[mapserver-msautotest]problem with shp2img on ogr_wkt.map in msautotest/misc directory

[nsavard@…]

When I execute shp2img on msautotest/misc/ogr_wkt.map on Windows, Windows
reports that a problem occured and cannot complete the request.

I work with ms4w 2.2 and the full command is:

shp2img -m ogr_wkt.map -o result/ogr_wkt.png

I tried also with ms4w 1.6 and the problem still occur.  I'll dig into the map
file to see what exactly is the problem.  But I suspect a compilation problem
because it works with a previous build Assefa provided me.

[nsavard@…]

I dug into the map file and the problem is related to the following block:

    FEATURE
      WKT # with any of POLYGON, POINT, MULTIPOINT, LINESTRING or MULTILINESTRING 
    END

[nsavard@…]

Could have a look to this one Jeff?  I think this related to ms4w build process.

[jmckenna@…]

Norm was the Assefa build that you mentioned based on mapserver 4.10.0 as well?
 Did you try older MS4W builds?  (1.6 is new)

[jmckenna@…]

BTW i see lots of WKT support changes listed in the MapServer 4.10.x release
notes
(http://www.maptools.org/ms4w/index.phtml?page=RELEASE_mapserver_ms4w2.x_dev.html).

[nsavard@…]

RE:  comment #3:  Yes Assefa's build is based on 4.10 release.

I'll check with a former ms4w.

[nsavard@…]

More info:

Same problem with ms4w 1.4.1 (Apache 2.0.55), 1.5.1 (Apache 2.0.55) and 2.1
(Apache 2.0.58).

But it works with ms4w 2.1 and with Assefa's build: 
mapserver-4.10.0_php5.1.4_gdal_cvs.zip

I could send you offline this zip file.

[jmckenna@…]

i am confused, which bug are you referring to? 

[jmckenna@…]

in terms of this bug only (WKT issue), are you saying that the problem exists
with all of those older MS4W versions except 2.1 and Assefa's build?

[nsavard@…]

Yes this is the case.  I was surprised as well.

RE:  comment #7:  Sorry but I don't understand your question.

[jmckenna@…]

Maybe it is possible that MS4W 2.1 used GDAL cvs by accident, and since Assefa's
build was based on GDAL cvs (and the other MS4W builds are GDAL 1.3.2 release)
maybe the WKT issue was fixed in GDAL cvs?  I'll have to check that.

Comment #7 explanation - you replied to both bug#1940 and bug#1941 with a single
message, and I have no idea what parts of the message are for which bug.  Please
specify further in bug#1941 for the line feeds issue.

[nsavard@…]

RE:  Comment #10:  I tested both issues with former ms4w and I specified in each
bug the result of my investigation. It appears that for each issue I came out
with the same result.  I investigated to know for which ms4w packages I was able
to replicate the problem.

I would add a note to bug 1941.

[jmckenna@…]

odd: with ms4w 2.2, ogr_wkt.map works with php_mapscript, but fails with shp2img

I even added CONFIG settings to the mapfile for the GDAL /data/ and /gdalplugins
folders just in case..and added those manually to the path...problem remains at
commandline.

Why would it work with php_mapscript and not at commandline, other than for
paths...i have no idea.

[jmckenna@…]

Well it looks like my guess in comment#10 was correct - after much pain, I
re-built GDAL with CVS (I never could get GEOS to work with GDAL CVS so i gave
up on that for now) and recompiled MapServer 4.10.0 - and yes ogr_wkt.map works
with shp2img.  

MS4W 2.1 probably had GDAL cvs by accident, which explains why other MS4W
versions did not work with this WKT test case (they contain GDAL 1.3.2).

However i am hesitant to include a CVS version of GDAL in MS4W (you can see my
problems building already).  I've filed an MS4W bug for this
(http://bugzilla.maptools.org/show_bug.cgi?id=1606).  But I am open for suggestions.

[nsavard@…]

RE:  comment #13:

Good work Jeff.  I think we should not include GDAL cvs in the ms4w package.  Ok
the WKT test does not pass but we know this problem is fixed in GDAL cvs and it
will disappear when you will include the new GDAL version in a future ms4w release.

I suggest that we keep this bug opened as a reminder to what the problem is.

[jmckenna@…]

note that the problem still exists with MS4W 2.2.2 (which contains GDAL 1.4.0).

So I have no idea what is going on here.  I guess the next test would be to try
MapServer CVS-head.

confusing.

[unicoletti]

Probably the same issue fixed by #2155?

[jmckenna]

Umberto is correct, I have no problems with that test with today's svn-trunk.