[Mapserver-users] WMS-GetMap problem in 4.0 (works in 3.6.6)

[Frieso ter Haseborg]

Hi,

okay I'm persuaded, no more relative pathes in the IMAGEPATH! And - well
- there are notes about this, at least in some tutorials. But I thought
being clever with using relative ones as a more flexible configuration.
;-)

Thx
Frieso

 
Frieso ter Haseborg
- Software Development - 
 
--

-------------------------------------------------- 
S&F Datentechnik GmbH&Co.KG 
Reimersstr. 41b 
26789 Leer, Germany


>-----Original Message-----
>From: Carl Anderson [mailto:carl.anderson at vadose.org]
>Sent: Friday, August 29, 2003 4:40 AM
>To: Frieso ter Haseborg
>Subject: Re: [Mapserver-users] WMS-GetMap problem in 4.0 (works in
>3.6.6)
>
>
>On 08/28/2003 06:24:11 AM, Frieso ter Haseborg wrote:
>
>> >Frieso ter Haseborg wrote:
>> >The handling of relative paths has changed between versions
>> >3.6 and 4.0.
>> >  It may be that you are running into a specific exception (or used
>> to
>> >take advantage of an exception).  If you consider this to be a bug
>> you
>> >could file it in bugzilla with a sample mapfile and a description of
>> >your installed directory structure (i.e. where the mapfile is
>> located,
>> >where you expect ../tmp to be located, etc.) and we could
>> investigate.
>> 
>> Hmm, still not sure - perhaps only a problem in a newbees head. ;-)
>>
>
>As a user who liberally used to use relative paths in mapfiles, I can  
>attest that there are problems.   
>http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=328
>(posted 5/25/2003, clarified today).
>
>But while playing with the PyIMS script I found that a relative  
>IMAGEPATH is a bad idea.
>
>why?
>
>because CGI accesses FONTSET, SYMBOLSET objects in relation to 
>the the  
>mapfile (relative links ok).  But IMAGEPATH is accessed in 
>relation the  
>the mapfile only for the map image but not for the reference map nor  
>scalebar nor legend when requesting mode=browse
>
>MapScript utilities are accessing FONTSET, SYMBOLSET, and IMAGEPATH  
>objects in relation to their own script.
>
>So absolute paths are the best, because they work in both  
>circumstances.  I still think that a note in the docs can help.
>
>
>
>C.
>
>
>