Opened 18 years ago

Closed 18 years ago

#1653 closed defect (wontfix)

shp2img : Graphic output corrupted with version 4.8.1 but correct with version 4.8.0-beta2

Reported by: abajolet@… Owned by: sdlime
Priority: high Milestone:
Component: MapServer CGI Version: 4.8
Severity: major Keywords:
Cc:

Description

Hello

The first joined image is generated with MapServer (shp2img) 4.8.0-beta2, the
output is correct.
The second joined image is gerenated with MapServer (shp2impg) 4.8.1, the output
is corrupted on some lines : Strange kind of antialiasing when using an STYLE...
ANTIALIASING statement on LINE layer with a vector symbol.

Mapfile and Symbolset files are *exactly* the same for the two images. The only
change is the MapServer version.

Mapfile extracts :

OUTPUTFORMAT
  NAME png
  DRIVER "GD/PNG"
  MIMETYPE "image/png"
  IMAGEMODE RGBA
  TRANSPARENT on
  EXTENSION "png"
END

[.../...]

LAYER
  NAME "rivieres"
  DATA tr_hydro_lambert
  STATUS off
  TYPE LINE
  CLASSITEM "LARGEUR"
  LABELITEM "TOPONYME"

  # PROJECTION
            # "proj=latlong"
            # "ellps=GRS80"
  # END # end of the projection object

  CLASS
        NAME "Fleuve"
        EXPRESSION "Plus de 50 mètres"
        MAXSCALE 1000000
        COLOR 100 100 255
        OUTLINECOLOR -1 -1 -1
        SYMBOL "rond3"
        SIZE 3
        LABEL
            MINDISTANCE 400
            POSITION auto
            BUFFER 2
            ANGLE auto
            ANTIALIAS true
            TYPE truetype
            FONT times-bold-italic
            SIZE 10
            OFFSET 0 3
            COLOR 50 50 255
        END # end of label object
  END # end of class object

  CLASS
        NAME "Rivière large"
        MAXSCALE 500000
        EXPRESSION "Entre 15 et 50 mètres"
        STYLE # I Guess the problem is there
            COLOR 100 100 255
            OUTLINECOLOR -1 -1 -1
            SYMBOL "rond3"
            SIZE 2
            ANTIALIAS true
        END
        
        LABEL
            MINDISTANCE 400
            POSITION auto
            BUFFER 2
            ANGLE auto
            ANTIALIAS true
            TYPE truetype
            FONT times-bold-italic
            SIZE 8
            OFFSET 0 2
            COLOR 50 50 255
        END # end of label object
  END # end of class object

  CLASS
        NAME "Rivière"
        MAXSCALE 200000
        EXPRESSION "De 0 à 15 mètres"
        COLOR 100 100 255
        OUTLINECOLOR -1 -1 -1
        SYMBOL "rond3"
        SIZE 1
        LABEL
            MINDISTANCE 400
            POSITION auto
            BUFFER 2
            ANGLE auto
            ANTIALIAS true
            TYPE truetype
            FONT times-italic
            SIZE 7
            OFFSET 0 1
            COLOR 50 50 255
        END # end of label object
  END # end of class object
END # end of layer object

Symbolset extract : 

SYMBOL
  NAME "rond3"
  TYPE ELLIPSE
  POINTS 5 5 END
  FILLED true
  TRANSPARENT 1
END

Attachments (4)

4.8.0-beta2.png (105.0 KB ) - added by abajolet@… 18 years ago.
Correct output with 4.8.0-beta2
4.8.1.png (121.1 KB ) - added by abajolet@… 18 years ago.
Bad output with 4.8.1
10km-16-angouleme-512x384.png (181.3 KB ) - added by abajolet@… 18 years ago.
Bad output with 4.8.3
testcase.tar.gz (15.0 KB ) - added by abajolet@… 18 years ago.
Testcase to reproduce the bug

Download all attachments as: .zip

Change History (11)

by abajolet@…, 18 years ago

Attachment: 4.8.0-beta2.png added

Correct output with 4.8.0-beta2

by abajolet@…, 18 years ago

Attachment: 4.8.1.png added

Bad output with 4.8.1

comment:1 by abajolet@…, 18 years ago

Note : The two versions of MapServer are compiled in the same way, and ldd
reports  exactly the same .so files.

Configure directive used :

./configure --with-ogr \
--with-gdal \
--with-geos \
--with-proj \
--with-postgis=/usr/local/pgsql/bin/pg_config \
--with-freetype  \
--with-gd \
--without-tiff \
--with-pdf \
--with-ming \
--with-png \
--with-jpeg \
--with-zlib \
--with-php=/usr/local

ldd reports :

	linux-gate.so.1 =>  (0xffffe000)
	libgd.so.2 => /usr/local/lib/libgd.so.2 (0xb7f64000)
	libjpeg.so.62 => /usr/lib/libjpeg.so.62 (0xb7f43000)
	libfreetype.so.6 => /usr/X11R6/lib/libfreetype.so.6 (0xb7ed8000)
	libpng12.so.0 => /usr/lib/libpng12.so.0 (0xb7eaf000)
	libiconv.so.2 => /usr/local/lib/libiconv.so.2 (0xb7dd7000)
	libpdf.so.1 => /usr/local/lib/libpdf.so.1 (0xb7d43000)
	libproj.so.0 => /usr/local/lib/libproj.so.0 (0xb7d0d000)
	libgdal.so.1 => /usr/local/lib/libgdal.so.1 (0xb7958000)
	libming.so => /usr/lib/libming.so (0xb7918000)
	libpq.so.4 => /usr/local/pgsql/lib/libpq.so.4 (0xb78fd000)
	libgeos.so.2 => /usr/local/lib/libgeos.so.2 (0xb7806000)
	libm.so.6 => /lib/libm.so.6 (0xb77e3000)
	libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xb7715000)
	libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0xb770c000)
	libc.so.6 => /lib/libc.so.6 (0xb75f2000)
	libdl.so.2 => /lib/libdl.so.2 (0xb75ee000)
	libz.so.1 => /usr/lib/libz.so.1 (0xb75df000)
	librt.so.1 => /lib/librt.so.1 (0xb75d6000)
	libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0xb759b000)
	libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0xb747a000)
	libcrypt.so.1 => /lib/libcrypt.so.1 (0xb744c000)
	libresolv.so.2 => /lib/libresolv.so.2 (0xb743a000)
	libnsl.so.1 => /lib/libnsl.so.1 (0xb7425000)
	/lib/ld-linux.so.2 (0xb7fc5000)
	libpthread.so.0 => /lib/libpthread.so.0 (0xb7412000)

comment:2 by sdlime, 18 years ago

Status: newassigned
Any chance you can build a test case (mapfile and shapefile) that generates the
problem? Hard to debug from just a mapfile snippet.

Steve

comment:3 by abajolet@…, 18 years ago

I will try to create one before the end of the W-E.

comment:4 by sdlime, 18 years ago

Component: MapScriptMapServer CGI
Antoine: Is this still an issue with 4.8.3?

Steve

by abajolet@…, 18 years ago

Bad output with 4.8.3

comment:5 by abajolet@…, 18 years ago

Steve : Yes, it is.
I still use the mapserver-4.8.0-beta2 version.

[Sorry but i lacked time to the testcase]

by abajolet@…, 18 years ago

Attachment: testcase.tar.gz added

Testcase to reproduce the bug

comment:6 by sdlime, 18 years ago

There was a bug in GD that was exposed around that time. I posted a fix to the 
users list (a fix for GD). Any chance you applied that? I can dig it out if 
necessary.

Steve

comment:7 by sdlime, 18 years ago

Resolution: wontfix
Status: assignedclosed
This is a GD error, marking as WONTFIX for the moment.
Note: See TracTickets for help on using tickets.