]
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> Thomas,
>
> I'll give this a try, and see what I get. Thanks. I'll let you how it
  > works. Just so I get this straight I add "ANTIALIAS FALSE" to my
layer
> after the patch and that layer won't be antialiased, correct?

nope, that's not it..
you create the symbol definition in your symbol file or inlined in
your mapfile. and then you set the layers you want rendered fast to
LAYER
NAME blablba
DATA ...
 CLASS
  STYLE
   COLOR r g b
   SYMBOL 'aliased'
  END
 END
END

it's slower than gd but did speed up dense line drawings by a factor
of 3 when i tried it, less with polygons.

tb

>
>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
>
>
> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Tuesday, September 18, 2007 12:31 PM
> To: Nolte, Tim
> Cc: [hidden email]
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> Tim,
>
> Attached is a patch you can apply to mapagg.cpp - if you can't apply
> it I'll mail you the whole file.
>
> it basically switches to aliased rendering if you use
> SYMBOL
> TYPE SIMPLE
> ANTIALIASED FALSE
> NAME 'aliased'
> END
>
> it'll work *only* on line and polygon layers, if you specify a
> STYLE
> NAME 'aliased'
> COLOR  r g b
> END
>
> nothing other is supported (line widths, or polygon outlines)
>
> I'd be interested to hear about timings if you can try it
>
> tb
>

[thomas bonfort]

ok , I'll see how we can incorporate that...
how much time does the same map take to be rendered with gd ?

tb

On 9/19/07, Nolte, Tim <[hidden email]> wrote:

> Thomas,
>
> You've totally made my day. I was finally able to get your patch applied
> today. I tested it out and it works wonderfully. My mapfile went from a
> whopping 17sec down to a little over 5sec. I don't know if this feature
> would ever actually make it into the full release but it's a keeper as
> far as I'm concerned.
>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
>
>
> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Tuesday, September 18, 2007 1:46 PM
> To: Nolte, Tim
> Cc: [hidden email]
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> > Thomas,
> >
> > I'll give this a try, and see what I get. Thanks. I'll let you how it
>   > works. Just so I get this straight I add "ANTIALIAS FALSE" to my
> layer
> > after the patch and that layer won't be antialiased, correct?
>
> nope, that's not it..
> you create the symbol definition in your symbol file or inlined in
> your mapfile. and then you set the layers you want rendered fast to
> LAYER
> NAME blablba
> DATA ...
>  CLASS
>   STYLE
>    COLOR r g b
>    SYMBOL 'aliased'
>   END
>  END
> END
>
> it's slower than gd but did speed up dense line drawings by a factor
> of 3 when i tried it, less with polygons.
>
> tb
>
> >
> >
> > ----
> > Timothy J Nolte - [hidden email]
> > Associate Network Planning Engineer
> >
> > iPCS Wireless, Inc.
> > 4717 Broadmoor Ave, Suite G
> > Kentwood, MI 49512
> >
> > Office: 616-656-5163
> > PCS:    616-706-2438
> > Fax:    [phone]
> > Web: www.ipcswirelessinc.com
> >
> >
> > -----Original Message-----
> > From: thomas bonfort [mailto:[hidden email]]
> > Sent: Tuesday, September 18, 2007 12:31 PM
> > To: Nolte, Tim
> > Cc: [hidden email]
> > Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
> >
> > Tim,
> >
> > Attached is a patch you can apply to mapagg.cpp - if you can't apply
> > it I'll mail you the whole file.
> >
> > it basically switches to aliased rendering if you use
> > SYMBOL
> > TYPE SIMPLE
> > ANTIALIASED FALSE
> > NAME 'aliased'
> > END
> >
> > it'll work *only* on line and polygon layers, if you specify a
> > STYLE
> > NAME 'aliased'
> > COLOR  r g b
> > END
> >
> > nothing other is supported (line widths, or polygon outlines)
> >
> > I'd be interested to hear about timings if you can try it
> >
> > tb
> >
>
>

[Nolte, Tim]

The difference between AGG (w/ patch) verses straight GD was about 1.5-2
sec difference. This is acceptable for me.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com
-----Original Message-----
From: thomas bonfort [mailto:[hidden email]]
Sent: Thursday, September 20, 2007 6:37 AM
To: Nolte, Tim
Cc: [hidden email]
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

ok , I'll see how we can incorporate that...
how much time does the same map take to be rendered with gd ?

tb

On 9/19/07, Nolte, Tim <[hidden email]> wrote:
> Thomas,
>
> You've totally made my day. I was finally able to get your patch
applied
> today. I tested it out and it works wonderfully. My mapfile went from
a
> whopping 17sec down to a little over 5sec. I don't know if this
feature

> would ever actually make it into the full release but it's a keeper as
> far as I'm concerned.
>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
>
>
> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Tuesday, September 18, 2007 1:46 PM
> To: Nolte, Tim
> Cc: [hidden email]
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> > Thomas,
> >
> > I'll give this a try, and see what I get. Thanks. I'll let you how

it

>   > works. Just so I get this straight I add "ANTIALIAS FALSE" to my
> layer
> > after the patch and that layer won't be antialiased, correct?
>
> nope, that's not it..
> you create the symbol definition in your symbol file or inlined in
> your mapfile. and then you set the layers you want rendered fast to
> LAYER
> NAME blablba
> DATA ...
>  CLASS
>   STYLE
>    COLOR r g b
>    SYMBOL 'aliased'
>   END
>  END
> END
>
> it's slower than gd but did speed up dense line drawings by a factor
> of 3 when i tried it, less with polygons.
>
> tb
>
> >
> >
> > ----
> > Timothy J Nolte - [hidden email]
> > Associate Network Planning Engineer
> >
> > iPCS Wireless, Inc.
> > 4717 Broadmoor Ave, Suite G
> > Kentwood, MI 49512
> >
> > Office: 616-656-5163
> > PCS:    616-706-2438
> > Fax:    [phone]
> > Web: www.ipcswirelessinc.com
> >
> >
> > -----Original Message-----
> > From: thomas bonfort [mailto:[hidden email]]
> > Sent: Tuesday, September 18, 2007 12:31 PM
> > To: Nolte, Tim
> > Cc: [hidden email]
> > Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
> >
> > Tim,
> >
> > Attached is a patch you can apply to mapagg.cpp - if you can't apply
> > it I'll mail you the whole file.
> >
> > it basically switches to aliased rendering if you use
> > SYMBOL
> > TYPE SIMPLE
> > ANTIALIASED FALSE
> > NAME 'aliased'
> > END
> >
> > it'll work *only* on line and polygon layers, if you specify a
> > STYLE
> > NAME 'aliased'
> > COLOR  r g b
> > END
> >
> > nothing other is supported (line widths, or polygon outlines)
> >
> > I'd be interested to hear about timings if you can try it
> >
> > tb
> >
>
>

[Nolte, Tim]

I guess I'm not totally out of the dark yet. I didn't notice it but when
using the patch the legend images for the layers I've added the SYMBOL
'aliased' to don't show up now. Any tip? I'm assuming that this patch is
looking specifically for the "SYMBOL 'aliased'" so I can't simply add
more symbols that have the correct color. I think this might be the
problem because it's trying to use that symbol to make the legend image.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com

> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Tuesday, September 18, 2007 1:46 PM
> To: Nolte, Tim
> Cc: [hidden email]
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> > Thomas,
> >
> > I'll give this a try, and see what I get. Thanks. I'll let you how

it

>   > works. Just so I get this straight I add "ANTIALIAS FALSE" to my
> layer
> > after the patch and that layer won't be antialiased, correct?
>
> nope, that's not it..
> you create the symbol definition in your symbol file or inlined in
> your mapfile. and then you set the layers you want rendered fast to
> LAYER
> NAME blablba
> DATA ...
>  CLASS
>   STYLE
>    COLOR r g b
>    SYMBOL 'aliased'
>   END
>  END
> END
>
> it's slower than gd but did speed up dense line drawings by a factor
> of 3 when i tried it, less with polygons.
>
> tb

[Nolte, Tim]

Ok, here is my symbol and layer definitions.

SYMBOL
NAME 'aliased'
TYPE SIMPLE
ANTIALIAS FALSE
END

  LAYER
    GROUP "Propagations"
    NAME "-98dBm to -104dBm (Outdoor Rural)"
    STATUS ON
    TILEINDEX "propagations/104/tileindex"
    TILEITEM "location"
    TYPE POLYGON
    # --------------------------------------------------
    # Class: -98dBm to -104dBm (Outdoor Rural)
    # --------------------------------------------------
    CLASS
      NAME "-98dBm to -104dBm (Outdoor Rural)"
      MINSCALE 0
      MAXSCALE 1200000
      STYLE
        SYMBOL "aliased"
        COLOR 255 128 0
        OUTLINECOLOR 255 128 0
      END
    END
  END

Does anyone see something that isn't right? My default application
interface is Chameleon, however just using the basic cgi index.html &
viewer-template.html ways of rendering are outputting the same thing (no
legend images for layers using SYMBOL "aliased"). Prior to this patch
those legend images showed up fine. I'm pretty sure it has something to
do with adding a symbol to the class.

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com
-----Original Message-----
From: thomas bonfort [mailto:[hidden email]]
Sent: Thursday, September 20, 2007 10:28 AM
To: Nolte, Tim
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

I can't reproduce this here... the legend icon shows up correctly (i.e.
aliased)
how are you drawing the legend icons?

tb

On 9/20/07, Nolte, Tim <[hidden email]> wrote:
> I guess I'm not totally out of the dark yet. I didn't notice it but
when
> using the patch the legend images for the layers I've added the SYMBOL
> 'aliased' to don't show up now. Any tip? I'm assuming that this patch
is
> looking specifically for the "SYMBOL 'aliased'" so I can't simply add
> more symbols that have the correct color. I think this might be the
> problem because it's trying to use that symbol to make the legend
image.

>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
>
> > -----Original Message-----
> > From: thomas bonfort [mailto:[hidden email]]
> > Sent: Tuesday, September 18, 2007 1:46 PM
> > To: Nolte, Tim
> > Cc: [hidden email]
> > Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
> >
> > On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> > > Thomas,
> > >
> > > I'll give this a try, and see what I get. Thanks. I'll let you how
> it
> >   > works. Just so I get this straight I add "ANTIALIAS FALSE" to my
> > layer
> > > after the patch and that layer won't be antialiased, correct?
> >
> > nope, that's not it..
> > you create the symbol definition in your symbol file or inlined in
> > your mapfile. and then you set the layers you want rendered fast to
> > LAYER
> > NAME blablba
> > DATA ...
> >  CLASS
> >   STYLE
> >    COLOR r g b
> >    SYMBOL 'aliased'
> >   END
> >  END
> > END
> >
> > it's slower than gd but did speed up dense line drawings by a factor
> > of 3 when i tried it, less with polygons.
> >
> > tb
>

[Nolte, Tim]

Thomas,

Thanks for getting back to me on this. I sort of suspected this could be
the issue as I've seen the posts come across the group. I was sort of
impatient so I found a workaround using KEYIMAGE and just specifying an
image to use for my legend. I can't actually run in production without
the MIN/MAXSCALE options, but when I get a chance I'll run through this
on my test environment. Thanks again!

- Tim

----
Timothy J Nolte - [hidden email]
Associate Network Planning Engineer

iPCS Wireless, Inc.
4717 Broadmoor Ave, Suite G
Kentwood, MI 49512

Office: 616-656-5163
PCS:    616-706-2438
Fax:    [phone]
Web: www.ipcswirelessinc.com


-----Original Message-----
From: thomas bonfort [mailto:[hidden email]]
Sent: Wednesday, September 26, 2007 9:03 AM
To: Nolte, Tim
Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!

there's a bug in the 5.0 code concerning maxscale and minscale. could
you try without these keywords and report if your legends are drawing?
if that's the case the fix will be in 5.0.1

tb

On 9/21/07, Nolte, Tim <[hidden email]> wrote:

> Ok, here is my symbol and layer definitions.
>
> SYMBOL
> NAME 'aliased'
> TYPE SIMPLE
> ANTIALIAS FALSE
> END
>
>   LAYER
>     GROUP "Propagations"
>     NAME "-98dBm to -104dBm (Outdoor Rural)"
>     STATUS ON
>     TILEINDEX "propagations/104/tileindex"
>     TILEITEM "location"
>     TYPE POLYGON
>     # --------------------------------------------------
>     # Class: -98dBm to -104dBm (Outdoor Rural)
>     # --------------------------------------------------
>     CLASS
>       NAME "-98dBm to -104dBm (Outdoor Rural)"
>       MINSCALE 0
>       MAXSCALE 1200000
>       STYLE
>         SYMBOL "aliased"
>         COLOR 255 128 0
>         OUTLINECOLOR 255 128 0
>       END
>     END
>   END
>
> Does anyone see something that isn't right? My default application
> interface is Chameleon, however just using the basic cgi index.html &
> viewer-template.html ways of rendering are outputting the same thing

(no
> legend images for layers using SYMBOL "aliased"). Prior to this patch
> those legend images showed up fine. I'm pretty sure it has something
to

> do with adding a symbol to the class.
>
> ----
> Timothy J Nolte - [hidden email]
> Associate Network Planning Engineer
>
> iPCS Wireless, Inc.
> 4717 Broadmoor Ave, Suite G
> Kentwood, MI 49512
>
> Office: 616-656-5163
> PCS:    616-706-2438
> Fax:    [phone]
> Web: www.ipcswirelessinc.com
> -----Original Message-----
> From: thomas bonfort [mailto:[hidden email]]
> Sent: Thursday, September 20, 2007 10:28 AM
> To: Nolte, Tim
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
>
> I can't reproduce this here... the legend icon shows up correctly

(i.e.
> aliased)
> how are you drawing the legend icons?
>
> tb
>
> On 9/20/07, Nolte, Tim <[hidden email]> wrote:
> > I guess I'm not totally out of the dark yet. I didn't notice it but
> when
> > using the patch the legend images for the layers I've added the
SYMBOL
> > 'aliased' to don't show up now. Any tip? I'm assuming that this
patch
> is
> > looking specifically for the "SYMBOL 'aliased'" so I can't simply
add

> > more symbols that have the correct color. I think this might be the
> > problem because it's trying to use that symbol to make the legend
> image.
> >
> > ----
> > Timothy J Nolte - [hidden email]
> > Associate Network Planning Engineer
> >
> > iPCS Wireless, Inc.
> > 4717 Broadmoor Ave, Suite G
> > Kentwood, MI 49512
> >
> > Office: 616-656-5163
> > PCS:    616-706-2438
> > Fax:    [phone]
> > Web: www.ipcswirelessinc.com
> >
> > > -----Original Message-----
> > > From: thomas bonfort [mailto:[hidden email]]
> > > Sent: Tuesday, September 18, 2007 1:46 PM
> > > To: Nolte, Tim
> > > Cc: [hidden email]
> > > Subject: Re: [UMN_MAPSERVER-USERS] MapServer 5.0.0 released!
> > >
> > > On 9/18/07, Nolte, Tim <[hidden email]> wrote:
> > > > Thomas,
> > > >
> > > > I'll give this a try, and see what I get. Thanks. I'll let you

how
> > it
> > >   > works. Just so I get this straight I add "ANTIALIAS FALSE" to
my
> > > layer
> > > > after the patch and that layer won't be antialiased, correct?
> > >
> > > nope, that's not it..
> > > you create the symbol definition in your symbol file or inlined in
> > > your mapfile. and then you set the layers you want rendered fast
to

> > > LAYER
> > > NAME blablba
> > > DATA ...
> > >  CLASS
> > >   STYLE
> > >    COLOR r g b
> > >    SYMBOL 'aliased'
> > >   END
> > >  END
> > > END
> > >
> > > it's slower than gd but did speed up dense line drawings by a

factor
> > > of 3 when i tried it, less with polygons.
> > >
> > > tb
> >
>
>