wms raster rquest crash!

Haiyan Liu hliu at JDBARNES.COM
Wed Sep 29 14:39:30 EDT 2004


Hi all,

The problem was fixed. My special thanks go to Frank who spent several
hours digging that 'bug' out. It happened because we compiled mapserver
with pdf support. The libpdf has its own libtiff which is different than
the one coming with gdal. So when I tested with wms raster display, it
crashed. (do not understand why regular mapfile was rendered ok?).
Anyways,Frank has a better explanation on this:

http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=888

It is such a tricky stuff and did give my colleagues and myself some
really hard time. Thanks again, Frank.

Haiyan

On Wed, 2004-09-15 at 13:59, Haiyan Liu wrote:
> Hi Frank and Ed, thank you for your responses.
> We are currently using GDAL 1.1.9.0, released 2003/06/27.
>
> Up till I posted this problem, our Admin has helped us try all the
> latest gdal versions in no vail. We do have successfully set up the the
> same wms on the same box and another box before. But since we recompiled
> everying of mapserver from the latest releases, we encountered the
> current problem. And now even we tried switching back to some earlier
> releases, it was still not working(well, it really should).
>
> Not sure whether or not data could be causing this the problem because
> we used the same data before. Here is the file anyways.
>
>         www.firstbasesolutions.com/tmp_download
>
> Sincerely,
>
> Haiyan
>
> On Wed, 2004-09-15 at 10:15, Frank Warmerdam wrote:
> > Haiyan Liu wrote:
> >   >>From this list, we noticed some postings on this problem. As some folks
> > > suggested that SegFault problem could be traced to mapserver's shp2img, so
> > > we tried shp2img and did get that segfault crash. When we tried using gdb to
> > > 'debug' the shp2img util with the same mapfile, we found it crashed on
> > > tif_read, here is the message before it died,
> > >
> > > ##########################################################################
> > > (gdb) run -m /var/www/html/zoom2it_mapserver/map_file/test_wms.map
> > > Starting program: /root/build/mapserver-4.0.1/shp2img
> > > -m/var/www/html/zoom2it_mapserver/map_file/test_wms.map
> > > (no debugging symbols found)...
> > > (no debugging symbols found)...
> > > (no debugging symbols found)...
> > > (no debugging symbols found)...
> > > (no debugging symbols found)...[Thread debugging using libthread_dbenabled]
> > > [New Thread -1084655424 (LWP 5180)]
> > > Program received signal SIGSEGV, Segmentation fault.
> > > [Switching to Thread -1084655424 (LWP 5180)]0x00a036ad in TIFFFillStrip
> > > (tif=0x8863628, strip=1) at tif_read.c:252252
> > > bytecount = td->td_stripbytecount[strip];
> > > ##########################################################################
> > >
> > > I hope i have explained my problem clearly. Any help/hint will be greatly
> > > appreciated! By the way, almost all mapserver versions above 4.0.1 have been
> > > triedwith the same configs as below and failed exactly the same way.
> >
> > Haiyan,
> >
> > When reporting crashes within gdb it is also helpful to do a "where" command
> > so we can see the whole call stack.
> >
> > This seems like it might be a bug in GDAL.  I would encourage you to report
> > it in the MapServer Bugzilla and ensure it is assigned to me.  You will need
> > to provide a dataset (ideally a smallish raster file that also demonstrates
> > the problem). Also, ensure you provide details on the version of GDAL you
> > are using.
> >
> > Best regards,
> --
> First Base Solutions Inc.
> 145 Renfrew Drive
> Suite 160
> Markham ON  L3R 9R6
>
> T. 905.477.3600
> F. 905.477.3882
--



More information about the mapserver-users mailing list