[MS4W-Users] The "!BOX!" statement doesn't work since update to ms4w 4.0.5

[Jeff McKenna]

Actually, I bet your test case could instead be based on a Spatialite db 
(file). But that takes time for you to verify that you can trigger your 
error with a Spatialite db also.  Likely, it is worth the effort to 
create the test case based on a Spatialite db, as all devs will be 
easily able to trigger your 'box hackery' problem with a simple shp2img 
command (and not have to setup a PostGIS database locally).

Hope my steps help you.

And thanks!  (you can see how much effort I have to put in to create 
these test cases)

-jeff





-- 
Jeff McKenna
GatewayGeo: MapServer Consulting and Training Services
co-founder of FOSS4G
http://gatewaygeo.com/


On 2021-02-10 4:03 p.m., Jeff McKenna via MS4W-Users wrote:
> Hello Evgeni,
> 
> You may have found a good one (this was an old hack literally named in 
> the source code "Box hackery" ha).  This will be tricky to reproduce, 
> but I am sure that you can help with this part :)
> 
> Please submit this as a new ticket at 
> https://github.com/MapServer/MapServer/issues  But most importantly, 
> please attach a small test case to that ticket (yes it can be done for 
> PostGIS). For example, your test case should be a zip file that contains:
> 
> 1) mapfile with 1 layer
> 2) commands.txt (containing the commands to reproduce, in this case the 
> createdb commands, load postgis extension, and command to call the 
> external SQL file, and the final shp2img command to throw error)
> 3) your SQL file (that is the create table command, for your sample 
> polygon geometry)
> 4) state your exact MapServer version (that fails, and the last version 
> that also works for you)
> 5) you could include a working full query (from the last MapServer 
> version that worked) by checking your PostgreSQL log file (hint: set 
> log_statement = 'all' in your postgresql.conf)
> 
>     or also check that (for the failing MapServer version) the 
> PostgreSQL log file does not in fact record any hit for that query 
> attempt (this is a sign that MapServer is failing).
> 
> Thanks, this may be a good find (I'm not 100% sure at the moment, until 
> I test with your nice test case).
> 
> -jeff
> 
> 
> 
> 
>