I would agree with Edi and recommend TIFF. It's easy to move around and
deal with, and is the fastest of the raster formats. (unless something
new has come up?)
I would also agree with Mark: a tileindex is for mosaicing many TIFFs,
and if your result is 1 single TIFF the tileindex is
I am unable to compile PHP MapScript 5.4 The build of mapserv and other
binaries works A-OK, but not PHP MapScript. It bails with a bunch of
linker errors, as if PHP couldn't be found or couldn't be loaded...
I can compile PHP MapScript on 5.2 and 5.0 and 5.6 just fine, if that's
a clue. Just
/var/www/cgi-bin/php: symbol lookup error:
/usr/lib/php/extensions/php_mapscript.so: undefined symbol: php_regcomp
Premature end of script headers: php
Any thoughts on this?
I was poking around today, and noticed that the PHP being used was
erroneously compiled as a module, and was thus
On 2/3/2011 12:55 PM, Daniel Morissette wrote:
I had not replied because I
didn't have much advice to provide other than make sure the
--with-regex=system switch really kicks in, sometimes it doesn't...
but maybe I should have given you at least that pointer.
Thanks to the long memory of the
On 2/3/2011 12:59 PM, Daniel Morissette wrote:
are you sure there you didn't trim any warning/errors in-between
the two that could provide a hint?
Positive.
Following, instead of my signature, is the whole output from
cd mapscript/php3 ; make
gcc -fPIC -O2 -fPIC -Wall-DCOMPILE_DL=1
On 2/3/2011 1:08 PM, Daniel Morissette wrote:
Notice the -I -c -o ... sequence
-I expects an argument which is an include directory
Sure, of course. And a blank one would be some sort of bug in the
Makefile construction process. I'll get back to you in a little bit,
after I've tracked down
I think I found it, see below:
-I/usr/include/php/TSRM -I -c -o php_mapscript_util.o
php_mapscript_util.c
Notice the -I -c -o ... sequence
-I expects an argument which is an include directory
This seems to correspond to $(PHP_REGEX_INC)
This is defined in mapscript/php3/Makefile as
Looks like this very bug has already been reported, and fixed in 5.6 and
in the upcoming 5.4.3
http://trac.osgeo.org/mapserver/ticket/3078
Good eye, Daniel; you're 2 for 2 so far!
--
HostGIS, Open Source solutions for the global GIS community
Greg Allensworth - SysAdmin, Programmer, GIS
Hey all.
I am hitting upon a weird problem with PHP MapScript: it will load
without issue, but crashes when trying to open a mapfile.
/var/www/cgi-bin/php: symbol lookup error:
/usr/lib/php/extensions/php_mapscript.so: undefined symbol: php_regcomp
Premature end of script headers: php
I've
To clarify GD is not being removed.
So, rendering issues with GD are and will be of concern.
OK, sounds good. Thanks for clarifying.
--
HostGIS, Open Source solutions for the global GIS community
Greg Allensworth - SysAdmin, Programmer, GIS Person, Security
Network+ Server+ A+
I came into this conversation a bit late, but...
So you're tiling, and using MS as your backend. Are you using TileCache
or similar in between? If so, TileCache supports meta-tiling which may
help your label problems significantly.
Using meta-tiling, TileCache requests a much larger tile
I was using TileCache, but I switched it off to debug this problem. If
TileCache generates larger tiles, will there still be clipped labels?
Just fewer clipped labels? I don't think this is a workable solution
for me, but if I have the time I'll try it.
Correct; there would be FEWER labels
On 1/26/2011 7:55 AM, Mike Stoddart wrote:
Thanks again. I don't think MapServer will work for me,
but thanks for your help.
Sure, you're quite welcome.
I am curious what else you would use that would eliminate tile-edge
artifacts. What map-serving software would you use, for example? I'm
On 1/26/2011 8:25 AM, Mike Stoddart wrote:
I set metaBuffer to 20 (I have SIZE=10) and there's no difference.
metaTile=yes
metaBuffer=20
Try metaTile=true, without the marks.
And of course, erase any existing cached tiles, so you're sure to be
generating new ones. I hate when I forget that;
Let's see:
For starts, set PARTIALS FALSE and FORCE FALSE in the mapfile. That
would prevent truncated labels, which would be worse than missing
labels. (presumably)
You may also want to add MAXSIZE 2048 to the mapfile, an element of the
MAP object. This would allow 5x5 (1280) tiles, as
My problem is that I need ALL labels to be complete and visible.
I don't think that's possible if you're using tiles. Somewhere, some
label will be close to a tile boundary and be truncated or omitted.
If you find a mapping engine that can place labels like that (drawing
half of a label on
On 1/26/2011 9:05 AM, Mike Stoddart wrote:
I admit I haven't checked but does Google et al suffer the same problem?
If you mean overlaying your data atop Google, say a small dataset
expressable as KML, that should work A-OK. If it's pure vector data
being rendered client-side, tiling
A while back someone, maybe Thomas Bonfort, did some performance tests
comparing GD to AGG, and if I recall AGG was only slightly slower and in
at least a couple of cases it was faster the GD.
In our case, GD renders in 1 second and AGG in 3 in these few tests
cases we're concerned with. We're
Hey guys. I have an interesting rendering issue here.
We are using the GD renderer for this project, since it's faster than
AGG to the degree that we want to offer it as a speedy option. But the
polygon outlines look bad in some cases, especially circles.
The center one in black, we render
This fella describes a hack he made to MapServer's source, to give him a
fuzzy matching for the OFFSITE value:
http://sites.google.com/site/bpederse/caliwms
It may help you out. It's a fairly simple set of changes.
--
HostGIS, Open Source solutions for the global GIS community
Greg
On 9/17/2010 10:38 PM, Mark Korver wrote:
I think your halos are blacks that are not pure black.
Not 0/0/0, just close.
Thanks so much for the link to bpederse's article. The simple addition
of OFFSITE 0 0 0 replicated the effect I wanted, and with his
fuzziness hack this will be even
Hey guys.
I have some NAIP imagery in MrSID format. It displays, but the black
area where counties meet is really ugly. The MrSIDs are 3 one-byte
bands: R/G/B The bands do not specify a NODATA but I figure that 0/0/0
(all black) may be usable as a NODATA value, via a PROCESSING directive.
Hey guys. Still no luck on this one, thought maybe I'd ask again.
The compile gets as far as linking, then bombs. The error seems to be
related to AGG, and if I give --with-agg=no the compiles finished
successfully. Though, of course we do want AGG!
The system does not have AGG installed at
Jump on my server gis.imaptools.com (it is Debian).
I just downloaded 5.6.5 and built it with no problem.
Good, good. Too bad I can't advise the customer to throw out everything
he's ever done and go with Debian. (ha ha) In the meantime, I used
--with=agg=no, so he can at least start testing
Hey Stephen, nice to hear from you. :)
You should not need to download, build and install agg anymore.
Sounds good. I have uninstalled it; that is, deleted libagg.a and the
include files.
But that doesn't solve the build issue. I still get that linking error:
In your mapserver build dir do:
make clean
./configure
make
Yeah, I know to do make clean before trying again. It hasn't helped.
also check to see if you have an errand mapserver.a somewhere
locate mapserver.a and remove them.
Good thinking, in case one somehow got placed somewhere.
Hey guys. Long time, no chat. :)
I'm trying to build 5.6.5 on Slackware/Slamd64. I have AGG installed (as
best one can since it doesn't do make install, copied libagg.a and the
include files). If I do not use --with-agg=no, MapServer's build process
tries to use AGG and fails miserably:
g++
Select ST_Transform(gpscoords,2276) from gpsdata;
But it is throwing error :
ERROR: transform: couldn't project point:-14 (latitude or longitude
exceeded)SQL state:XX000
It means what it says: one of your points has a invalid ordinate, so it
cannot be transformed.
Try this to debug it.
MS_MAPFILE=/var/www/html/theDir/theFile.map
export MS_MAPFILE
QUERY_STRING=map=${MS_MAPFILE}zoomdir=0zoomsize=2layer=countieslayer=states...
/var/www/cgi-bin/mapserv
it accesses the
mapfile in /theDir/, and /theDir is supposed to be password protected
now by Apache.
Correct. But the
Latest versions of MapServer allow you to set an env variable called
MS_MAPFILE_PATTERN
Holy cow!
SetEnv MS_MAP_NO_PATH 1
SetEnv WMS1_MAPFILE 'some path'
SetEnv WMS2_MAPFILE 'some other path'
Wow! Wow!
Super cool. I hadn't even heard of these. Thanks for the tip!
--
HostGIS, Open
Don't use MapServer's built-in TIFF support; GDAL's is a lot nicer
anyway. GDAL's even has BigTIFF if you build GDAL using its internal
support for most formats, which is what I'd advise at that stage too.
--
HostGIS, Open Source solutions for the global GIS community
Greg Allensworth -
Are any plans to include compressed GML as an output in the windows version
of mapserver?
Typically that would be handled silently by your webserver, if the
browser supplies an Accept-encoding header including deflate or gzip.
Alternately, you can have more control over the process with some
Hey Gonzalo. Still wrestling with that puppy, huh? We'll get it hammered
into place. :)
Warning: [MapServer Error]: msLoadMap(): Undefined symbol
Hidro.shp-0-0 in class 0, style 0 of layer Hidrografia. in
/home/orsep/public_html/gis2/incphp/globals.php on line 60
Undefined symbol means
And the new error:
Warning: [MapServer Error]: msLoadSymbolSet(): First token must be
SYMBOLSET, this doesn't look like a symbol file. in
/home/orsep/public_html/gis2/incphp/globals.php on line 60
That means exactly what it says: symbol files must now start with
SYMBOLSET and end with END.
Warning: dl() has been disabled for security reasons in
/home/orsep/public_html/test/pmapper-stable/incphp/globals.php
on line 51
That says it all: Your PHP installation does not allow you to load PHP
modules, such as MapScript. So running a MapScript-based app, just isn't
going to happen
dl() its allow in the php.ini the extension=php_mapscript.so load ok
I'd have to disagree.
If you call dl() you're told that you're not allowed to use dl() That
means that dl() is not allowed in the php.ini
If you leave it out, mapscript is obviously not loaded via an extension=
line in
I'm a big fan of Komodo. Not only is it free, but it handles FTP and
SFTP connections. I dig it.
--
HostGIS, Open Source solutions for the global GIS community
Greg Allensworth - SysAdmin, Programmer, GIS Person, Security
Network+ Server+ A+ Security+
jchampagne2 wrote:
Thank you! I did a little research, and writing a .bat file to execute as a
scheduled task is exactly what I needed
Cool.
del *.png
Just for future planning, you may want to add *.jpg and *.gif to it.
Otherwise, three years from now when you're also generating JPEGs
Jeff McKenna wrote:
- wherever possible, avoid using special characters in the NAME and
GROUP parameters, because these might/will cause grief for OGC services
Yeah, that's the part what seems most problematic!
This isn't going to get what you want!
LAYERS=parksrecreationSRS=EPSG:4326
This
Andrew Ayre wrote:
Hi! I can't seem to get AGG to work. Currently my output is not
antialiased. Can anyone give me some hints? Thanks!
loadOutputFormat(): General error message. OUTPUTFORMAT clause
references driver AGG/PNG, but this driver isn#39;t configured.
That says it all: Your
Andrew Ayre wrote:
Now I have SUPPORTS=AGG. Also the error has gone. But the output isn't
anti-aliased. Is there something else that I need to add
OUTPUTFORMAT
NAME 'AGG_Q'
DRIVER AGG/PNG
IMAGEMODE RGB
FORMATOPTION QUANTIZE_FORCE=ON
FORMATOPTION QUANTIZE_DITHER=OFF
FORMATOPTION
all I need to do is recompile the existing Apache installation
No, no; no need for that. MapServer is a simple CGI program, so if your
Apache has support for CGI programs you're good to go. And while support
for CGI can be disabled, it's probably enabled.
Anyone experienced in installing
I want
to confirm that LAYER NAME is not supported in run-time substitution
before I re-work everything.
It sure isn't, confirmed.
A recompile, though, should be able to enable it. If you have elementary
C skills and are used to compiling MapServer from source, that may
present little
I have a point in -8.0572,42.4702 (lon/lan). I like a bounding box
arround 1 mile of this site. Can use any SRS like EPSG:4230 or another.
Hit up spatialreference.org
You can look up a SRS, then move a point over a map to get conversions.
Very handy.
--
HostGIS, Open Source solutions for
Jennifer Shanks wrote:
FILTER (access_cod = 'Restricted Access' or access_cod = 'Open Access')
I suggested the parens as disambiguation, when I saw this in
PostgreSQL's query log.
WHERE code = 'Restricted' or code = 'Open' and gid = 12345
As such, her own FILTER clause seems just fine but
Steve Lime wrote:
Probably a result of some of the PostGIS driver work that was done for 5.4. A
bug would be
wise I think. Make sure to assign to the PostGIS component.
Done
http://trac.osgeo.org/mapserver/ticket/2937
--
HostGIS, Open Source solutions for the global GIS community
Greg
Steve Lime wrote:
Will probably be late in the week...
Okay. Thanks for the update.
--
HostGIS, Open Source solutions for the global GIS community
Greg Allensworth - SysAdmin, Programmer, GIS Person, Security
Network+ Server+ A+ Security+
___
Moen, Paul T. wrote:
Can anyone get this WMS layer to reproject to a projected coordinate
system like epsg:2266?
Add to your WEB block, this. You can then make requests to your
mapfile-as-WMS in the given SRS.
METADATA
wms_srs EPSG:2266
END
This does mean fetching the raster from them
John Mitchell wrote:
If instead of using compressed ECW images as input for MapServer WMS you
had GeoWebCache in front of MapServer which means that the WMS input is
cached tiled png's instead of ECW images from MapServer.
Then I think that you get around the ECW SDK license agreement,
Rui Gomes wrote:
I expect recived the follow error Message:
msLoadMap(): Unable to access file. (/testing_error/hello.map)
I would expect that from the command line. From a webserver, I would
expect a 500 response and to find that above message in the error_log
--
Gregor Mosheh / Greg
And a few more question: what libraries i need to put in the server?
Where i can find some documentation about the libraries mapserver
need?
A technique I find handy for debugging mapserver problems, a bit more
useful than Internal server error, is to go into the Unix shell and
use shp2img.
Hey guys. This picture says the first 1000 words:
http://maps.hostgis.com/gregor/cg-esri-compare.jpg
In this case it's a UTM projection for South America, and they're zoomed
out inappropriately, so the rest of the world looks ugly.
The client's request, is whether MapServer has a mechanism
Alter your geometry DATA selection to be ST_Union(geom,
bbox_of_projection)?
Not sure I get you. I see that union would trim to a square. What would
be the effective box of a given projection?
--
Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+
System Administrator,
Apache is crashing? Ouch.
Question: Are you using PHP/MapScript and PHP as a DSO instead of a CGI?
That has a habit of causing Apache to effectively freeze but not crash.
--
Gregor Mosheh / Greg AllensworthBS, A+, Network+, Security+, Server+
System Administrator, Lead Programmer
HostGIS
Amiya Patra wrote:
*PHP/MapScript and PHP has a habit of causing Apache to effectively freeze*
But how to solve that one...
If you suspect that that's the problem with your server, do what
everyone else does: compile and run PHP as a CGI program, not as a DSO.
It is
Perhaps I'm not clear on what constitutes 8-bit pseudocolored map
generation. I am using the AGG/PNG driver with TRANSPARENT ON.
FWIW, here is my favorite IMAGETYPE block for generating AGG PNGs with
transparency:
OUTPUTFORMAT
NAME png
MIMETYPE image/png
DRIVER AGG/PNG
Venkat Rao Tammineni wrote:
How can I retrieve layers by Group Name. And how to loop these
layers based on group. Where I can able to switch on and swatch off the
layers based on Group.
A group is just a fake layer with many layers. You request a group in
the same way as you would a
maxphe...@netscape.net wrote:
Anyone have a mapfile with the symbology corresponding to the OSM layers
shown on the new MapServer main page? TIA.
Ooh, that'd be neat. If so, can someone post it to the Symbology
Exchange pages of the website's wiki? We do have one google-style
streets
58 matches
Mail list logo